INT-2038
, Comisión Económica para América Latina y El Caribe
XII SEMINARIO REGIONAL DE POLITICA FISCAL
NACIONES

Organizado por CEPAL
Con el copatrocinio del FMI, Banco Mundial y BID
y el auspicio del Ministerio de Hacienda de Chile
Santiago de Chile, 24 al 26 de enero del 2000

C E P A L

XII SEMINARIO REGIONAL DE POLITICA FISCAL
COMPENDIO DE DOCUMENTOS
2000

UNIDAS

CEPAL
Comisión Económica para América Latina y El Caribe
XII SEMINARIO REGIONAL DE POLITICA FISCAL
Organizado por CEPAL
Con el copatrocinio del FMI, Banco Mundial y BID
y el auspicio del Ministerio de Hacienda de Chile
Santiago de Chile, 24 al 26 de enero del 2000

mm

NACIONES

C E P H L

XII SEMINARIO REGIONAL DE POLITICA FISCAL
COMPENDIO DE DOCUMENTOS
2000

UNIDAS

XII SEMINARIO REGIONAL DE POLITICA FISCAL
Organizado por CEPAL con el copatrocinio del FMI, Banco Mundial y BID
y el auspicio del Ministerio de Hacienda de Chile
Santiago de Chile, 24 al 26 de enero del 2000

El programa del XII Seminario Regional de Política Fiscal refleja los criterios que tradicionalmente han
presidido la selección y tratamiento de temas en anteriores ediciones de este foro anual:
i)

propiciar una visión conceptual amplia respecto al contenido y gama de problemas por considerar
en la formulación, diseño e implementación de la políticafiscal, como también en la coordinación
y armonización de esta última con las demás políticas macroeconómicas, en particular con la
política monetaria y de administración de la deuda pública

ii)

focalizar en el examen de problemas, diagnósticos y soluciones referidas al campo fiscal de las
economías de América Latina y el Caribe, comparando experiencias internacionales, identificando
mejores prácticas y destacando los límites que las especificidades propias de cada realidad
imponen a la transferencia de enseñanzas de terceros países

iii)

promover el uso de diferentes perspectivas de análisis -macroeconómico, microeconómico o
institucional de manera flexible, según el caso- a fin de orientar el proceso de diagnóstico,
análisis y toma de decisiones gubernamentales respecto a las funciones de asignación de recursos,
estabilización, distribución y crecimiento económico

Tal amplitud de miras necesariamente exige que el Programa del Seminario abrigue una amplia
gama de tópicos, como, también, que estos puedan ser tratados mediante enfoques teóricos diversos y por
capacidades tecnico-profesionales no menos variadas. De la libre competencia de ideas en todos los planos
se espera suijan las luces que mejor podrán orientar la toma de decisiones gubernamentales en el plano
fiscal de cada país de nuestra región respecto a las convencionales funciones de asignación de recursos,
estabilización, distribución y crecimiento económico que son responsabilidad del sector gobierno.

Sesión 1
A la luz de recientes experiencias de acceso a los Mercados Financieros
por parte de Gobiernos Sub-Nacionales: ¿Qué diseño de Marco Regulatorio
sería más apropiado para evitar sobre-endeudamiento? Comando central?
Compatible con incentivos? Combinación de ambos?
Antecedentes
Existe una comprensible tensión entre la responsabilidad de las autoridades económicas de los gobiernos
centrales por asegurar equilibrios macroeconómicos (especialmente, ante la relativa falta de incentivos para
asegurar disciplina fiscal en los gobiernos sub-nacionales) y las aspiraciones de autonomía en los demás
niveles de gobierno (en especial, para decidir autónomamente sobre las formas definanciar sus compromisos
de gasto público).- Esto genera un conflicto entre oferta de regulación (originada en el gobierno central y
destinada a minimizar riesgos de sobre-endeudamiento subnacional) y demanda por regulación (vista como
lobby para legitimar el acceso de los gobiernos sub-nacionales a los mercadosfinanciero y, en la medida de
lo posible, para asegurar el beneficio de garantías públicas).
La orientación regulatoria hoy vigente en muchos países de la región tiene características propias de
una estructura de comando central.- Sin embargo, la creciente importancia política de los procesos de
descentralizaciónfiscal y el hecho de que numerosos gobiernos subnacionales de la región y de Europa, ya se
encuentran operando activa y exitosamente en los mercados de deuda, sugiere la conveniencia de repensar
dicho enfoque. Esto se ve reforzado por el hecho de que: (i) la discusión en curso acerca de una nueva
arquitectura del sistemafinanciero internacional, destaca que el sector privado deberá tener una mayor y
creciente responsabilidad para prevenir y resolver eventuales crisis futuras; (ii) el enfoque de comando
central que informa el marco regulatorio hoy vigente en muchos países de la región, sufre importantes y
conocidas debilidades (básicamente, asimetrías de información, experiencia y motivación, entre las
autoridades reguladoras y reguladas).Proyósito
Revisar la experiencia reciente de acceso de gobiernos subnacionales a los mercadosfinancieros y discutir las
condiciones bajo las que el referido enfoque de comando central podría ser complementado por diseños
regulatorios capaces de alinear los incentivos de los gobiernos subnacionales (para acceder a los mercados
financieros locales y externos), con los objetivos perseguidos por las autoridades regulatorias del gobierno
central (esto es, limitar el 'riesgo moral' y evitar sesgos hacia el sobre-endeudamiento provincial, estadual y/o
municipal).-

Sesión 2
Incentivos Tributarios para atraer Inversión Extranjera Directa:
Experiencias Recientes y Evaluación de Diseños Alternativos
Antecedentes
La IED hacia América Latina y el Caribe viene experimentando aumentos sin precedentes durante los 90's
(de poco menos de 10.000 millones de dólares en 1990 pasó a poco mas de 67.000 millones en 1998).- Entre
las explicaciones de tal auge destaca un sensible cambio de actitud en los países respecto al rol de la IED.En un notable ejemplo de cambio de paradigma, respecto al que prevalecía durante las primeras tres décadas

posteriores a la II Guerra Mundial, la IED es hoy percibida como una manera superior de dinamizar el
comercio internacional e impulsar el crecimiento económico.- En principio ella permite que la economía
receptora acceda a nuevos y privilegiados mercados (de insumos y productos), a tecnologías y asistencia
técnica más avanzadas, así como a nuevas formas de conocimiento (en materia de producción, administración,
comercialización, transporte, etc.), y por cierto, a la generación de nuevos empleos con productividad y
remuneraciones superiores a la media.- Ello explica en parte por qué las autoridades económicas de la región
están dispuestas a — de hecho se esfuerzan por- promover la radicación de la IED dentro de sus respectivas
y
jurisdicciones, mendiante la oferta de diversos tipos de incentivos.- Entre estos últimos destacan diversas
modalidades de incentivos tributarios horizontales y/o selectivos (por sectores, industrias o empresas).
De hecho, los gobiernos llegan a competir entre sí -de maneras a veces predatoria- para asegurar la
localización del mayor volumen posible de IED dentro de sus respectivas fronteras (para lo cual conceden
exenciones o vacaciones impositivas por diferentes horizontes temporales, deducciones especiales por cuenta
de depreciaciones aceleradas, tasas preferenciales, zonas libres de impuestos, etc.), descuidando hasta cierto
punto su composición .- Este último punto viene generando abundante reflexión en algunos países, donde los
incentivos tributarios están siendo diseñados no apenas para atraer IED, sino para asegurar que ésta se
canalice hacia actividades y/o sectores prioritarios.- Además, estos temas vienen despertando fuerte interés en
las llamadas 'economías pequeñas', que para competir exitosamente con las de mayor tamaño relativo son
impulsadas a atender demandas de las empresas responsables por la IE, con pesadas consecuencias en materia
de 'renuncias' tributarias para el Tesoro Nacional.- Esto último también puede llegar a tener un significativo
impacto sobre lasfinanzas públicas de los gobiernos subnacionales (y por carácter transitivo, sobre el Tesoro
Nacional) en economías grandes y de tradición política federal, donde la 'guerra' de incentivos entre niveles
de gobierno subnacionales ha venido ganando espacio hasta transformarse en un grave problema fiscal.Proyósito
Promover el análisis y discusión entorno a diversas modalidades de incentivos tributarios para atraer IED en
uso en la región, desde una perspectiva comparativa en cuanto al diseño de tales instrumentos, al costo fiscal
de su utilización y los resultados que vienen siendo alcanzados.-

Sesión 3
Administración de Riesgos y manejo de Pasivos Contingentes del Sector Público:
¿Cómo incorporar el tema de las Garantías Gubernamentales
-explícitas e implícitas- a la programación presupuestaria anual y plurianual?
Antecedentes
El sector gobierno, como cualquier agente económico, debe hacer frente a muchos tipos de imprevistos con
claras consecuenciasfinancieras.- Así cabe mencionar los pagos que el Tesoro público debe efectuar en cada
ejercicio fiscal como consecuencia de fallos judiciales adversos, o por reconocimiento de deudas
inadecuadamente contabilizadas en el pasado (ej. reconocimiento de 'esqueletos' fiscales), o por
indemnizaciones (ej. por despidos de trabajadores), compensaciones (ej. por expropiación de propiedades)
y/o reparaciones (ej. por daños privados causados por agentes públicos).- De igual modo, el Estado debe
hacerse cargo del pago de diversos seguros implícitos [ej. por quiebres de cosechas, inundaciones, terremotos
u otras catástrofes naturales, a los que debe adicionarse el 'rescate' de organizacionesfinancieras -estatales y
privadas- insolventes, agencias de gobiernos subnacionales técnicamente en bancarrota, o en general, de
'proyectos' que por diversos motivos cuentan con garantías gubernamentales implícitas (ej. la seguridad

social de grupos especialmente vulnerables)].- A lo anterior debe adicionarse los compromisos financieros
derivados de la oferta de avales u otros tipos de garantías públicas explícitas (ej. volúmenes mínimos de
ventas o tasas mínimas de rentabilidad a concesionarios privados de servicios de utilidad pública).Elementales principios de administración deriesgosy normas contables de aceptación universal, sugieren la
necesidad de que los pasivos contingentes del sector público reciban un tratamiento técnico especial, de
manera análoga a lo que el sector privado hace con los suyos- Ello incluye la oportuna y transparente
identificación, estimación y presupuestación de los pasivos latentes, así como el continuo monitoreo y
evaluación de su 'calidad' (afin de recalcular el valor esperado de los compromisosfinancieros a medida que
las circunstancias hagan variar la probabilidad de su efectivación).
Aun reconociendo que en algunos países se viene dedicando atención al tema (ej. a través de la
creación de fondos de estabilización, formación de reservas técnicas, provisiones por pérdidas derivadas de
garantías otorgadas, normas para el oportuno reconocimiento público de pérdidas, etc.), parece evidente que:
(i) falta mucho por avanzar en materia de administración deriesgosen el campo fiscal (especialmente si se
lo compara con las iniciativas y avances ya logrados en el campofinanciero); (ii) se conoce poco sobre lo
que los países de la región vienen haciendo al respecto (en principio, no parece constituir práctica corriente el
mantener adecuado registro de los compromisos contingentes por niveles de gobierno, según los tipos o
modalidades que adoptan, el perfil temporal de los valores involucrados y el valor estimado de la probabilidad
de que se hagan efectivos, etc.).Propósito
Promover un intercambio de ideas y experiencias sobre administración de riesgos asociados con pasivos
contingentes del sector público. Además, se busca conocer lo que constituirían 'mejores prácticas' para
identificar, estimar, supervisar e incorporar tales pasivos a los presupuestos públicos -anual y plurianualde diferentes niveles de gobiemo.-

Sesión 4
Papel, alcance y costos fiscales de las Redes de Protección Social.
Antecedentes
Durante la era de Bretton Woods — bajo supuestos convencionales a la Krugman-- una crisis de balanza de
y
pagos generalmente conducía a un ajuste fiscal contraccionario sin que las autoridades gubernamentales
hicieran mayores esfuerzos por neutralizar, o al menos compensar, las consecuencias sociales negativas que
de aquel se derivaban.- En aquellos tiempos, los problemas económico-financieros de un país eran vistos
como eventos básicamente aislados, producto de la implementación de malas políticas domésticas que
intentaban 'hacer vivir a la sociedad, más allá de sus medios'.- En dicha era, los organismos de cooperación
financiera internacional -cuando convocados a colaborar en la corrección de un desequilibrio en las cuentas
externas— usualmente recomendaban comprimir el gasto público (como un medio de reducir la absorción
global de recursos), dejando a cargp de las autoridades domésticas decidir la composición de los cortes.- En
los años 90 el escenario es bastante diferente al anterior.
En primer lugar, los déficit públicos dejaron de ser abiertamente elevados,frecuentes y generalizados
como en décadas precedentes. En segundo lugar, la combinación deflexibilidad cambiaría (propiciada por el
quiebre del modelo de Bretton Woods) y movilidad del capitalfinanciero internacional (promovida por las
políticas de liberalización y desregulaciónfinanciera iniciada en los años 80), tuvieron como resultante una

fuerte y generalizada volatilidad cambiaría que posibilitó crisis bancarias y financieras cuyos efectos reales
adversos son hoy bien conocidos.- La gravedad y escala internacional de las recesiones ocurridas en Asia y
América Latina a partir de 1997 se expresa en notables aumentos no transitorios de los niveles de desempleo
abierto (en el sector formal) y de subempleo (en el sector informal) de numerosos países.- En un corto periodo
de tiempo y en diferentes partes del mundo, muchos millones de personas volvieran a caer debajo de los
niveles de pobreza de los que habían conseguido escapar en décadas anteriores y de manera especial, durante
fase de rápido crecimiento registrada a comienzos de los 90's.
La simultaneidad y escala internacional con que se sintió el impacto social adverso de las crisis
financieras recientes: contribuyó a minar el paradigma de que los problemas económicos nacionales pueden
ser vistos como fenómenos básicamente aislados, fortaleció la preocupación de la comunidad internacional
con las posibles repercusiones sistémicas de problemas inicialmente puntuales, y llevó aquella a percibir con
mayor claridad que la sostenibilidad política de los programas económicos depende de la minimización de
sus costos sociales.- Todo ello ha venido contribuyendo para gestar un nuevo consenso (del que participan
las economías del G-7, las agencias de Naciones Unidas, las instituciones de Bretton Woods, y por cierto,
los propios países afectados por dichas crisis), respecto a la necesidad de priorizar la protección de los grupos
más vulnerables en futuros programas de estabilización, ajuste y de reformas estructurales.Propósito
Examinar opciones teóricas alternativas para el diseño de 'redes de protección' de corto plazo para grupos
socio-económicos más vulnerables, analizar los resultados de experiencias recientes en economías de América
Latina y Asia, como también, evaluar el costofiscal correspondiente a tales intervenciones. Además, se
tratará de clarificar en qué sentido y medida las propuestas más recientes para integrar 'redes de seguridad' al
diseño de una nueva arquitecturafinanciera internacional, constituyen una nueva doctrina para el manejo del
gasto público en épocas de crisis (respecto a las recomendaciones de la ortodoxia convencional). Más
específicamente: ¿De qué manera dichas 'redes' se vinculan a lafijación de metasfiscales en Programas de
Estabilización y Ajuste del FMI?, y ¿Cuál es el papel de los bancos de asistencia multilateral en esta
materia?.-

Sesión 5
El nuevo Manual de Estadísticas de Finanzas Gubernamentales del FMI:
Presentación y Discusión
Antecedentes
El Departamento de Estadística del FMI se encuentra revisando la edición de 1986 del Manual que entrega
los conceptos, definiciones, clasificaciones y reglas contables básicas para la generación e interpretación del
sistema de estadísticas gubernamentales de mayor uso internacional.- Para esta nueva edición se vienen
considerando importantes cambios que esta sesión permitirá conocer en primera mano, creando al mismo
tiempo un espacio para que especialistas enfinanzas públicas y política fiscal de AL&C entreguen sus
reacciones , sugerencias y recomendaciones.- Los capítulos ya disponibles del Manual se encuentran en la
página web del Fondo (http://www.imf.org), donde es posible apreciar las principales novedades.- Entre éstas
destaca el propósito de compatibilizar el marco analítico del Manual con el Sistema de Cuentas Nacionales de
Naciones Unidas de 1993.

Como consecuencia, cambia la caracterización del sector gobierno (desde la óptica de funciones a la
de agrupación de unidades), el criterio para registrar transacciones (de caja a devengado), y aumenta el grado
de integración entre flujos y estoques (la noción deriquezaneta y hoja de balance del sector gobierno
adquieren gran importancia).- Estos últimos y otros importantes avances conceptuales propuestos en el
Manual, contrastan con la sensación de que los indicadoresfiscales más utilizados (digamos, sobre déficit y
deuda pública) están afectados de una cierta dosis de opacidad que convendría disipar pues perjudica las
posibilidades de evaluación y análisis de las propias autoridades económicas nacionales y, por cierto, la
comparabilidad internacional de las finanzas públicas y de la políticafiscal.- Por este motivo, la CEPAL y el
Banco Nacional de Desarrollo Económico y Social (BNDES) de Brasil vienen trabajando en una iniciativa
conjunta para mejorar la transparencia y comparabilidad de los indicadoresfiscales de los países de AL&C,
tema este que será objeto de discusión en un Taller ad-hoc a realizarse el día posterior a esta sesión.Propósito
Crear un espacio para que los especialistas enfinanzas públicas y política fiscal de la región, quiera sean de
gobierno, instituciones académicas especializadas u organismos internacionales, puedan tomar conocimiento
detallado y en primera mano, sobre los cambios que vienen siendo propuestos para esta nueva edición del
Manual y para que tengan la oportunidad de expresar sus observaciones, sugerencias y recomendaciones.
Establecer un puente con el Taller sobre Metodologías para Estimación y Cálculo de Indicadores Fiscales
organizado por CEPAL-BNDES

/¿s

Juan Carlos Lerda
. Coordinador
XII Seminario Regional de Política Fiscal

INDICE

DISERTACION INAUGUI^L
'The Role of the State and the Quality of the Public Sector"
Vito Tanzi

SESION 1;

3 tS

¿ QUE DISEÑO DE MARCO REGULATORIO ES MAS ADECUADO
PARA EVITAR EL SOBRE-ENDEUDAMIENTO SUBNACIONAL ?

"Responsabilidad Fiscal a Nivel Subnacional: Argentina y Brasil"
George Kopits, Juan Pablo Jiménez y Alvaro Manoel

23

^

"Sub-national Governments and Credit Markets: Selected Cases "
María E. Freire, Marcela Huertas, Benjamin Darche

59

^

"El Endeudamiento de los Gobiernos Subnacionales: el Caso de las Provincias Argentinas"
Alejandro Zentner y Juan Sanguinetti

SESION 2:

111

INCENTIVOS
TRIBUTARIOS
PARA
ATRAER
INVERSION
EXTRANJERA DIRECTA: EXPERIENCIAS RECIENTES Y EVALUACION
DE DISEÑOS ALTERNATIVOS

"Globalización, IED y Nuevas Estrategias Empresariales en América Latina"
Michael Mortimore
"Tax Incentives for Foreign Investmentin Latin America"
Richard M. Bird and Duanjie Chen
"Gravación Impositiva e Incentivos Fiscales a la Inversión Extranjera Directa la Opinión del
Servicio de Asesoría sobre Inversión Extranjera"
Joel Bergsman

155

169 ^

y
217

SESION 3:

ADMINISTRACION
DE
RIESGOS
Y
MANEJO DE
PASIVOS
CONTINGENTES DEL SECTOR PUBLICO: ¿ COMO INCORPORAR EL
TEMA DE LAS GARANTIAS GUBERNAMENTALES «EXPLICITAS E
IMPLICITAS-- A LA PROGRAMACION PRESUPUESTARIA ANUAL Y
PLURIANUAL?

"Government Contingent Liabilities: a Hidden Risk to Fiscal Stability"
Hana Polackova

231VX

"Managing Fiscal Risk in Bulgaria"
Hana Polackova, Sergei Shatalov and Leila Zlaoui

261 /

"Risk Management of Contingent Liabilities Within a Sovereign Asset-Liability Framework"
Elizabeth Currie and Antonio Velandia

307"

SESION 4:

PAPEL, ALCANCE
Y
PROTECCION SOCIAL

COSTOS

FISCALES DE

LAS

REDES DE

"Social Issues in IMF - Supported Programs"
Sanjeev Gupta, Louis Dicks-Mireaux, Ritha Khemani, Calvin Mcdonald and Marijn Verhoeven
"Economic Shocks, Safety Nets and Fiscal Constraints: Social Protection for the Poor
in Latin America"
Norman Hicks and Quentin Wodon
"Labor Markets and Income Support: What did we Learnfrom the Crises?"
Gustavo Márquez

¡
337 -

í
381
,
409 v

Disertación Inaugural - 1

Disertación Inaugural - 2

"What is the best government? That
which teaches us to govern
ourselves"

-Johann Wolfgang von Goethe

THE ROLE OF THE STATE AND THE QUALITY
OF THE PUBLIC SECTOR *

by
Vito Tanzi
January 5, 2000

* This paper's objective is to contribute to the Iiterature on second generation reforms. An earlier versión
was presented at the Fund "Conference on Second Generation Reforms" (November 8-9, 1999) and at a
conference on "Controlling Public Activities" organized by the Bank of Italy, in Perugia (December 2-3,
1999). Specific comments on the earlier draft received from Era Dable-Norris, Barry Potter and Howell
Zee were much appreciated. I would also like to thank Patricio Castro, Jerome Foumel, Eduardo Ley,
Istvan Székely and various participants in an internal FAD seminar. The views expressed are strictly
personal and not official Fund views.

Disertación Inaugural - 3

Table of Contents
I. The Role of the State
II. The Importance of Rules
A. The role of constitutions
B. Laws
C. Regulations
III. Political and Procedural Rules
IV. The Quálity of Public Institutions
A. Synergy
B. Enforcemeiit mechanisms
V. Measuring the General Quality ofthe Public Sector
VI. Concluding Remarles
References

Disertación Inaugural - 4

I. The Role of the State
Human beings are social animals, as a consequence they tend to aggregate in groups. Aggregation
generates markets and the need for social institutions. Markets and economic activities are characterized
by specialization on the part of the participants. Specialization leads to exchanges, and exchanges involve
contracts. Contracts may be implicit or explicit. Implicit contracts prevail in primitive societies and are
often enforced by existing moral and social codes. As societies develop, they tend to need more explicit or
formal contracts which in turn require institutions to register them, when registration is required, and to
enforce them. Individuáis also accumulate assets and this leads to the need for property protection.
Furthermore, they need personal protection against criminal elements within their groups or protection
against outsiders. Thus any aggregation of individuáis requires the existence of at least rudimentary
institutions that perform these functions and that can be considered a minimal state. This is the essential
role of the state in a primitive society or the minimal role in any society.1
As societies become more complex, and, perhaps, as the groups that constitute them become
larger and less homogeneous, the state must assume additional responsibilities if it wishes to promote fully
the welfare of the individuáis that comprise it. For example, markets are not efficient when they are
distorted by monopolies or when essential information is not available to the participants, or is too costly
for them to obtain. This leads to the need for the state to regúlate markets and to provide at least some
essential information. Thus, institutions must be created to satisfy these needs. See World Bank (1997).
There are some goods (defense, city streets, etc.) that only the state can provide because prívate
individuáis would not find in their interest to provide them. An important reason is that once these goods
are provided, the provider will not be abíe to exelude from their use individuáis who do not wish to
contribute to the cost of providing them. Another reason is that once they are produced they can be
consumed jointly by everyone so that it would be inefficient to exelude some individuáis from their use.
Thus, if these goods are essential, they must be provided (but not necessarily produced) by the state. These
are the public goods of the public economics literatee. Also, the production or consumption of some
goods or activities generate positive or negative extemalities which are not captured by those who
consume or produce them. Thus, if the state does not intervene, the market is likely to underproduce or
overproduce these goods or activities. Once again, there may be a need for the public sector to intervene.2
All the above are examples of a general role of the state that in the public economics literature
goes under the ñame of allocation. See Musgrave (1959). Although some problems of allocation can be
dealt with by the use of regulations, often, if the state must perform this role, it will need resources. In
principie the state could directly appropriate these resources by forcing individuáis to contribute their time
or their wealth for the production of the required goods—as it does, for example, with compulsory military
service or with some forms of expropriation. In medieval times roads were often built in this way.
However, a more efficient alternative is to use resources raised through taxation. Thus, there is a need for
a tax system and for institutions that collect taxes. And, of course, there is a need for institutions that
spend the money and keep track of that spending. In all these activities, in a market economy the state is
There is now a extensive literature that argües that even gangs or criminal associations of individuáis require within
n
themselves s m organization that is a rudimentary f r of govemment. See Skaperdas and Syropoulos (1995), Charap a d
o e
om
n
Harm (fortheoming), a d various papers in Fiorentini a d Zamagni, editors (1999). For the concept of a minimal, or even of
n
n
a ultraminimal, state see Robert Nozick (1974).
n
The traditional state intervention is to tax those who generate negative extemalities and to subsidize those who generate
positive extemalities.

Disertación Inaugural - 5

expected to minimize the cost that its activities impose on the market and on society at large. Thus, a
public sector of high quality requires an efficient tax system and an expenditure system that minimizes
inefficient and unproductive spending. It also requires a budget large enough to allow the state to perform
its role in a satisfactory way. See Tanzi and Zee (1998).
In addition to the essential or fundamental role of the state associated with the allocation of
resources, a role that has been explicitly recognized and described by economists at least since Adam
Smith's The Wealth ofNations, in this century two or, perhaps, three additional economic roles have been
assigned to the state. These are (a) redistribution of income, (b) stabilization of economic activity, and (c)
promotion of growth and employment. These latter tasks are lessfirmly based in economic theory than the
allocation role but they have been important in the practical implementation of policies in the second haft
of the 20a1 century.3
Markets produce goods and services and provide incomes to those who particípate in them.
Depending on the initial positions of the individuáis as to ownership of real assets, of financial resources,
and of human capital and talent in general, and depending on effort, propensity to save, luck and risk
taking, a given pattern of income (and wealth) distribution is determined. This distribution may or may not
match the prevailing societal perception of the degree of inequality that is considered acceptable.
Furthermore, because of physical or mental handicaps, or because of other characteristics or situations,
such as oíd age, unemployment, number of dependents in the family and so on some individuáis may not
be able to generate an income sufficient to sustain themselves or their families. In these cases, in this
century, and in some countries, the state has been expected to provide them with transfers aimed at raising
their consumption over an essential minimum.4
The assignment of a redistributive role for the state has dramatically changed the character of its
intervention in the economy because it has introduced in its actions purely political considerations or
objectives. In fact, while, at least in principie, the state's role in allocation can be derived from technical
and objective analysis, the proper redistributive role cannot be established objectively. For example, there
is no objective way to determine the ideal Gini coefficient or the ideal ratio of the top quintile to the
bottom quintile of the income distribution of a country. Thus, these decisions are unavoidably influenced
by the biases of the individuáis constituting the government in power.
Programs aimed at redistributing income often require a lot of resources and the presence of
institutions charged with administrating these programs. For this reason, they tend to lead to large
increases in public spending and in the level of taxation as they have in the so called welfare states.5 At
times, the redistribution objective has been pursued through progressive taxation. But more often it has
required large bureaucracies and much public spending. In many countries a large proportion of the public
employees is now engaged in the provisión of education and health services and a large share of the total
public spending of industrial countries goes for social programs including pensions. And in some cases
public employment has had the specific objective of redistributing income. See Alesina, et al (1999).
Markets do not operate smoothly but are characterized byfluctuations that at times become
recessions or even depressions. These recessions lead to growing unemployment and a loss of output and,
in the second part of the 20m century, they created another major justification for state intervention
namely, the maintainment of fall or at least high employment and the stabilization of output. The
These roles had not affected m c the behavior of goveniments in m r e economies until this century.
uh
ak t
In previous centuries religious groups played a similar role in helping t e poorest.
h
The World Bank, op.cit., has argued that t e process of taking over these new functions has distracted the state f o its m r
h
rm
oe
f n a e t l role.
u d m na

Disertación Inaugural - 6

promotion of this Keynesian objective is carried out by government officials and requires some relevant
institutions. These officials must have the capacity to change tax revenue and/or public spending in order
to influence aggregate demand in desirable directions and must have the information and the technical
knowledge and wisdom to make the right decisions.
Finally, in recent decades, governments have also promoted policies aimed at raising the rate of
growth or at creating employment. Import substitution and industrial policies have been examples of
attempts at raising the growth rate. Manipulation of the work week and public investments in particular
activities or regions are examples of attempts at raising employment. Regardless of the wisdom of these
policies, particular institutions must be responsible for executing them.
1.

We can summarize the role of the state in a market economy in the following way:
It must establish and enforce formal rules in the economy, including the enforcement of contracts and
the protection of propertyrights,as well as rules governing the extraction and the use of public
revenue;

2.

It must provide a legal and regulatory framework that reduces transaction costs—that is the cost of
dealing with other individuáis in economic matters-and promotes market efficiency—through
strategic market intervention in cases of market failure or through the production of some essential
information;

3.

It must provide public goods and deal with obvious cases of externalities which cannot be resolved
by negotiations between the involved prívate parties;

4.

It must promote macroeconomic stabilization; and

5.

It must promote a distribution of income that is consistent with the prevailing view of society.

For all of these functions the state needs some rules and institutions:
The role of the state in the economy has changéd over time and varíes across countries because it
is influenced by the changing views of society and by changing technologies. Therefore, the quality of the
public sector should be assessed in light of the objectives and priorities of the government at any given
point in time. These objectives may vary depending on the current needs of a particular country. At times
these objectives may even be in conflict with one another. The state plays its role through the set of rules,
laws, and institutions that make up the public sector. The higher is the quality of the public sector, the
easier it will be for the state to promote its role.
The quality of the public sector is defined here as the characteristic that allows the state to pursue
its objectives in the most efficient way. Although it is difficult to separate the two, the quality of the public
sector is not necessarily the same thing as the quality of economic policy. Thus, the quality of the public
sector is distinguished from the quality of the policies that a given government may be pursuing at a given
moment. A high quality public sector is simply the instrument that facilitates the formulation and the
implementation of the government's policies. A good public sector makes it easier to pursue good policies.
But, even a high quality public sector cannot guarantee consistently good economic policies because it
cannot prevent policymakers from occasionally pursuing poor policies. However, one would have some
hesitation to cali a public sector of high quality if poor policies were frequently pursued and one would
expect tofind a high relation between the quality of the public sector as defined here and the quality of
economic policy. In other words, a high quality public sector is likely to promote good policies and over
time a low quality public sector is likely to promote poor policies.

Disertación Inaugural - 7

The definition of high quality public sector used in this paper differs from that of "good
government" used by other writers. For example, La Porta et al (1998) define good government as one that
is "good for capitalist development." They do not distinguish between quality of public sector and quality
of public policy. A poor public sector will, of course, render the pursuit of good policy more difficult
because it will not provide the government with the needed information and it will not guarantee that the
policy decisions will not be distorted in the implementation stage. See on this, Tanzi (forthcoming).

II. The Importance of Rules
In the previous section I described the broad categories of state intervention in the economy. Each
of these categories requires specific programs. In democracies with market economies the mandate for
intervention in the economy given to the state is often specified, first, by a constitution and then by many
laws and regulations that give a specific content to the normally general principies enunciated by the
constitution. Thus, the constitution and the laws and regulations establish the rules of the game that guide
the actions of individuáis and enterprises on one hand and public institutions on the other. But, as North
and Weingast have pointed out, "[a] critical political factor is the degree to which the regime or sovereign
is committed to or bound by these rules"6 (1994, p.312).

A. The role of constitutions
The principies expressed by the constitution may be clear but are often not very specific. The
constitution must be a living document that guides action but it cannot be one that addresses specific
situations or that anticipates activities or situations that do not exist at the time the constitution is written.
It has been reported that Napoleon rejected the draft of the constitution that the best legal minds of his
time had prepared at his request because he felt that the document was too specific and, thus, potentially
too restrictive and likely to become obsolete with the passing of time. He preferred a statement of general
principies that could be interpreted with some flexibility over time. To a large extent the American
constitution is a good example of such a statement. Only rarely it has required amendments.
Examples of constitutional principies that are not too clear abound. A well known example is
Article 81 of the Italian Constitution that seems to restrict déficit financing but that, over the years, was
interpreted in ways that did not prevent the occurrence of largefiscal déficits. Examples of too restrictive
constitutional limits are also not uncommon. An important example is the 1988 Brazilian Constitution. In
recent years that document has prevented the national government from making important and badly
needed reforms infiscal federalism and in the pensión system. Other examples are, the Indian Constitution
that has prevented the introduction of a national sales tax, or the Pakistani Constitution that has limited the
scope of any sales tax imposed by the national government to goods only and has prevented the taxation
of agricultural incomes.
A problem with constitutions is that they tend to reflect the preoccupations and the political forces
of the time they were written. A good example of this characteristic is the Italian Constitution. Its first
article sets the tone by stating that "Italy is a democratic Republic founded on labor." It subsequently
states that all citizens have the right to labor. See Article 4. The third chapter (Title El) of the
In authoritarian govemments or in dictatorships t e constitution m y play a marginal role. Stalin's constitution for t e Soviet
h
a
h
Union was a very good one but it did not prevent him f o doing whatever he wanted. He was obviously not bound by t e
rm
h
constitutional limitations.

Disertación Inaugural - 8

Constitution, which specifically deals with "economic relations," lists many rights of workers and only
limitations to property rights. It specifies that the state can put limits to the use of property and can
expropriate it for the national interest. It also states that the state can impose various limits to the use of
property in order to "establish equitable social relations." If the proper economic role of the state in a
market economy requires the protection of propertyrights,this document is surely reluctant to assign that
role to the state. It should, thus, not be surprising if economic policies and institutions in Italy have
developed to conform to this constitution and have allowed policies (rent controls, expropriation of land
with very low compensation, etc.) that are contrary to the protection of property.
In conclusión, the role that constitutions play in determining the quality of the public sector
cannot be exaggerated.7 It is the constitution that allows or disallows certain actions on the part of the
governing body and on the part of the citizens. Thus, at least in principie, the laws and the regulations that
govern a country must be consistent with the constitutional principies as interpreted by a Supreme Court
or by an equivalent body. In recent years, there has been a trend in many countries to try to modernize
their constitutions. However, the results have not been always very good. At times the revised
constitutions have been poorly formulated, complex and confusing.
Some economists, including James Buchanan and Francesco Forte have argued that the economic
orfiscal role of the constitution should be one of establishing limits to governmental action as for example
done by the Swiss Constitution. See Forte (1998). The constitution should stress what the state cannot do
rather than what it should do. Thus, these authors might favor limitations to tax rates, to levels of public
spending or tofiscal déficits. The Maastricht Stability and Growth Pact might be seen as an example of
such a constitution even though it is silent about the size of public spending.
In some circumstances, and again Brazil comes to mind, a change in the constitution that removed
obstacles to desirable and important reforms may be a necessary condition for raising the quality of the
public sector and of public policy.
B. Laws
While the constitution sets, or at least should set, the general principies that guide the country's
policies, the latter are permitted and directed by specific laws. It is possible to generalize that the quality
of the public sector is enhanced when the laws are relatively few, are clearly written~and thus not subject
to conflicting interpretation—are comprehensive, and are not in conflict among themselves. Difficulties
have arisen when the laws are too many, are not clear, do not cover all relevant areas of economic
activities, or provide conflicting signáis.
Among European countries it has been reported that while some of them have only few thousands
active laws others may have tens of thousands. See for example, Ferro et al (1999). It may thus be difficult
tofind one's way in this legal jungle. The legal system may reflect a problem conceptually similar to that
associated with Y2K. It occurs because at each time that a new law is enacted, all the existing laws should
be scrutinized and, if necessary, revised to make sure that all the elements of these laws are consistent
with the new law. When the existing laws are many, and especially when they are not clear, it is almost
impossible to ascertain this. Thus, occasionally elements imbedded in laws enacted in the past, but still on
the books, conflict with the new laws. When this happens the directives to the citizens as well as to the
institutions charged with implementing the programs contemplated by the laws get confused. This
problem could be referred to as one of legal inconsistency. It often characterizes the relations between
This of course assumes t a t e rule of law is well established in t e country so t a t e constitution is t k n seriously.
ht h
h
ht h
ae

Disertación Inaugural - 9

national and subnational governments, or between, say, pensión and health programs on one hand and
annual budget laws on the other.8 Zoning laws and laws dealing with the environment have also suffered
from this problem. At times, a law and, thus, a given institution authorizes a certain land use but another
law and another institution prohibits its. These confliets may lead to costly mistakes.
Thus, to repeat, the quality of the public sector is enhanced when the laws are written clearly and
cover all the necessary areas, when they do not lead either the public or the public officials to conflicting
interpretations, when their number is as small as feasible, and when they are not in conflict among
themselves. Over the years there have been attempts on the part of some countries either to codify the
existing laws or to simplify them. However, these actions, which would give benefits only over the long
run, do not get much political support. Thus, the attempts are often abandoned before they produce the
necessary results. See, for examples, Guy (1996), Braibant (1996), Mattarella (1994), Smith and
Richardson (1999) and Tan and Tower (1992).
C. Regulations
Laws are often accompanied by specific regulations.9 These can be classified in three groups:
economic regulations, safety regulations, and information regulations. Regulations explain procedures or
elabórate on the content of the laws or simply impose rules on individuáis and enterprises. In some cases
the laws are so complex as requiring a very large regulatory body. For example, the regulations covering
the U.S. Internal Revenue Code cover 35,000 pages. The Internal Revenue Code itself is some 1500 pages
long. Regulations issued by an executive authority or by a regulatory agency can themselves be very
complex, may not be easily accessible by the public, and may overlap with other regulations. In some
countries they may not even be published.10
Because many forms of regulations do not require budgetary appropriations or formal approval by
the legislature, they tend to be less scrutinized. The interpretation of regulations is often left to the
bureaucrats who administer them. Therefore, the possibility of confusion or even abuse is very high.
Regulations have been identified by the relevant literatee as one of the major causes of corruption
because the bureaucrats in charge may abuse them to derive personal advantage. See Tanzi (1998b). They
have also been shown to impose very large welfare costs on the economy.11
Because of the dynamic character of economies, and because of fast technological change, it is
common tofind countries with too many anachronistic, useless, or even damaging regulations and too few
necessary regulations related to new economic activities.12 This is an area where excess and scarcity often
coexist. Many countries are now struggling to create needed regulations for the financial and banking
sectors, for the use of the information superhighway (internet), for genetic research, for the use of some
drugs, and for other similar, important, new areas.

9
10
11
12

For example, the budget law m y assign a given budgetary allocation for t e health sector but t e laws t a determine the
a
h
h
ht
performance of the health sector m y imply a m c higher spending level. See Reviglio (1999) for a specific example f o
a
uh
rm
Italy.
Regulations m y be legislated and t u be laws themselves or t e m y be issued by public agencies.
a
hs
hy a
In a country in which I worked m n years ago tax incentives regulations were not publicly available in a y form. Thus, t e
ay
n
h
bureaucrats making the decisions had all the knowledge a d their decisions could not be challenged by t e taxpayers.
n
h
See 1 8 Report of t e U.S. Grace Commission. See also Dixit (1996) and Laffont and Tiróle (1993).
94
h
A couple years ago, it was reponed in the American press t a in California, bread had to be sold only in pieces of a exactly
ht
n
well specified weight. This regulation, which was still on t e book, had been introduced at t e beginning of t e century when
h
h
h
m s buyers did not have scales so that t e could be easily cheated by unscrupulous bakers. In Italy and Germany discounts
ot
hy
o sales by shops m s be approved by the municipality a d m s be limited in time.
n
ut
n ut

Disertación Inaugural - 10

In conclusión, a high quality public sector must have enough clear rules to guide economic (and
other) activities but not rules that are so many or so vague as to give excessive power to the bureaucrats or
to create uncertainty among those who make economic decisions. In general, the rules should specify what
is not allowed rather than authorize what is allowed. Discretion by bureaucrats should be kept to a
minimum. It has been reported that, in some countries, particular activities~say, a request for a tax
incentive or an application to open a small enterprise—may need 30 or 40 authorizations signed by public
employees in as many agencies or offices. See, inter alias, De Soto, 1989.
An important step would be to make a periodic inventory, say, every ten years, of all the existing
regulations, so that a regulatory budget could be established and could be pruned of redundant and
anachronistic regulations while complementing it with new required ones and while clarifying the
confusing ones. Such an attempt was made by a few countries including Hungary and Argentina. Such a
process, while costly, would raise the overall quality of the public sector. Another useful step is the
creation of a one-stop, or a "single window," center where individuáis can obtain all the permits and
authorizations they need. Such centers have been created in a few places and are reported to have led to
the elimination of many existing regulations.13
In addition to the formal, constitutional rules and the rules specified in the laws and in the
regulations, the quality of the public sector may be affected by informal norms or arrangements that
influence the economic behavior and the behavior of the public sector. Such norms (a) influence the
choice of presidents, ministers, and other high level officials; (b) characterize appointments in the civil
service; and (c) influence contacts between the state and the prívate sector.14 These informal norms may
be based on religious, social, or political considerations. Being of an informal nature and being based on
cultural characteristics, these norms are difficult to change. Still the application of arm's length
relationships and the rule of law to all aspects of public sector behavior should be a goal in the search for a
higher quality public sector. See Tanzi (1995).

III. Political and Procedural Rules
The constitution, the laws, and the regulations establish the broad legal powers of the public
sector, or, putting it differently, they set the rules of the game that should determine the behavior of the
public sector and the regulation of the market. The importance of these rules cannot be underestimated.
Several authors including Buchanan, Alesina, von Hagen, Poterba, Tabellini, and Persson have argued that
political arrangements such as fiscal federalism and fiscal decentralization, proportional or
nonproportional representation in parliaments, the frequency of elections, the choice of presidential versus
nonpresidential types of govemments, the role of the ministry of finance as a super ministry, the rules that
apply to the budgetary process~for example, whether it starts with a macroeconomic constraints which
reflects a collective view on priorities or whether it allows pressures for spending to be determined
through the political influence of each minister-whether parliament can modify the content of budgetary
proposals, or whether it must vote on the whole budget, whether the central bank is independent and so on
have a significant impact onfiscal and macroeconomic outcomes.15 These situations have been modeled
13
14

15

Examples of these centers now exist in Bologna (Italy) and in Salvador de Bahia (Brazil). In Bologna an attempt is being
made to allow access to this one-stop center through the internet.
In some countries high public positions are almost inherited. Some jobs in particular institutions are passed from one to
another member of the same family. In others, party affiliation facilitates access to government jobs and a change in
government creates a large number of vacancies because many jobs are considered as political.
Ricardo Hausmann has argued in support of a politically independent Fiscal Council that would restrict annual public debt
accumulation to agreed levels. The analogy to a central bank independence is obvious.

Disertación Inaugural - 11

using sophisticated game theories and have been subjected to some empirical testing. See inter alias,
Tabellini and Persson (forthcoming) and Poterba and von Hagan (1998). The guiding assumption is that
economic policy is much influenced by these arrangements.
I will not review this literatura which is still evolving and which, at times, has given conflicting
results. While recognizing its importance and its potential contribution to economic policy, in this paper
the focus is on other aspects and, specifically, on the quality of the public institutions. Political and
procedural rules are more likely to affect policies than the quality of the public sector or of public
institutions. It is the institutions that confront the citizens and that implement the policies. But, of course,
by changing the incentives under which policy makers and institutions operate, the political and
procedural rules may affect the behavior and, thus, the quality of the institutions and vice versa. The rules
are just a set of instructions. They are not yet the policies. Until the game is played, these rules remain just
pieces of papers; and the game is played by the institutions charged with carrying out these instructions.16
They may or may not carry the instructions in a faithful and efficient way.17 The public sector is composed
of many institutions, some more important than the others. It is the performance and the efficiency of
these institutions that, to a large extent determine the quality of the public sector.

IV. The Quality of Public Institutions
The quality of the public sector may be affected by the absence of some essential institutions or by
the poor performance of the existing institutions. For example, in many countries, there are no institutions
responsible for enforcing competition or for forcing full disclosure on the part of financial institutions or
for publishing good accounts on the part of enterprises whose shares are traded in the stock market. As a
consequence the market may function less well because of cronyism and monopoly powers or because of
lack of essential information. The performance of public institutions depends on many factors including
(a) tradition and reputation, (b) the resources they have available and the discretion over their use (c) the
clarity of their mandate, (d) their organizations, (e) the incentives that they face, (f) the quality of their
leadership and staff, and (g) the freedom they have over reorganizational matters.
Take one of the most fundamental institutions as an example, namely the tax administration. Its
performance will depend, in part, on its tradition and its reputation. A tax administration that has been
efficient, honest, and proud in the past is likely to continué to be so in the future unless it faces truly
fundamental shocks. By the same token it is very difficult in the short run to change a corrupt and
inefficient administration.18 Its performance will also depend on the resources that it has available to hire
capable employees and to pay good salaries, to invest in new computer technology, to carry necessary
audits and so on. The clarity of its mandate-for example to enforce fairly and objectively the tax laws-is
important. Its day-to-day independence from political pressures is essential. When the mandate becomes
unclear, either because the laws are not transparent or because the institution is subjected to political
interference that forces it to accommodate the special circumstances of some taxpayers, problems develop.
This, for example, has been the case in some transition economies and in some developing countries.19 In
these countries, political interference has reduced the quality of the tax administration. The organization of
the tax administration is also important and so is the set of incentives that it faces. If an institution is
16

19

It can be argued that there are two sets of games taking place. Thefirst is the one that sets t e rules and the second is t e one
h
h
that implements them.
In a recent paper I have argued t a public policy is often distorted by the prevalence of principal-agents problems. See Tanzi
ht
(forthcoming).
Occasionally countries have closed comipt administrations and have opened new ones, t u starting from scratch.
hs
See for example, Tanzi (1998a).

Disertación Inaugural - 12

poorly organized, or if good or bad performances are equally compensated, the contribution of this
institution to the quality of the public sector will be low.20
There are current attempts at strengthening incentives either by making tax administration
politically independent like central banks, 2 1 or by negotiating explicit contracts between the govemment
and the tax administration that require quantitatively specified performances. In Australia, for example,
the govemment guarantees to the tax administration a given level of resources over a three-year-period. In
exchange the tax administration commits itself to deliver some outcomes and some quantified outputs 2 2
This brings into the picture two other important and related aspects of public institutions namely
(a) the synergy among public institutions and (b) the enforcement mechanisms. These are treated as two
separate aspects although, to a large extent, they are two faces of the same coin.
A. Synergy
Like different elements of an ecological system, public institutions work together and support one
another so that it may not be possible to have, say, a first-class tax administration in an environment where
other institutions, such as the treasury or important ministries or the judiciary, or even the post office, do
not function well.23 Often the same weaknesses affect different institutions. This implies that attempting to
improve just one institution, when the others need equal attention, is not likely to generate the desired
results in the long run. This has been the experience in transition economies where, for example, the
establishment of a good treasury system has not improved much the quality of public expenditure
management when the budgetary process has continued to generate budgets that were so unrealistic that
no treasury couldfinance or manage them. In some of these cases the result has been the accumulation of
arrears on the part of govemment and, as a counterpart, the accumulation of tax arrears by the taxpayers.
See Potter and Diamond (1999).
Inter-institutional extemalities (either positive or negative) are very important and must be
recognized and dealt with in any attempt at improving the quality of the public sector. Unfortunately, to
my knowledge this aspects has not been addressed in the literature. For example, when the judiciary does
not work well, many other institutions suffer. A holistic approach that addresses problems in different
institutions at the same time is likely to be necessary. However, such an approach, which is inevitably
difficult to follow, must be guided by a clear strategy and by the proper sequencing of the changes
required and made. If this approach requires more time to implement than the political horizon of the
govemment that introduces it, it is less likely to be fully implemented. This is the reason why the quality
of the public sector changes only slowly over time.

B. Enforcement mechanisms
The quality of the public sector will depend to a considerable extent on the existence of controls
and enforcement mechanisms. Some of these mechanisms must operate within the institutions themselves.
For example, efficient internal auditors' offices can improve the functioning of the institutions and can
20
21
22
23

Rules that apply to hiring, promotions,firing and similar issues for public servants obviously play an important role in the
determination of the quality of the public sector.
In exchange the tax administration receives more discretion over the use of resources.
The cost of collection ratio may play some role in this context. See, for example, Highfield (1999).
For example, in countries in which tax evaders are not punished by the courts, it will be more difficult for a tax administration
tofight tax evasión.

Disertación Inaugural - 13

provide some guarantees that the latter will not stray away from their basic mandate. But these
mechanisms may not be sufficient. In other cases the enforcement mechanisms must cut across
institutions. This happens when supreme audit institutions specialized in controls and in enforcement have
the responsibility of controlling spending and revenue collection. These institutions often receive their
mandate from the constitution and act as independent bodies.
Examples of these supreme audit institutions are the General Accounting Office (GAO) in the
United States, the Cour des Comptes in France and La Corte dei Conti in Italy, the General Auditor's
Offices in Latin American Countries and so on. Historically, these audit institutions have focused too
much on whether the institutions have complied with the letter of the law rather than with the spirit of it.
Thus, in many cases, the performance of the institutions in terms of outputs and outcomes has attracted
less attention than whether they have complied with legalistic requirements. This kind of auditing is of
limited valué except, perhaps, for insuring accountability because it does not promote the quality of the
public sector in its fundamental objective of delivering services to the public. It does not guarantee that the
public is getting valué for the money spent by the public sector.
In recent years a movement to focus on performance and on output, rather than on formalities and
on input, has started. This movement tries to assess public spending in terms of its economy, efficiency
and effectiveness. It requires quantitative indicators of public performance and of the cost of public sector
activities. It has had its strongest expression in New Zealand and Australia and, in some form, it is
spreading to other countries.24 It has brought with it many changes in contractual arrangements and in
organizations in public institutions. For example, in the countries that have adopted it, tenure on civil
service jobs has gone and many constraints on the actions of those who run departments have been
removed. The goverament, as the principal, now makes contracts with a public agency on what it must
deliver and the head of that agency becomes personally responsible for the outcome. If the agreed goals
are not met, the contracts of the heads of the agencies are not renewed or their salaries are reduced. Thus
strong economic incentives have been linked to performance. The final impact of this change on cultures
different from those prevailing in New Zealand and Australia remains to be assessed.25 For example, in
societies where it is difficult tofire workers, almost regardless of their performance, it would be difficult
to implement these changes. It is easy to see a conflict between this approach and the spirit of the Italian
Constitution.
This discussion of enforcement mechanisms and controls would not be complete if two
fundamental points were not made.26 Thefirst is the weakness of the cash accounting approach that has
characterized the measurement of the operation of the govemment in providing support for good
efficiency controls. The second is the developing field of measuring consumer reactions and preferences
with respect to the services rendered by the public agencies. However, before discussing these two points,
it is important to mention that any concept of quality of govemment services must be related to the cost of
those services and the resources available to generate them. The higher are the resources available, the
better one should expect the public services to be. However, given the level of resources available, the
higher is the quality of the public sector, the higher will be the quality of the services.

24

25
26

While the Australian system initially focused on the outcome from public spending, the New Zealand system has always
concentrated on measurable output. It should be clear that outputs and outcomes are not the same thing. Outcomes may be
more difficult to measure and may be affected by public spending only with long lags. For example, educational spending
may be assessed in terms of students in school (an output) or in terms of growth in human capital (an outcome). Health
spending may be assessed in terms of operations performed or in terms of impact on the length and the quality of life.
For a skeptical assessment of this change see Schick (1998).
I am grateful to Barry Potter for calling my attention to these two points.

Disertación Inaugural - 14

Thefirst question, then, is how to measure the level of resources used. Traditionally, the fiscal
accounts (and thus the budget) have relied on cash accounting or cash transactions. Thus, the cost of, say,
agency X is measured by the cash transfer that agency X receives from the budget. However, cash
transactions have many weaknesses now widely recognized by accountants and economists. The need for
accrual accounting is now well established, at least, in principie, although many practical difficulties with
its use are likely to delay for some time the transition from cash accounting.27 Of course, accrual
accounting measures much better the true costs of public sector activities. It would, thus, allow audits that
are more meaningful in comparing outcomes to true costs. For example, cash accounting ignores the
opportunity cost of using public sector assets if this use does not result in cash transfers. See Tanzi and
Prakash (forthcoming). Thus, an activity that receives little cash transfers but uses very valuable land or
buildings is now assessed to have low costs. This is obviously wrong.
The second question relates to the role of citizens, as taxpayers and consumers of public services,
in evaluating the latter. Most of the reforms to date, particularly those agency-based reforms that pursue
the split between principal (the government) and agent (the public agencies), represent actual or implicit
contracts between two producer interests. A reasonable parallel from the prívate sector might be that
between a regulator in the prívate sector and the monopoly industry that it regulates. The regulator is
trying to represent the consumer in some sense: but his direct role is really concerned with the producers
surplus (avoiding any super-normal profits) rather than consumer interests. What is missing in the
Australia/New Zealand arrangements is the voice of the consumer. Unlike the prívate sector where the
consumer can state his preference in a very obvious way (providing there is some adequate degree of
competition within the market), this role is denied in public services. Some reforms in the 1990s have
been directed at trying to find a surrogate measure for consumer preferences. There are perhaps three
approaches that are worth mentioning.
® First, some evaluation work has deliberately sought to canvas the views of particular
consumers, not so much in terms of an organized consumer survey, but as part of a wider
exercise that tries to understand what, for example, parents make of the education that is given
to their children. However, such exercises have tended to be very partial and confined to
individual sectors.
*

•

27
28

A second approach has been that of consumer surveys on a more organized basis that may
cover a number of different services. The World Bank and individual countries have launched
some initiatives in this area.28 While an advance on the past, the limitations of the approach
are also well understood. First, it is essentially opinion-related rather than measuring some
actions or providing specific responses by the consumer to quality concems. Second, it is
subject to the usual problem that a well crafted survey can produce more positive views of
public sector provisión, than might be justified.
These weaknesses in both evaluation and consumer survey-type approaches led the United
Kingdom in the early 1990s to seek some third surrogate approach. From this emerged the
"Citizens' Charter," which tried to involve the citizens' voice in the setting of standards. This
has been done in a number of ways. Citizens' representatives or other consumer interests are
involved in the setting of standards and performance contracts between government and
agencies, rather than just the ministry of finance. Moreover, having set standards, consumers

The Department of statistics of t e IMF is in t e process of producing a new manual of governmentfinance statistics based on
h
h
accrual accounting.
For example, users of public hospitals have been asked to assess t e quality of t e services. Or taxpayers have been asked
h
h
their opinions of t e tax administrations.
h

Disertación Inaugural - 15

are also given arightand the means to complain through the provisión of telephone hot lines,
etc. The consumer response is taken into account when assessing the overall performance of
an agency manager. The compulsory use of ñame badges for officials so that consumers,
through their response, can both reward good and punish bad service is a small but significant
element. This is still another example of how incentives are slowly being introduced in the
public sector. Most important, in certain cases, consumers are entitled to some form of
compensation, if the public service provider fails to meet the required standard. The simplest
example is that season ticket holders on particular rail services were entitled to refunds, if
delays in the running of trains exceeded a certain mínimum.
None of the approaches is wholly satisfactory. But the injection of a consumer voice in the setting
of standards, in commenting upon the observations or otherwise of standards and, thus, in influencing the
future allocation of public sector resources, is likely to be a major theme over the next few years. It is also
likely that incentives and penalties will play an increasing role in promoting a higher quality public sector.
In addition to the internal controls and those performed by auditing institutions such as the GAO
or the Court des Compts there are some other institutions whose working and efficiency are an essential
ingredient in the quality of the public sector. Among these the system of justice is of the greatest
importance. The role of the justice system in all its manifestations in enforcing contracts, in protecting
property, in ensuring the safety of individuáis, in keeping corruption under control, and in improving the
efficiency of other institutions is fundamental. Thus, it is not an exaggeration to state that the quality of
the public sector of a country and the functioning of its market depend significantly on the quality of its
justice system. It is for this reason that the system of justice is receiving a lot of attention in many
countries including European countries, such as, France, Italy, and Portugal, and the majority of Latin
American countries.
In many countries the justice system has been in crisis in recent years. Individuáis and property
are not protected, contracts are not enforced, processes take years or even decades to be concluded, and so
on. In many cases individuáis who break the law are not punished, are punished lightly, or are not
punished until much later when the deterrent effect of the punishment is lost. In some countries, the slow
moving or even corrupt system of justice has provided an implicit incentive to tax evasión, to corruption
and to other crimes because those caught committing these crimes could count on the inefficiency or the
corruptibility of the justice system to escape punishment. In some countries it takes ten years for the
government to determine whether someone accused of tax evasión ought to pay the unpaid taxes. And,
often, the penalties imposed are insignificant. This is a good example of the cross-institutional externality
mentioned earlier. The same may happen with those who do not live up with the terms of contracts. For
example, the proliferation of bad loans and of financial crises is in part a direct consequence of the low
and much delayed penalties on those who do not live up to theirfinancial contracts.29 In some countries
apartments are left empty rather than being rented out because of the difficulty that the owners would have
in reclaiming them when needed.
When justice is corrupt or inefficient, it becomes unjust because some people are more adroit at
taking advantage of its weak points. The poor and the more law abiding citizens are the ones who end up
paying the higher price and economic activity suffers. Equal access to justice, and to a justice that is
administered in a timely fashion must remain one of the fundamental goals of the state. If this goal is not
met the quality of the public sector will remain poor. See Guigou (1999).

29

In m n countries bankruptcy has become almost a trivial m t e with little consequences.
ay
at r

Disertación Inaugural - 16

V. Measuring the General Quality of the Public Sector
Although knowledgeable individuáis have some a priori notions of the quality of the public sector
of countries, it would be difficult or even impossible to get objective measures of that quality. In principie
one could conduct surveys of perceptions of such quality using the same techniques adopted, for example,
in the surveys of corruption. However, the informational requirements on the part of the respondents to
give acceptable results would be extremely high; thus, the quality of the responses would be low. It might
be easier to evalúate the quality of the major institutions that make up the public sector and somehow
weigh their importance to the general quality of the public sector. However, given the number of such
institutions and the knowledge required to assess them, such an enterprise would also be extraordinarily
costly and not necessarily successful in achieving the desired results. An alternative way would simply
measure the economic and social performance of a country. This alternative would focus on the output or
better the outcome and attribute the results to the quality of the public sector. However, this approach
would also have its limitations.
In recent years some institutions and some scholars have begun to focus on particular features that
capture some significant aspects of the quality of public sectors. The IMF, for example has been focusing
on data that countries could produce and make available to the public. Some of these data relate to the
public sector. The assumption would be that countries that are willing and able to generate good public
sector data, and are willing to make them available to the public, have a higher quality public sector.30 The
Fund has also started to assess the transparencv of fiscal policy andfiscal institutions. This evaluation is
done against a set of principies of fiscal transparency. The assumption is that lack of transparency is an
indication of a lower quality of the public sector and that this lack of transparency promotes inefficiency,
poor policies, and various problems of governance. In time there might be transparency reports for most
countries. These would provide proxies for informal assessments of the quality of the public sector.
In recent years, there has been a lot of attention paid to issues of governance and corruption in
public institutions. These issues also bear on the quality of the public sector. It has been recognized that
lack of transparency in the way institutions operate promotes inefficiency and corruption. There is now a
large literature on transparency and on corruption. Australia and New Zealand, have once again led the
way in promoting techniques aimed at increasing transparency. The new "architecture" of the international
financial system has recommended that countries become more transparent in their policymaking. More
transparency would probably mean less corruption and generally a higher quality and better performing
public sector. However, much needs to be leamed about these relationships. It must also be recognized
that tests of transparency may be superficial or deep. Superficial tests would not be very useful in
understanding what is going on in a public sector and deep test may require a lot of highly specialized
resources.
Corruption is a sign of a public sector's lower quality. Various groups have been generating
corruption indexes for a large number of countries. See Tanzi (1998b). These indexes purport to measure
"perceptions" of corruption. It is likely that to some extent, and assuming that the perceptions reflect
reality, these indexes of corruption can also be taken as proxies for the quality of the public sector.
However, caution is needed because, quite apart from the quality of the corruption indexes, a country
could have a totally honest bureaucracy or even political leadership and very inefficient policies and
institutions. While important, corruption is only one aspect of poor public sector quality.
Some authors, such as Rauch and Evans (forthcoming) have provided estimates of bureaucratic
performance for many developing countries thus providing measures for yet another variable that has a
30

The data themselves might tell more about the quality of public policy than t e quality of the public sector.
h

Disertación Inaugural - 17

bearing on the quality of the public sector. On the other hand, Keefer and Knack (1997), after defining
institutional quality as "objective evaluations...of the institutions that protect property and contractual
rights" attempt a measurement based on various indicators. It is questionable whether what they measure
is the quality of public sector institutions.
There are other variables that may provide information on the quality of the public sector. Some
would relate to its efficiency, some to the policies followed. For example, the relationship between
spending in a given category-say health and education~and the outcome of that spending-such as lives
saved, successful operations, reduction in the incidence of certain diseases, educational achievementswould be an indication of efficiency. See Gupta et al (1997). On the other hand, measurements of fiscal
sustainability might be indications of poor policy rather than of poor quality of the public sector.

VI. Concluding Remarks
The quality of the public sector can be assessed only against the role of the state. If the public
sector allows the state to promote its goals in an efficient and successful way, it can be argued that the
public sector is of high quality. However, the goals must be realistic because even a very efficient public
sector will be unable to cope with unrealistic goals. Thus, in general the quality of the public sector
cannot be measured by the quality of the policy outcome although the two are obviously closely linked.
In general, an efficient public sector should be able to achieve the state's objectives with the
minimum degree of distortion of the market, with the lowest burden of taxation on the taxpayers, with the
smallest number of public employees and with the lowest absorption of economic resources by the public
sector. The public sector must be transparent in its processes and in its outcome. Corruption should play
no role in the decisions made by the bureaucrats and the political leaders. And the resources in the hand of
the public sector should be put to a use that maximizes their social rate of return. Such a public sector
would be clearly "market augmenting."
The equity aspects of the public sector are also important because the pursuit of equity, as
distinguished from efficiency, is now one of the fundamental goals of the public sector. Given its impact
on the use of resources and on growth, a public sector that, through all its aspects, promotes equity must
be deemed to be of higher quality than one that doesn't. However, as argued earlier, it is difficult to
determine the optimal role of the public sector in this area and it is easy to identify disincentive effects that
may be created by some policies that redistribute income. See, for example, Alesina, et al (1999).

Disertación Inaugural - 18

References
Alesina, Alberto, Stephan Danninger and Massimo Rostagno, 1999, "Redistribution Through Public
Employment: The Case of Italy," IMF Working Paper/99/177 (December).
Braibant, G., 1996, "Utilité et Difficultés de la Codifícation," in Droits, No. 24 (December), pp. 61-72.
Charap, Joshua and Christian Harm, forthcoming, Institutionalized Corruption and the Kleptocratic State,"
in Institutions, Contracts and Organizations: Perspectives from New Institutional Economics,
edited by Claude Menard (Edward Elgar).
De Soto, Hernando, 1989, The Other Path (New York: Harper and Row).
Dixit, Avinash K., 1996, The Making of Economic Policy (MIT Press).
Ferro, Pasquale, Stefano Lo Faso and Giancarlo Salvemini, 1999, "L'Azione della Publlica
Amministrazione per la Competitivita Internazionale in Presenza di Vincoli di Finanza Pubblica."
In Concorrenza Fiscale in Un 'Economía Internazionale Inte grata, edited by Massimo Bordignon
and Domenico da Empoli (Franco Angelí).
Forte, Francesco, editor, 1998, Le Rególe della Costituzione Fiscale, Politeia, Anno 14-N. 49/50.
Guigou, Elisabeth, 1999, "Justice:
September 1, p. 12.

Du Ministére des Affaire au Ministére du Droit," Le Monde,

Gupta, Sanjeev, Keiko Honjo, and Marijn Verhoeven, 1997, "The Efficiency of Government Expenditure:
Experiences in Africa," IMF Working Paper 97/153.
Guy, S., 1996, "Une Utopie: La Codifícation," in Revue Franqaise de Droit Constitutionnel, No. 26, pp.
273-310.
Highfield, Richard, 1999, "Tax Administration: Understanding and Using the cost of Collection Ratio,"
(Mimeo, November).
Laffont, Jean-Jacques and Jean Tiróle, 1993, A Theory of Incentives in Procurement and Regulation
(Cambridge: MIT Press).
La Porta, Rafael, Florencio Lopez-de-Silanes, Andrei Shleifer, and Robert Vishny, 1998, "The Quality of
Government" (mimeo, Second Draft, August).
Keefer, Philip and Stephen Knack, 1997, "Why Don't Poor Countries Catch Up? A Cross-National Test
of An Institutional Explanations," Economic Inquiry, Vol. XXXV, July, pp. 590-602.
Mattarella, B.G. 1994, "La Codifícation du Droit: Réflexions sur L' Expérience Franqaise
Contemporaine," in Revue Franqaise de Droit Administrative, Vol. 10, No. 4, July-August, pp.
668-685.
Musgrave, Richard A., 1959, The Theory of Public Finance (New York: McGraw-Hill).

Disertación Inaugural - 19

North, Douglass C. and Barry Weingast, 1994, "Constitutions and Commitment: The Evaluation of
Institutions Governing Public Choice in Seventeenth Century England," in Monetary and Fiscal
Policy, Vol. 1: Credibility, edited by Torsten Persson and Guido Tabellini (MIT Press), pp. 311343.
Nozick, Robert, 1974, Anarchy, State and Utopia (Basic Books).
Poterba, J.M., and J. von Hagen, editors, 1999, Fiscal Institutions and Fiscal Performance (Chicago,
NBER, University of Chicago Press).
Potter, Barry and Jack Diamond, 1999, Guidelines for Public Expenditure Management, (Washington:
International Monetary Fund), pp. 3-50.
Rauch, James E. and Peter B., Evans, forthcoming, "Bureaucratic Structure and Bureaucratic Performance
in Less Developed Countries," forthcoming in Journal of Public Economics.
Reviglio, Franco, 1999, "Budget Transparency for Public Expenditure Control: The case of Italy"
(Mimeo: December).
Schick, Alien, 1998, "Why Most Developing Countries Should Not Try New Zealand's Reforms," The
World Bank Research Observer, Vol. 13, No. 1 (February) pp. 123-31.
Skaperdas, Stergios and Constantinos Syropoulos, 1995, "Gangs as Primitive States" in The Economics of
Organized Crime, edited by Gianluca Fiorentini and Sam Peltzman (Cambridge: Cambridge
University Press), pp. 61-82.
Tabellini, Guido and Thorsten Persson, forthcoming, Fiscal Policy in Representative Democracies (MIT
Press).
Tan, Lin Mei and Greg Tower, 1992, "Readability of Tax Laws: An Empirical Study in New Zealand,"
Australian Tax Forum, 9, No. 3, pp. 355-372.
Tanzi, Vito, 1995, "Corruption: Arm's Length Relationships and Markets" in The Economics of
Organized Crime, edited by G. Fiorentini and Sam Peltzman (Cambridge: Cambridge University
Press).
, 1998a, "Essential Fiscal Institutions in Selected Economies in Transition," Discussion Paper No.
53 (November 1998), Collegium Budapest, Institute for Advanced Study.
, 1998b, "Corruption Around the World," StaffPapers, International Monetary Fund (December).
, Forthcoming, "Rationalizing the Government Budget: Or Why Fiscal Policy is So Difficult" in
Economic Policy Reform: The Second Stage, Edited by Anne O. Krueger (The University of
Chicago Press).
Tanzi, Vito and Tej Prakash, forthcoming, The Cost of Government and the Use of Public Assets (Mimeo).
Tanzi, Vito and Howell Zee, 1997, "Fiscal Policy and Long-Run Growth," IMF StaffPapers, Vol. 44, No.
2 (June).

Disertación Inaugural - 20

United States Congress, Senate Committee on the Budget, 1984, "President's Prívate Sector Survey on
Cost Control (Grace Commission): Hearing Before the Committee on the Budget, United States
Senate, Ninety-Eighth Congress, First session, November 7.
World Bank, 1997, The State in a Changing World, World Development Report, 1997 (Washington, DC:
The World Bank).

Disertación Inaugural - 21

Disertación Inaugural

SESION 1
¿ QUE DISEÑO DE MARCO REGULATORIO
ES MAS ADECUADO PARA EVITAR EL
SOBRE-ENDEUDAMIENTO SUBNACIONAL ?

Sesión 1 - 2 3

Sesión 1

RESPONSABILIDAD FISCAL A NIVEL SUBNACION AL:
ARGENTINA Y BRASIL *

George Kopits, Juan Pablo Jiménez y Alvaro Manoel
Fondo Monetario Internacional

Versión Preliminar

* A presentarse en el XII Seminario Regional de Política Fiscal, CEPAL, Santiago de Chile, 24 al 26 de
enero de 2000. Este trabajo en gran parte ha sido motivado por Teresa Ter-Minassian; los autores quieren
agradecer su apoyo y sugerencias. Las opiniones expresadas son de los autores y no necesariamente
coinciden con las de las autoridades argentinas y brasileñas o con las del Fondo Monetario Internacional.

Sesión 1 - 25

Indice
I.

Modelos de responsabilidad

fiscal

II.

27

Desempeñofiscal en las provincias argentinas

30

III. Desempeñofiscal en los estados brasileños

37

IV. Perspectivas para el futuro

44

V.

Referencias

47

Sesión 1 - 26

La evolución en las finanzas públicas en Argentina y Brasil en las ultimas décadas se ha
caracterizado por dos tendencias importantes. Por un lado, ambos países han sufrido recurrentes déficits
fiscales en un contexto de desequilibrios macroeconómicos. Por el otro, como en muchos otros países con
sistemas federales, tanto en Argentina como en Brasil se han observado presiones centrífugas hacia la
descentralización fiscal1. La descentralización del gasto público no siempre estuvo acompañada por el
consiguiente aumento de recursos a nivel subnacional, con lo cual las tendencias mencionadas se
alimentaron mutuamente.
A principios de los años noventa en la Argentina, y hacia mediados de la década en el Brasil, se han
lanzado iniciativas de ajuste en el sector público. Más recientemente, ante la necesidad de respaldar los
programas de estabilización—en el marco de la Ley de Convertibilidad en Argentina y el Plan Real en
Brasil—y con elfin de eludir el efecto contagioso de futuras crisis cambiarías (tipo Tequila) en mercados
emergentes, ambos países decidieron adoptar reglas macro-fiscales de manera de alcanzar el necesario
equilibriofiscal. La legislación de responsabilidadfiscal o convertibilidadfiscal, que se lanzó este último
año, brinda una oportunidad para promover la disciplinafiscal no solamente en el ámbito nacional, sino
también a nivel subnacional. En efecto, la premisa de base de este estudio es que para que la
responsabilidadfiscal se haga realidad, se la debe ejercer en todas las jurisdicciones donde tenga lugar la
toma de decisiones de políticafiscal, sea cual fuere el grado de descentralización.
Con estas consideraciones en mente, el presente trabajo propone explorar esquemas alternativos que
puedan asegurar una conducta prudente de políticafiscal por gobiernos subnacionales en estos países. En
otras palabras, se trata de encontrar soluciones que sean eficaces y al mismo tiempo respondan a la
realidad institucional de cada país, para resolver la problemática de las tendencias mencionadas. Con tal
propósito, la primera sección de este trabajo consiste en una breve reseña de la experiencia internacional
con instrumentos de responsabilidad fiscal bajo sistemas federales. En las secciones que siguen se
examinan el desempeñofiscal y el proceso de endeudamiento público en las provincias argentinas y los
estados brasileños durante los últimos años. Finalmente, con este telón de fondo, se evalúan diversos
esquemas que emergen en estos países y futuras perspectivas para consolidar la disciplina fiscal a nivel
subnacional.

I.

Modelos de responsabilidad fiscal2

En principio, la responsabilidad fiscal es un concepto amplio que se puede aplicar por medio de
reglas o de una política discrecional3. En la práctica, es un sinónimo de reglas fiscales, es decir,
restricciones permanentes sobre el déficit fiscal, sobre el stock de deuda pública, o sobre algún otro
indicador global de conductafiscal. Por lo general, tales reglas—que se pueden implantar en diferentes
grados de formalidad institucional—están acompañadas por criterios de transparencia y de un marco

1

2
3

La tendencia hacia la autonomía subnacional se ha documentado en Tanzi (1995) y Ter-Minassian (1997a). Las presiones
centrífugas m s violentas se han manifestado en los países integrantes de la antigua Unión Soviética y la antigua Yugoslavia;
á
véase Kopits y Mihajlek (1993).
El término modelo se usa e este trabajo en el sentido genérico institucional; de ninguna manera se le atribuye u significado
n
n
analítico c m de modelo económico teórico o econométrico.
o o
La ley de Honestidad Presupuestaria de 1 9 promulgada en Australia, con fuerte énfasis sobre transparencia por medio de
98
informes frecuentes y confiables de performance e intenciones en lasfinanzas públicas, es un buen ejemplo de responsabilidad
fiscal sin reglas. Pero el término se originó c n la ley de Responsabilidad Fiscal de 1 9 en Nueva Zelanda, que además de
o
94
n r a de transparencia establece reglas de política fiscal.
om s

Sesión 1 - 2 7

presupuestario plurianual consistente con previsiones macroeconómicas4. Es en este sentido que usamos el
concepto en nuestra reseña.
Más allá de esta aclaración, un aspecto fundamental en la implementación de un programa de ajuste
fiscal o de una reglafiscal tiene que ver con el nivel de gobierno que deba llevarlo a cabo. En un sistema
unitario la responsabilidad obviamente se concentra en el gobierno nacional, por lo cual la política fiscal
suele ser más eficaz, especialmente en lo que se refiere a la estabilización, que en un sistema
descentralizado. En cambio, en un sistema federal la responsabilidad fiscal es difusa ya que reside a cada
nivel de gobierno. De ahí que se necesite establecer un esquema para que la responsabilidad sea en efecto
asumida por el gobierno nacional (federal, central) y a su vez por cada gobierno subnacional (provincial,
estatal, local, municipal, etc.). En caso contrario puede que ocurra una falta de responsabilidad con
extemalidades sobre otros gobiernos que a la larga se refleja en un alza en el costo de financiamiento y un
déficit mayor para el gobierno general.
Por lo general, en sistemas federales, la tendencia hacia la descentralizaciónfiscal por intermedio de
la devolución de funciones de gasto y de recursos tributarios a entes subnacionales ha sido acompañada
por el establecimiento de límites o la prohibición alfinanciamiento de déficits de gobiernos subnacionales
por parte del banco central y por fuentes extranjeras5. Pero más allá de estos límites se pueden identificar
dos modelos básicos para promover la responsabilidadfiscal a nivel subnacional: el modelo autónomo y el
modelo de coordinación.
En el modelo de autonomía el gobierno subnacional se impone a sí mismo una restricción al
endeudamiento y/o al saldo presupuestario (Cuadro 1). El ejemplo quizá más conocido es el de EE.UU.,
donde ya antes de la Guerra Civil algunos estados habían introducido la llamada regla de oro (que requiere
equilibrio entre los ingresos y gastos corrientes). Actualmente esta regla está en vigencia en casi todos los
estados (con la excepción de Vermont) con diferentes grados de severidad6. A principios de los años
noventa, en Canadá, varias provincias adoptaron diversas reglas que requieren un saldo presupuestario
global equilibrado o limitan el déficit en proporción al nivel de ingresos, mientras otros siguen la regla de
oro7. En Suiza, la mayoría de los cantones ha seguido de facto o de jure diferentes versiones de la regla de
oro8.
Además de su carácter completamente voluntario, estos esquemas tienen varios aspectos en común.
Primero, las reglas fueron introducidas por gobiernos subnacionales que desean ganar credibilidad ante los
mercadosfinancieros; de hecho, en ninguno de estos países existe una regla macro-fiscal a nivel federal9.
Segundo, las relaciones intergubernamentales (es decir, la distribución subnacional de recursos,
transferencias, responsabilidad por gastos) generalmente están bien definidas y son estables. Tercero, las
sanciones por no cumplimiento de estas reglas generalmente consisten en pérdida de prestigio político por
parte de los gobernantes y su partido político y, más concretamente, por un aumento en el costo de interés
del gobierno en el mercadofinanciero; con menor frecuencia, en algunas provincias canadienses, la falta
de cumplimiento involucra la dimisión o la imposición de una multa al funcionario responsable.
Kopits y Symansky (1998) contiene u análisis detallado de reglas macro-fiscales. Véanse las reseñas en Kopits (1999a) desde
n
el punto de vista argentino, y en Tavares, Manoel, Afonso y Nunes (1999) con una óptica brasileña.
5 Véase Ter-Minassian y Craig (1997).
6 Por ejemplo, mientras que en algunos estados la regla solo se debe observar ex ante, en otros la obligación es también ex post.
También en algunos estados se han creado fondos de contingencia o se permite el traspaso de gastos no utilizados de u año a
n
otro. Suits y Fisher (1985) describen abusos cometidos bajo esta regla e los estados de Michigan y Nueva York.
n
7 Véase Millar (1997).
8 Por ejemplo, Basilea, Fribvcgo y St. Gallen están obligados a cumplir ex ante y ex post con la regla, bajo una definición estricta
de lo que constituye gasto de capital.
9 En el caso de Suiza se acaba de votar e u referéndum una regla de equilibrio presupuestario a nivel federal, que todavía no ha
n n
entrado en vigencia.
4

Sesión 1 - 28

Cuadro 1
Reglas fiscales a nivel subnacional
Fecha de
introducción

Restricción

Estatuto

Sanción por no
cumplimiento

Modelo de autonomía
Canadá
Estados Unidos
Suiza

varios
varios
varios

equilibrio corriente
equilibrio corriente
equilibrio corriente

ley
constitución, ley
ley

judicial
judicial
judicial

Modelo de coordinación
Alemania
Australia

1949
1927 V

equilibrio corriente
límite al
endeudamiento
límite al
endeudamiento
límite ai crédito
de banco central
equilibrio global

constitución
acuerdo
informal
ley

judicial
reputacional

tratado
internacional

financiero

tratado
internacional

financiero

Colombia

1997

CFA (zona del franco)

1973

UE (bajo UEM)

1997

judicial

V Abandonado en 1992.

El modelo de coordinación entre los gobiernos subnacionales generalmente cae bajo la tutela del
gobierno nacional, o supranacional en el caso de una unión cambiaría entre países. En Australia, desde los
años veinte, guiados por el gobierno central, los estados estuvieron sujetos a un acuerdo informal
("gentlemen's agreement") sobre la asignación de límites definanciamiento del déficit subnacional. Con
el tiempo este acuerdo se convirtió en un mecanismo más formal bajo la autoridad del Consejo de
Préstamos, antes de ser abandonado por su ineficacia dada la evolución de varios canales de evasión por
parte de los estados10. El caso de Alemania, donde en la constitución de posguerra se adoptó la regla de
oro a nivel federal y en los Lánder para asegurar la estabilización económica, pertenece al modelo de
coordinación—a pesar de que a primera vista se asemeja al modelo autónomo11. Mas recientemente, en
Colombia se estableció un límite al endeudamiento de gobiernos territoriales. Las entidades territoriales
no pueden contraer nuevos créditos si su deuda supera 80% del ingreso corriente y sus pagos de interés
superan 40% del superávit corriente12. Además en muchos países, como se ha mencionado, se le prohibe a
los gobiernos subnacionales, al igual que al gobierno general,financiar su déficit con créditos del banco
central.

Véase la descripción del sistema en Scott (1983) y una evaluación crítica de su funcionamiento en Craig (1997).
S a n y FOttinger (1997) demuestran la falta d transparencia en la contabilidad y en elfinanciamiento de déficits en los
ph
e
Lánder, no obstante la utilidad de la regla como instrumento de coordinación.
12 Según la ley 358 de 1 9 , el incumplimiento de estas restricciones implica suspención de nuevo endeudamiento, salvo que la
97
entidad territorial prescriba a u plan de aporte c n el Ministerio de Hacienda; la Superintendencia de Bancos podrá imponer
n
o
m l a financieras a entidades que otorguen créditos a estos gobiernos. Pero estos incentivos y recursos no h n sido suficientes
ut s
a
para que todos los gobiernos subnacionales c m l n con la regla.
u pa
10

11

Sesión 1 - 29

Otro ejemplo que también tiene características de coordinación es el de la Unión Europea, si se
considera cada estado miembro como un gobierno subnacional que funciona dentro de una federación o
confederación. En este caso, los valores de referencia del déficit global (límite de 3% del PBI) y la deuda
pública (límite de 60% del PBI) se imponen en virtud del Tratado de Maastricht que estableció la Unión
Económica y Monetaria (UEM). También bajo el tratado, los estados miembros del CFA zona franca están
sujetos a un límite de endeudamiento con el banco central equivalente a 20% de sus ingresos tributarios—
una restricción que incentiva la evasión.
A diferencia del modelo autónomo, el modelo de coordinación debe abarcar a todos los gobiernos
subnacionales—no solamente aquellos que asumen la obligación unilateralmente. Adicionalmente, estas
reglas uniformes se imponen desde arriba, habitualmente según un acuerdo entre los entes subnacionales y
bajo la tutela nacional (o supranacional en el caso de la Unión Europea). Por consiguiente, existe una
autoridad nacional (o supranacional) encargada de velar por el cumplimiento de la regla y autorizada a
imponer sanciones (inclusive multasfinancieras) por no cumplimiento. Además, la experiencia comprueba
que la credibilidad del modelo se apoyafielmente en la transparencia de las relacionesfinancieras entre la
autoridad nacional y las autoridades subnacionales, y en los criterios de contabilidad. La falta de
transparencia peijudica la eficacia de la regla. Aun así, en general, las reglas bajo el modelo de
coordinación han sido menos eficaces, con excepción de la UEM que entró en vigencia hace poco.
Otra diferencia clave es que mientras en el modelo de coordinación la razón de ser de la regla fiscal
es en parte la de reducir el riesgo moral que puede surgir por la existencia de una garantía de rescate
("bailout clause") implícita o explícita, en el modelo de autonomía los gobiernos subnacionales no tienen
acceso al rescate financiero por parte del gobierno nacional en caso de insolvencia o incapacidad de pago.
En otras palabras, en el modelo de coordinación, la regla sirve para evitar la externalidad negativa de la
falta de disciplinafiscal de un gobierno subnacional altamente endeudado sobre el costo de financiamiento
de otros gobiernos del mismo nivel pero menos endeudado dentro del mismo espacio económico13.
Por consiguiente, desde el punto de vista de un país altamente endeudado con un sesgo al déficit
presupuestario, se puede argüir a favor del modelo de coordinación. Volviendo a la premisa de base, en
estas circunstancias, es difícil implementar un ajustefiscal o mantener la prudenciafiscal a nivel nacional
sin restricciones que se apliquen de manera uniforme y transparente por todos los gobiernos
subnacionales.

II.

Desempeño fiscal en las provincias argentinas

Los noventa han sido una década de significativos cambios en la economía argentina. Numerosas
reformas estructurales sumadas a un contexto internacional favorable permitieron lograr un crecimiento
sostenido, sin precedentes en las últimas décadas. Aun así, la situación fiscal no pudo dejar de lado la
tendencia que la caracterizó en los últimos años; solo obtuvo superávit en el año 1994. En lo que se refiere
a las provincias, la evolución de la situación fiscal ha sido extremadamente volátil en los últimos años
(Cuadro 2). El déficit consolidado de las provincias subió desde un 0,7% del PBI en 1993 a 1,32% del PBI
en 1995, para luego acercarse al equilibrio y volver a subir a 0,6% del PBI en 1998. Así pues, el año 1995
divide este período en dos subperiodos. Esta evolución de la situación fiscal provincial es producto de
diferentes cambios que se produjeron en lasfinanzas provinciales durante estos años, entre los cuales
Véanse los argumentos basados en la presencia o ausencia de cláusulas de rescate con referencia a la Unión Europea, e
n
Eichengreen y von Haguen (1995) y McKinnon (1996).

13

Sesión 1 - 30

sobresalen importantes modificaciones en la relación entre el gobierno central y las provincias. Por
supuesto que este resultado global observa comportamientos diferenciados cuando se observa provincia
por provincia14.
El favorable contexto macroeconómico determinó un importante crecimiento de los recursos
tributarios, ya sean de origen nacional (transferidos a las provincias vía el régimen de coparticipación de
impuestos) como de origen provincial. Los recursos coparticipables observaron un fuerte aumento en esos
años, no sólo por el importante aumento en el nivel de actividad sino también por cambios en la estructura
impositiva15. Este fuerte crecimiento de los recursos a transferir resultó clave, permitiendo al gobierno
central reformular los montos de transferencias hacia las provincias. Los fondos detraídos al sistema de
coparticipación fueron canalizados fundamentalmente para financiar el creciente déficit del sistema de
seguridad social 16 . Por el lado del gasto, a fines del año 1991 se dispuso la transferencia de los servicios
educacionales17. Este proceso de descentralización (fortalecido por presiones políticas), que se realizó sin
la correspondiente transferencia de recursos18, explica en parte el crecimiento de los gastos de personal
observados en las jurisdicciones provinciales.
La estabilidad de precios y el fuerte auge en actividad, en el periodo 1992-94, permitió al gobierno
central reformular de manera significativa la relación financiera entre la administración central y las
provincias y avanzar en el proceso de descentralización. Pero a principios de 1995 la crisis financiera
impacta en forma abrupta sobre las finanzas provinciales. La crisis repercute en las cuentas provinciales
de dos maneras: por un lado, la pérdida de depósitos en los bancos provinciales determinó la necesidad de
La particular configuración del ordenamientofiscal en Argentina, con una fuerte asimetría entre las responsabilidades de gasto
y de recaudación tributaria, hace que los cambios en los recursos de origen nacional determinen en forma significativa la
situaciónfiscal de las provincias. Para m s detalle acerca de las características del sistema federal argentino, inclusive u
a
n
análisis e forma desagregada, véase Banco Mundial (1996).
n
15 La nueva estructura se caracterizó por u fuerte a m n o de las tasas del IVA y ganancias, en ese m m no totalmente
n
u et
o et
coparticipables, y una considerable pérdida de importancia de los recursos de comercio exterior, que no se transfieren a los
niveles subnacionales. Adicionalmente las provincias recibieron recursos extraordinarios provenientes de la cancelación de
obligaciones entre la Nación y las provincias (Ley 2 1 3 y 24154) y el pago de deudas del gobierno federal por regalías
43
hidrocarburíferas.
16 Las transferencias a las provincias sufrieron sucesivas modificaciones desde 1 9 . Los cambios afectaron, principalmente el
91
sistema de coparticipación regulado por la ley 2 5 8 (BO: 26/1/88), configurando u e q e a crecientemente complejo. Entre
34
n sum
las modificaciones realizadas sobresalen aquellas querecayeronsobre el IVA y el impuesto a las ganancias (afectando una parte
de lo producido a diversas asignaciones específicas), y los dos "pactosfiscales" que determinaron, e esencia, la afectación del
n
1 % de la recaudación coparticipada alfinanciamiento del Sistema Nacional de Previsión Social.Estas modificaciones, a e á
5
dms
de cambiar la distribución efectiva de los recursos entre el gobierno central y las provincias, generaron distintas transferencias
con asignación específica - hacia lafinalidad Cultura y Educación en el caso de activos, hacia el sistema de seguridad social, el
Fondo del Conurbano Bonaerense, las transferencias según el índice de Necesidades Básicas Insatisfechas (NBI) y el Fondo de
Aportes del Tesoro Nacional en el caso del impuesto a las ganancias, hacia el sistema de seguridad social nacional y provincial
e el caso del IVA- las cuales hicieron m c o m s complejas las relaciones entre el gobierno central y los gobiernos
n
uh á
provinciales. Para u análisis pormenorizado del tema, véase Cetrángolo y Jm n z (1995).
n
i ee
17 Se descentralizaron aquellos que aun quedaban e manos del Ministerio de Cultura y Educación de la Nación
n
(fundamentalmente secundarios) y el Consejo Nacional de Educación Técnica (CONET), algunos hospitales y programas
sociales a los gobiernos provinciales y Municipalidad de la Ciudad de Buenos Aires Vale recordar que el proceso de
descentralización de servicios sociales se había iniciado con la transferencia en 1978, cuando se transfirieron varios de los
hospitales que a dependían de la Nación según ley 2 8 3 (BO: 2 1 / 8. En cuanto a servicios educacionales, léase Ley
un
18
/ 17 )
24049 (BO: 111192). Para más detalle sobre el proceso de descentralización véase Bisang y Cetrángolo (1997) y Carciofi,
Cetrángolo y Larrafiaga (1996).
18 En el mensaje del Poder Ejecutivo Nacional que acompaño el proyecto de la Ley 2 0 9 se justifica el proceso de
44
descentralización afirmando que el ms o "ha tenido principio de ejecución a partir de la mayor dotación de recursos
im
financieros que se originan como efecto de la ley 23548, por la que se dio u nuevo régimen de coparticipación federal y que
n
significó u sacrificio en la participación del Gobierno Nacional". En realidad, la f r a en que sefinanció esta transferencia
n
om
consistió e la retención, por parte de la Secretaria de Hacienda de la Nación, de la participación de cada provincia e la
n
n
distribución secundaria de los montos coparticipados de una s m equivalente al costo de esos servicios. A medida que las
u a
provincias se hicieron cargo de los servicios fueron recibiendo los m n o allí establecidos.
ots
14

Sesión 1 - 3 1

un importante flujo de aportes desde los tesoros provinciales; por el otro lado, la caída en el nivel de
actividad significó un menor ingreso por impuestos provinciales y por transferencias de origen nacional,
vía el régimen de coparticipación. Ante esta situación, el gobierno central asumió un papel activo y utilizó
las diversas herramientas con las que contaba para el rescatefinanciero de las provincias más golpeadas
por la crisis19, ya sea utilizando los Aportes del Tesoro Nacional, los anticipos de coparticipación, o los
préstamos en títulos públicos.
Cuadro 2
Argentina: Ejecución presupuestaria consolidada de las provincias,1993-98
(en porcentaje del PBI)
1993

1994

1995

1996

1997

1998

Ingresos
Tributarios
De origen provincial
De origen nacional
Ingresos corrientes
Ingresos de capital

10,8

10,6

10,4

10,7

9,3
3,7
5,6
1,2
0,3

9,1
3,8
5,4
1,3
0,2

8,8
3,6
5,2
1,3
0,3

8,9
3,6
5,3
1,4
0,4

11,1
9,3
3,7
5,6
1,2
0,6

11,4
9,5
3,9
5,6
1,2
0,4

Gastos
Gastos corrientes
Gastos de consumo
Pagos de interés
Transferencias corrientes
Gastos de capital

11,5
9,9
7,2
0,2
2,6
1,6

11,5
9,7
7,1
0,2
2,4
1,8

11,6
9,8
7,1
0,3
2,4
1,8

11,1
9,4
6,6
0,4
2,5
1,7

11,2
9,4
6,6
0,4
2,5
1,8

11,8
10,0
6,9
0,4
2,7
1,8

Saldo primario (-déficit)

-0,5

-0,7

-0,9

-0,1

0,3

-0,3

Saldo global (-déficit)

-0,7

-0,9

-1,2

-0,4

-0,1

-0,7

Fuente: Dirección Nacional de Coordinación Fiscal con las Provincias

En el plano fiscal, el gobierno central tomó diversas medidas de emergencia para enfrentar la
Estas medidas no fueron compartidas con las provincias, lo que permitió al gobierno central
mejorar su posición fiscal relativa. El déficit fiscal en las provincias alcanzó su pico en la serie bajo
análisis, y a diferencia de crisis anteriores, las condiciones de acceso al crédito bancario se tornaron
particularmente difíciles para las provincias. Los gobiernos provinciales tuvieron que endeudarse a
elevadas tasas y comprometer porcentajes crecientes de su coparticipación como garantía de los
préstamos. La pérdida de depósitos de los bancos provinciales, producto de la crisisfinanciera, agravó la
crisis fiscal de las provincias. Los gobiernos provincialesfinanciaron la banca oficial, en un primer
momento para evitar su caída y, luego con el objetivo de capitalizarla para su privatización21.
crisis20.

Las provincias m s afectadas fueron Río Negro, Mendoza y Córdoba; véase Banco Mundial (1996) y Damill (1997)
á
Entre ellas una sobretasa al IVA del 3% y u a m n o de los aranceles sobre el comercio exterior
n u et
21 Producto de la crisis, durante ese periodo las provincias avanzaron a u ritmomayor e el proceso de privatizaciones E los
n
n
n
últimos años se produjo la privatización de catorce bancos provinciales (además uno licitado y cuatro e proceso de
n
privatización), 1 empresas de energía eléctrica (adicionalmente 2 e proceso de privatización) y 7 empresas de agua potable
4
n
(además 2 en proceso de privatización y 4 licitadas). Como resultado de este proceso puede observarse que salvo los bancos de
19

20

(continued...)

Sesión 1 - 3 2

El proceso de ajuste realizado por las provincias se reflejó en una notable reducción del déficit en
1996. En ese sentido, las partidas que observaron un mayor ajuste fueron las transferencias a las empresas
públicas provinciales (producto de las privatizaciones) y a los organismos de seguridad social22. Como
efecto de estas medidas, el déficit provincial consolidado se redujo significativamente hasta 1997. Pero ya
nuevamente en 1998 las provincias experimentaron un deterioro de la situación fiscal, producto de la
acción combinada de un empeoramiento del contexto macroeconómico y de decisiones de gasto tomadas
por las provincias. En efecto, el empeoramiento en la situación fiscal de las provincias se explica por
causas de tipo endógeno y exógeno. Por el lado endógeno pueden nombrarse: la caducidad de las normas
de emergencia que se aplicaron en 1995 y 1996 para superar la crisis; la implementación de políticas
específicas, sobretodo en educación; y cierta laxitud en lo referente al control del gasto debido a la mejora
en el acceso al mercado financiero. Mas tarde, como causa exógena puede señalarse la crisis financiera
internacional a mediados de 1998 que produjo una caída en el nivel de actividad y consecuentemente en la
recaudación tributaria.
En lo referente al endeudamiento provincial, en 1992 la reforma de la Carta Orgánica del Banco
Central introduce dos modificaciones que tuvieron un fuerte impacto en la evolución de las deudas
provinciales. Primero, el Banco Central no puede tomar ningún nuevo activo interno, lo que significa que
las provincias no podían contar mas con el Banco Central para descontar préstamos de los bancos
provinciales a los gobiernos nacionales. Segundo, se prohibe al Banco Central garantizar depósitos
bancarios, con lo cual los bancos provinciales tienen que apoyarse en la confianza de los depositantes para
mantener liquidez. Ambas medidas endurecen la restricciónfinanciera de las provincias, limitando su
habilidad para pedir prestado a los bancos oficiales.
A su vez, cambios en la situación de los mercados de crédito permitieron que las provincias tengan
un acceso mas fluido a fuentes financieras internas y externas. Durante los inicios de la década este
proceso se desenvolvió sin mayores inconvenientes. Pero como se observó antes, la crisis de 1995 tuvo un
efecto significativo en lasfinanzas provinciales. La crisis significó una fuerte fuga de depósitos y el
agravamiento de la situación de los bancos provinciales, lo cual se reflejó también en un notable aumento
del endeudamiento provincial. La subida de las tasas de interés sumado al difícil acceso a los mercados
financieros significó una pesada cargafiscal, motivando a la mayoría de las provincias a avanzar en un
masivo proceso de privatización de bancos y empresas estatales23. Con el acceso al Banco Central cerrado
por la reforma de la Carta Orgánica, las provincias tuvieron que utilizar sus propios recursos para
recapitalizar los bancos, preparándolos para su privatización.
Como resultado de la crisis de 1995, el stock de la deuda del conjunto de provincias se incrementó
significativamente, en 18% comparado al año anterior. Luego, durante el periodo 1996-99, el nivel de
endeudamiento muestra pequeñas variaciones, aunque, analizado en forma individual, se observa que el
comportamiento de las provincias no ha sido uniforme24. Algunas provincias se encuentran sumamente
comprometidas cuando se compara su endeudamiento con sus ingresos (Cuadro 3). Se destaca en este
grupo la delicada situación de Corrientes, Formosa y Río Negro con un stock de deuda que supera los
las provincias de Buenos Aires, Córdoba, La Pampa, Neuquen, Tierra del Fuego y el Gobierno de la Ciudad de Buenos Aires,
los restantes bancos provinciales h n sido privatizados o están en proceso de privatización.
a
2 2 El Pacto Federal (Ley 24130) incluye u a cláusula por la cual el gobierno central acepta el traspaso de los sistemas
n
previsionales provinciales al Sistema Integrado de Jubilaciones y Pensiones. Afines de 1 9 se habían transferido las cajas
99
previsionales de Catamarca, Santiago del Estero, Ciudad de Buenos Aires, Salta, San Juan, Mendoza, La Rioja, Río Negro,
Jujuy, T c m n y San Luis. Este proceso significó para las provincias u importante ahorrofiscal, ya que los déficit de estas
uu á
n
cajas fueron absorbidos por el gobierno central.
2 3 En ese año se creó el Fondo Fiduciario para el Desarrollo Provincial (Decreto 266/95), que tuvo como objetivo ayudar a las
provincias en este proceso.
2 4 Para u análisis detallado de la evolución de la deuda provincial, ya sea en lo que se refiere a su nivel c m su composición,
n
o o
véase Secretaría de Programación Económica y Regional (1999).

Sesión 1 - 33

ingresos provinciales. Mientras tanto, en Buenos Aires, la Pampa, San Luis y Santa Cruz el
endeudamiento no alcanza ni un quinto del nivel anual de ingresos.
Cuadro 3
Argentina: Deuda de las provincias, 1996-98
(en porcentaje de ingreso provincial)
1996
Total
G.C.B.A.
Buenos Aires
Catamarca
Córdoba
Corrientes
Chaco
Chubut
Entre Ríos
Formosa
Jujuy
La Pampa
La Rioja
Mendoza
Misiones
Neuquen
Río Negro
Salta
San Juan
San Luis
Santa Cruz
Santa Fe
Santiago de Estero
Tucumán
Tierra del Fuego

1997

1998

47,9
78,1
36,9
65,4
40,6
95,6
67,6
55,1
43,1
101,7
54,7
18,6
58,7
73,5
67,4
21,2
78,1
55,7
44,2
12,4
9,2
14,1
42,6
78,1
18,1

40,8
77,6
11.8
62,2
39,8
96,7
79,0
63,3
39,6
107,0
62,6
16,5
40,0
86,0
79,1
18,4
101,2
58,6
49,6
11,2
4,1
14,1
39,1
83,1
36,8

44,1
67,8
14,1
69,2
35,8
116,7
92,9
75,2
44,5
111,5
86,0
15,4
49,1
54,8
83,8
28,6
123,1
57,2
53,3
10,6
18,2
20,0
38,9
85,0
32,1

Fuente: Dirección Nacional de Coordinación Fiscal con las
Provincias

La evolución del endeudamiento obedece en gran parte a la reglamentación de los gobiernos
provinciales. Algunas normas se encuentran en la órbita del gobierno central y otras en jurisdicción de las
provincias. Por su lado, dentro del gobierno central la competencia sobre coordinación y control se divide
entre diversas entidades: Banco Central, Ministerio de Economía, y Jefatura de Gabinete.
En este sentido, en el caso argentino se observa una idiosincrática forma de control del gobierno
subnacional, en que se combinan los modelos de autonomía y de coordinación.

Sesión 1 - 3 4

La Constitución Nacional otorga amplias facultades a las provincias para el financiamiento de
fuentes extranjeras25. Dentro de este marco, el acceso de las provincias al uso del crédito se encuentra
regulado por diferentes mecanismos. En la órbita provincial, en forma autónoma, las constituciones
provinciales establecen límites al endeudamiento provincial, la autorización necesaria para hacer uso del
crédito y la manera de utilización26. Estos límites no siempre han sido respetados por las jurisdicciones
provinciales. En lo que se refiere a los créditos en moneda extranjera, estos caen bajo la tutela del
gobierno central27.
Las operaciones de crédito de las provincias o municipios generalmente resultan garantizadas
mediante la cesión de los recursos que perciben diariamente las provincias de acuerdo con el Régimen de
Coparticipación28. Este mecanismo produjo que en los últimos años la afectación de la Coparticipación
Federal haya sido en promedio del 30%, destacándose los casos Corrientes, Formosa, Río Negro y
Tucumán, que superan el 50% (Cuadro 4). Este tipo de afectación de los recursos a coparticipar, que para
las entidades bancarias resulta un acceso privilegiado al colateral, distorsiona la necesaria evaluación
crediticia. Este mecanismo incentiva a los bancos a prestar sin evaluar la situación de solvencia de los
presupuestos provinciales. Por otro lado, al ser la transferencia afectada en el origen distorsiona el normal
proceso presupuestario en cuanto a las decisiones de gasto.
Cuando se evalúa el proceso de endeudamiento de los subniveles provinciales y su relación con el
gobierno central, deben mencionarse los mecanismos definanciamiento de emergencia con los que cuenta
el gobierno central, los cuales impactan en la performance fiscal y en la evolución del endeudamiento29.
La Constitución Nacional autoriza al Congreso otorgar subsidios para cubrir el déficit en las operaciones
TA

ordinarias de las provincias.
Con este propósito, se estableció el Fondo de Aportes del Tesoro Nacional (ATN),financiado por el
1% del total a coparticipar31, para ser distribuido por el Ministerio del Interior32. De acuerdo con una
evaluación de las transferencias hechas por este Fondo en el período 1988-96 en relación a los principios

El articulo 1 5 establece que entre las diversas atribuciones que tienen las provinciasfigura la posibilidad de importar capitales
2
extranjeros. Mas detalladamente, el artículo 1 4 establece: "Las provincias podrán crear regiones para el desarrollo económico
2
y social y establecer órganos con facultades para el cumplimiento de susfines y podrán también celebrar convenios
internacionales en tanto no sean incompatibles con la política exterior de la Nación y no ae t n las facultades delegadas al
f ce
Gobierno Federal o el crédito público de la Nación, con conocimiento del Congreso Nacional."
2 6 En el Anexo 1 se resumen las exigencias incluidas e las Constituciones Provinciales de las 2 jurisdicciones para poder
n
4
acceder al crédito público.
27 Las provincias están sujetas a la "...supervisión, coordinación y aprobación del Ministerio de Economía y Obras y Servicios
Públicos, respecto de todo acto, contrato, gestión o negociación de créditos por los que se originen obligaciones de pago e
n
m n d extranjera" bajo la Resolución 1 7 (BO: 2 / / 3. La autoridad de aplicación e el Ministerio de Economía es la
oea
05
399 )
n
Secretaria de Programación Económica y Regional (Resolución 1 9 (BO: 31/8/98)).
00
28 Estas operaciones están sujetas a la autorización bajo la Resolución 1 7 . El Banco de la Nación Argentina funciona c m
05
o o
agente de recaudación y retención de las transferencias tributarias.
29 Existe numerosa bibliografía sobre la posibilidad de rescate de los gobiernos provinciales y su impacto en el comportamiento
fiscal provincial; para el caso argentino, véase Nicolini, Posadas, Sanguinetti y Tommasi (1999).
3 0 El Artículo 75, inciso 9 establece que corresponde al Congreso "acordar subsidios del Tesoro Nacional a las provincias cuyas
rentas no alcancen, según sus presupuestos a cubrir sus gastos ordinarios" Para m s detalle sobre el funcionamiento de los
á
Aportes del Tesoro Nacional, véase Cetrángolo y Jm n z (1997).
i ée
31 Este Fondo sufrió posteriormente algunas modificaciones: a partir del a o 1992, de acuerdo al Decreto 8 9 9 (BO: 9 6 9 ) se
ñ
7/2
/ / 2,
incorporó al Fondo de ATN el 2% del impuesto a las ganancias; la Ley 2 6 1 (BO: 9 1 9 ) dispuso que a partir de 1 9 , el
42
//6
96
Fondo sufre una detracción anual de 6 millones de pesos; de acuerdo con la Ley 2 6 9 (27/9/96) el impuesto a las ganancias
49
sufre una detracción (previa a su distribución) de 5 0 millones anuales, de estos, 20 millones se destinan al Fondo de ATN.
8
32 De acuerdo con la Ley 2 5 8 (BO: 2 / / 8, el Fondo de ATN "...se destinará a atender situaciones de emergencia y
34
618 )
desequilibriosfinancieros de los gobiernos provinciales y será previsto presupuestariamente en jurisdicción del Ministerio del
Interior, quien será el encargado de su asignación."
25

Sesión 1 - 3 5

expresados en la ley de creación de este Fondo 33 , se ha constatado que las provincias favorecidas con
mayor proporción de los ATN no son las que presentaban problemas fiscales más graves. Los
desequilibrios fiscales parecen explicar poco acerca de las justificaciones utilizadas en la distribución de
los ATN. En ese sentido, este mecanismo definanciamiento de emergencia no parece cumplir con el fin
con que fue creado. Un avance importante ha sido el recorte en los recursos con lo que cuenta este fondo,
por parte de las nuevas autoridades nacionales, poniéndole un límite a este mecanismo de rescate
discrecional.
Cuadro 4
Argentina: Retención de ingresos provinciales
provenientes de la Coparticipación Federal, 1998
(en porcentaje de ingreso por coparticipación)
Afectaciones
Bancarias

Total
Afectaciones

Total

17,3

27,9

Buenos Aires
Catamarca
Córdoba
Corrientes
Chaco
Chubut
Entre Ríos
Formosa
Jujuy
La Pampa
La Rioja
Mendoza
Misiones
Neuquen
Río Negro
Salta
San Juan
San Luis
Santa Cruz
Santa Fe
Santiago del Estero
Tucumán
Tierra del Fuego

0,3
22,4
14,3
47,7
19,6
13,6
18,1
24,0
40,3
1,4
41,6
6,8
21,7
1,4
63,2
22,5
3,4
1,7
1,5
1,8
45,4
58,7
32,2

0,8
52,4
14,8
48,3
20,1
14,2
18,6
24,6
71,2
2,0
77,1
44,9
22,2
2,0
96,5
53,9
33,1
22,4
2,0
2,4
67,3
95,8
32,8

Fuente: Dirección Nacional de Coordinación Fiscal con las Provincias

En Cetrángolo y Jm n z (1997) se encuentran diversas medidas para evaluar la situaciónfiscal de cada provincia. Ya sea
i ée
considerando la proporción del déficit (antes de recibir ATN) cubierto en cada provincia por ATN; ya sea utilizando la relación
porcentual entre el déficit y el gasto total en cada jurisdicción; ya sea considerando la relación entre el déficit y el total de
recursos de la provincia restadas las transferencias del Fondo de ATN; en todos los casos, aquellas provincias favorecidas con
mayor proporción de ATN no se corresponden con las que presentaban problemasfiscales m s graves. Las cinco provincias con
á
peor desempeñofiscal e el período 1988-96 son Chubut, Catamarca, Jujuy, San J a y Formosa. En cambio las m s
n
un
á
favorecidas por los ATN durante ese período fueron La Rioja, Tucumán, Salta y Mendoza. La Rioja recibió, e promedio, e el
n
n
período 1988-96, el 33.5% de estos recursos.

33

Sesión 1 - 3 6

La falta de un mecanismo claro y transparente de distribución de recursos entre el gobierno central
y las provincias ha influido en la evolución fiscal desfavorable de las provincias. Además, durante la
ultima década, se desaprovechó un contexto macroeconómico en general sumamente favorable, que
hubiese facilitado la discusión, diseño e implementación de un nuevo esquema de relacionesfiscales entre
el nivel federal y las provincias. También faltaron medidas para incentivar y transparentar el
funcionamiento de los mecanismos de acceso al crédito por parte de las jurisdicciones provinciales. La
tarea pendiente en este sentido es introducir instrumentos que conduzcan a que los mercados financieros
internalicen de manera adecuada el riesgo crediticio. Por supuesto que estas medidas deben ser
acompañadas por políticas que incentiven un comportamientofiscal responsable por parte de las propias
provincias.

III.

Desempeñofiscal en los estados brasileños

El sistema federal brasileño se ha desarrollado en ciclos claramente definidos, tanto de
centralización como de descentralización. Desde 1987 se experimenta un ciclo claro de descentralización
fiscal. En tanto la apertura política disminuía el poder del gobierno central, la Constitución del98834
acentuaba la tendencia hacia la desconcentración de los ingresos tributarios; ya en 1987 la porción federal
correspondía a poco menos del 60% del total.
La crisis económica y la hiperinflación también pusieron de manifiesto la falta de adecuación del
sistema tributario y abrieron un espacio político para introducir cambios sustanciales en el sistema federal.
Probablemente el hecho de que la carga tributaria haya disminuido de 25% del PBI a principios de los
años setenta a 20% a losfines de los ochenta promovió las propuestas con miras a la descentralización.
Una de las consecuencias de esa descentralización fue un mayor peso de los gobiernos estatales sobre el
déficit y sobre la deuda pública, lo que constituye el principal problema de la relación federativa actual.
Actualmente los estados y municipios recaudan directamente el 32% del total de impuestos que se
cobran en el país y, tras efectuar las distribuciones constitucionales de los impuestos recaudados, pasan a
disponer del 43% de la recaudación tributaria nacional. Esta división entre las esferas de gobierno se ha
mantenido prácticamente estable desde el gran cambio acaecido en el periodo 1989-9335. Se destaca en
este contexto el papel de las transferencias constitucionales de ingresos recaudados por el gobierno
federal. En efecto, aproximadamente el 27% de los ingresos de la Unión se destinan a sufragar gastos de
los estados y municipios36.
La importancia de los gobiernos subnacionales en el gasto público también se pone de manifiesto en
las principales categorías del gasto (Cuadro 5). El componente mas dinámico consiste en salarios
(inclusive pensiones a los jubilados) pagados que fueron creciendo como proporción del PBI hasta
alcanzar el 7,3% en 1998. Se verifica, en cambio, un estancamiento e incluso una reducción relativa de los
gastos de capital.

34La.

Constitución de 1 8 desconcentró significativamente el sistema: transfirió autoridad tributaria a los estados y municipios;
98
les devolvió responsabilidades de gasto, aunque no bien definidas; consolidó la independencia del endeudamiento de los estados
y municipios, quedando sujetos nominalmente a límites impuestos por el Senado; y posibilitó la ampliación del número de
municipios.
3 5 LOS principales beneficiarios de la descentralización h n sido los municipios cuando se consideran los recursos tributarios
a
disponibles para el gasto público. También se registraron cambios e la división horizontal a lo largo deltiempodespués de que
n
fueron transferidos m s recursos a las regiones m s pobres. Véase al respecto Serra y Afonso (1999).
á
á
36Véase Mendes (1999). El análisis de las transferencias puede encontrarse además en Ter-Minassian (1997b).

Sesión 1 - 37

Cuadro 5
Brasil: Ejecución presupuestaria de los estados y municipios, 1994-98
(en porcentaje del PBI)
1994

1995

1996

1997

1998

Ingresos
Tributarios
ICMS (IVA)
Otros
No tributarios
Transferencias

13,6
8,5
6,8
1,7
1,7
3,4

13,6

12,0

12,9

13,2

9,0
7,2
1,8
1,4
3,2

9,3
7,5
1,8
0,5
3,3

9,2
7,4
1,8
0,5
3,3

9,3
7,5
1,8
0,5
3,5

Gastos
Gastos corrientes
Salarios Brutos
Materiales y suministros
Pagos de interés real 1/
Otros gastos corrientes
Gastos de capital
Efectos en cobro y ajustes

14,5
12,9
5,9
2,3
1,4
3,4
1,9
-0,3

15,9
13,5
7,0
2,1
2,1
2,3
1,9
0,6

14,9
13,4
7,4
2,3
1,3
2,4
1,6

15,1
13,4
7,3
2,2
1,5
2,4
1,8

15,0
13,6
7,3
2,3
1,6
2,4
1,5

Saldo primario (-déficit)

0,5

-0,2

-0,6

-0,7

-0,2

Saldo global operacional (-déficit)

-1,0

-2,3

-1,9

-2,2

-1,8

Saldo global nominal (-déficit)

-19,0

-3,5

-2,7

-3,0

-2,0

Fuente: Banco Central de Brasil y Ministerio de Hacienda.
1/ Los pagos de interés real líquidos representan pagos de interés sobre la deuda externa más el
componente real de los pagos de interés sobre la deuda interna.

Como resultado, los gobiernos subnacionales contribuyeron de manera importante al déficit y a las
necesidades definanciamiento del sector público. Se esperaba que la descentralización de la recaudación
tributaria Xsobre todo del ingreso disponible que queda bajo el control de los gobiernos subnacionalesX
ocurrida después de iniciada la década de los años noventa, diera lugar a un alivio de las cuentas públicas
y se comenzara a producir un equilibrio, al menos, en esas esferas de gobierno. Lamentablemente, se ha
constatado que los estados y municipios siguen teniendo un importante impacto negativo. Mientras que el
gobierno centralrindió,en cifras acumuladas, un superávit primario del 4,4% del PBI durante el periodo
1994-98, los estados y municipios produjeron un déficit primario acumulado del 1,2% del PBI (Cuadro 5).
El efecto de los desequilibrios fiscales, asociados a la ausencia de mecanismos efectivos para
controlar el endeudamiento de los estados y municipios y la generación de déficit primarios y de
operaciones en esos niveles de gobierno provocaron, a lo largo de los años, problemas con respecto al
flujo y al monto de la deuda pública de los gobiernos subnacionales. Una consecuencia grave es que esa
situación dio lugar a una serie de rescates y mecanismos de asistencia ad hoc de parte del gobierno central
para los estados y municipios37.
37Rangel

(1999) y Mendes (1999) presentan los análisis m s recientes de estos rescates y mecanismos de asistencia.
á

Sesión 1 - 3 8

El desglose de la deuda pública de los estados y municipios brasileños38 permite constatar en primer
lugar el nivel reducido del componente externo, de menos de 5% del total. Ese resultado obedece a los
efectos de las negociaciones de la deuda externa con los acreedores internacionales de principios de los
años noventa y a la emisión de los bonos Brady y su concentración en el gobierno federal—deudor
soberano de los títulos extemos en función del esquema de renegociación utilizado. La porción mas
importante, que supera el 80% del total, se debe al Tesoro Nacional, además de las obligaciones que
también fue incluida en las renegociaciones39. El resto, de mas del 10%, corresponde a la deuda contraída
con la red bancaria nacional (Cuadro 6).
El mayor aumento de la deuda pública de los estados tuvo lugar en los años setenta y ochenta, con
los conocidos resultados problemáticos de los años noventa. Esta evolución se puede caracterizar por tres
fenómenos: internalización mediante la sustitución de deuda externa frente a bancos internacionales,
denominada en dólares, por deuda interna frente al Tesoro Nacional, en su mayor parte denominada en
moneda nacional; crecimiento acelerado de la deuda en obligaciones, debido al deficiente control ejercido
por el gobierno federal y por el Senado; y federalización debido a las varias asunciones,
reescalonamientos y rescates realizados por el gobierno federal40. En las dos últimas décadas, la Unión
negoció en condiciones favorables deudas de los estados, contraídas originalmente con el sector financiero
público y privado (equivalentes a 34% del PBI), en tanto que las compensaciones y los pagos por
privatización, redujeron significativamente el total41.
En la comparación de gobiernos subnacionales se observa que, en general, los estados más ricos,
pero con una tradición arraigada de endeudamiento público (especialmente en obligaciones emitidas
recientemente) como Sao Paulo, Minas Gerais, Rio de Janeiro, Rio Grande do Sul, presentan niveles de
endeudamiento que no llegan al 30% del PBI. Por otra parte, muchos de los estados más pobres, entre los
cuales de encuentran Acre, Alagoas, Goiás, Maranháo, Mato Grosso, Roraima, alcanzan niveles por
encima del 40% del PBI (Cuadro 6).

38La

deuda pública brasileña, y en particular la de los estados, se desarrolló recién a partir de los años sesenta. La deuda de los
estados a m n ó bastante en los años setenta (deuda externa). Las Constituciones de 1 6 y 1 8 atribuyeron al Senado la
u et
97 98
fijación de límites globales al endeudamiento de los estados, y correspondía al Banco Central establecer n r a para ponerlos
om s
e práctica. No obstante, siempre hubo resquicios en la legislación, lo que hizo que los órganos de control fiscalizaran
n
básicamente la utilización de los recursos (la inversión, por ejemplo) y no el m n o ni la sostenibilidad de la deuda. A partir de
ot
1 7 hubo algunas innovaciones en esos controles vinculándose la capacidad de pago a los ingresos líquidos de los estados
95
(Resoluciones del Senado Federal 6 / 9 5 y 9 / 9 6 Véase a ese respecto Almeida (1996).
21 7
31 7 .
39Es importante destacar que del total de la deuda en obligaciones quefigura e el Cuadro 6 (R$24.600 millones), casi la mitad de
n
esa cantidad correspondía a la deuda del estado de Rio de Janeiro (R$l 1.300 millones), cuyo contrato de renegociación con el
gobierno federal quedó formalizado afin de 1999. Por lo tanto, en los datos que serán proporcionados en el año 2000,
seguramente ese m n o de deuda en obligaciones deberá trasladarse a la columna "Deuda contractual frente al Tesoro
ot
Nacional". Por otro lado, la otra mitad aproximada de la deuda e obligaciones (R$10.400 millones) de Sao Paulo no
n
corresponde a la deuda del estado sino a la del municipio de Sao Paulo y probablemente será renegociada c n el gobierno
o
federal en el programa de reprogramación de la deuda existente.
""Véase Almeida (1996).
41 Véase Mendes (1999).

Sesión 1 - 39

Cuadro 6
Brasil: Deuda consolidada de los estados y municipios, septiembre de 1999
(en porcentaje del PBI estatal)
.
n
Deuda en
Obligaciones 1/

Total
Acre
Alagoas
Amapá
Amazonas
Bahia
Ceará
Distrito Federal
Espirito Santo
Goiás
Maranháo
Mato Grosso
Mato Grosso do Sul
Minas Gerais
Pará
Paraíba
Paraná
Pernambuco
Piauí
Rio de Janeiro
Rio Grande do Norte
Rio Grande do Sul
Rondónia
Roraima
Santa Catarina
Sao Paulo
Sergipe
Tocantins

2,5
12,1
-

3,1
-

10,3
-

0,1
-

3,4
3,0
-

Deuda Contractual Interna
Tesoro
Nacional

_
B a n c a r , a

Otras
2,3/

17,1

2,8

0,5

41,6
28,9
1,9
12,7
16,5
14,8
0,0
9,0
40,7
45,6
38,9
29,0
21,8
8,3
25,4
9,1
12,9
29,7
6,0
12,3
19,5
19,3
16,2
9,2
20,1
20,9
5,0

2,3
1,7
5,7
0,2
3,1
3,0
5,5
2,0
1,9
1,5

4,6
1,7

1,9

0,1
1,0
0,1

1,0
1,0

0,7
1,8
3,0
1,2
11,4
11,5
2,9
0,8
5,4
27,3
1,2
1,4
2,4
6,7

-

0,0
0,2

0,0
0,0
1,1
3,8
0,6
1,2

0,9
0,9
0,4
0,1

0,1

0,4
0,8
4,9

1,0
1,3

0,0
0,5

0,1

,
_
,
Deuda
Deuda
Externa 3/ Total

1,1

24,0

.

48,5
44,9
8,1
12,9
21,7
22,2
6,3
13,6
46,7
47,8
42,1
31,3
25,1
10,4
28,4
15,3
18,6
42,1
29,3
15,9
21,9
29,6
44,5
16,2
25,2
25,5
25,7

0,5
0,4

0,0

1,9
4,4
0,7
1,5
0,3

0,1
1,0
0,4

1,3
1,2
0,3
2,3
0,9
0,9
1,4
0,2
0,9
-

1,1
0,7
1,7
13,9

Fuentes: BACEN/DEDIP, BACEN/DEMAB, CETIP, IBGE, gobiernos estatales del Distrito Federal y de los
municipios capitalinos. El valor del PBI (diciembre de 1997) se ha recalculado a septiembre de 1999.
1/ No incluye los pagarés que no están bajo custodia de CETIP, ni las obligaciones de empresas estatales
registradas en Bovespa.
2/ Incluye las deudas frente a INSS, FGTS, PASEP y otras entidades.
3/ Proyección con base en los últimos datos de los que disponen los estados, el Distrito Federal y los municipios
capitalinos. No incluye deudas de administración indirecta (todos los niveles) ni de administración directa de los
municipios no capitalinos.

Si se compara la deuda en proporción de los ingresos líquidos de los estados Xla suma de recursos
tributarios mas transferencias constitucionales que quedan a disponibilidad de los estadosX, ciertamente
un indicador apropiado para determinar la responsabilidad y la sostenibilidadfiscal, es evidente una vez
mas el alto porcentaje registrado por los estados con una tradición arraigada de endeudamiento, pese a que
son los de economías másricas.En este sentido, se destacan Sao Paulo, Minas Gerais, Rio de Janeiro y
Rio Grande do Sul, con un nivel superior a 320% de su ingreso liquido, muy por encima de la una media

Sesión 1 - 40

del 274% para el conjunto de los estados (Cuadro 7). Aunque no del todo comparable, es interesante
observar que el valor de este indicador en los estados brasileños es muy superior al valor medio de 4 4 % en
las provincias argentinas (Cuadro 3).
Cuadro 7
Brasil: Deuda consolidada de los e s t a d o s , septiembre de 1999
(en porcentaje del ingreso estatal líquido real) 4/
Deuda Contractual Interna

Deuda en
Obligaciones
1/
Total

Tesoro
Nacional

Bancaria

Otras
2/3/

19,9

208,2

27,0

5,8

Deuda
Externa
3/

Deuda
Total

13,0

273,9

„
150,5
14,7
129,1
6,7
Acre
96,9
221,4
2,8
9,8
4,4
335,3
Alagoas
6,7
15,7
1,4
23,8
Amapá
152,7
2,0
0,1
0,0
154,9
Amazonas
177,9
27,1
1,7
22,8
229,6
Bahia
125,6
21,8
0,4
38,2
186,0
Ceará
0,1
30,3
0,1
4,1
34,5
Distrito Federal
100,3
17,7
11,5
17,3
146,8
Espirito Santo
440,1
17,5
40,9
3,4
502,0
Goiás
287,7
6,0
2,7
0,9
297,2
Maranhao
297,6
4,4
8,8
7,9
318,6
Mato Grosso
365,9
20,5
4,9
391,3
Mato Grosso do Sul
308,3
4,7
15,0
18,5
346,4
Minas Gerais
70,0
3,8
1,1
10,4
85,3
Pará
163,2
6,5
4,5
1,6
175,9
Paraíba
126,5
14,3
13,0
29,8
183,5
Paraná
36,9
143,5
11,3
3,9
11,0
206,6
Pernambuco
176,2
61,5
0,6
5,1
243,4
Piauí
184,3
47,8
195,1
1,3
16,8
445,2
Rio de Janeiro
71,6
13,8
2,4
1,4
89,2
Rio Grande do Norte
1,2
311,8
7,5
12,4
13,8
346,7
Rio Grande do Sul
119,1
33,6
31,1
183,8
Rondónia
30,3
48,7
1,9
80,8
Roraima
61,7
169,3
13,8
23,9
20,4
289,0
Santa Catarina
3,8
298,5
6,6
0,1
8,8
317,8
Sao Paulo
141,8
16,1
12,2
170,1
Sergipe
15,1
17,1
41,6
73,8
Tocantins
Fuentes: BACEN/DEDIP, BACEN/DEMAB, CETIP, gobiernos estatales del Distrito Federal y de los municipios
capitalinos.
1/ No incluye los pagarés que no están bajo custodia de CETIP ni las obligaciones de empresas estatales
registradas en Bovespa.
2/ Incluye las deudas frente a INSS, FGTS, PASEP y otras entidades.
3/ Proyección con base en los últimos datos de los que disponen los estados, el Distrito Federal y los
municipios capitalinos. No incluye deudas de administración indirecta (todos los niveles) ni de administración
directa de los municipios no capitalinos.
4/ Ingreso líquido real (Orden No. 491 de STN/MF, del 29-II-99) anualizado, que incluyen el ingreso tributario
propio más las transferencias gubernamentales.

Sesión 1 - 4 1

La evolución del sistema federal brasileño en la última década se ha caracterizado por una serie de
déficits y un alto endeudamiento público a nivel subnacional, que fue aliviado por numerosos rescates por
parte del gobierno federal. Esta situación que aún subsiste emana de tres causas especificas42. Primero, los
estados y municipios tienen un gran peso dentro del sector público en lo que se refiere a los ingresos,
gastos y la deuda pública, en función de descentralizaciónfiscal; al mismo tiempo existe una asimetría en
la distribución de los derechos y deberes entre los tres niveles de gobierno, con un sesgo a favor los
gobiernos subnacionales, en cuanto la división de responsabilidades de gasto no están bien definidas.
Segundo, los bancos estatalesfinanciaban a los estados, y algunos todavía mantienen el potencial de
financiamiento43. Tercero, los gobernadores tienen una fuerte influencia política en el Senado, que por
mandato constitucional debería controlar el endeudamiento de los estados y municipios. Como regla
general, el Senado siempre aprobó incrementos a los límites de endeudamiento—por lo menos mediante la
capitalización de intereses—de los estados y municipios44.
Una consecuencia negativa de este cuadro institucional ha sido la enorme dificultad de realizar
ajustesfiscales en los niveles inferiores de gobierno, necesarios para la estabilización macroeconómica,
recurriendo en vez el aumento de la deuda pública y a repetidas operaciones de rescate y otorgase
asistencia a los gobiernos subnacionales a fin de hacer viables la continuidad y el funcionamiento del
sector público. Se presentaron también otros impedimentos importantes al ajuste fiscal que también
emanan de la reciente evolución del sistema federal brasileño reciente. En primer lugar, hubo una
interferencia de los estados en el comercio exterior (variable típica de control nacional) mediante la
tributación de las exportaciones de productos primarios y semielaborados por medio del ICMS (Impuesto
a la Circulación de Mercancías y Servicios). Esta cuestión fue resuelta mediante el compromiso del
gobierno federal de compensar posibles pérdidas líquidas de recaudación de ese impuesto por parte de los
estados. Pese al acuerdo logrado y a las compensaciones realizadas por el gobierno federal, a éste se le
acusa repetidamente de disminuir los ingresos de los estados en función de esta medida tributaria que tiene
un gran impacto sobre el comercio internacional y sobre la eficiencia del sistema productivo45.
En segundo lugar, la estructura tributaria provocó asimismo una guerra entre los estados mediante
la manipulación de las alícuotas del ICMS y la concesión de beneficios ocultos tras empréstitos
subsidiados e inclusive participaciones de capital en acciones de empresas privadas. A raíz de la guerra
tributaria, ha disminuido el ingreso tributario estatal disponible lo cual termina por involucrar al gobierno
federal, ya sea desempeñando su función de intermediario entre los estados o bien compasando algunos
gobiernos subnacionales afectados.
En tercer lugar, dada la magnitud de la devolución de la autoridad tributaria a los estados y
municipios bajo mandato constitucional, la Unión se vio obligada a dejar de hacer uso de los impuestos
tradicionales para el ajuste fiscal, induciendo de esa manera la creación de nuevos impuestos, como por
42Véase

el análisis m s detallado que presentan Serra y Afonso (1999).
á
significativo reordenamiento del sistemafinanciero de los estados que tuvo lugar en los últimos años con el cierre de u
n
gran número de bancos y la transformación de varias institucionesfinancieras estatales en simples agencias de desarrollo (sin
derecho a la captación de recursosfrente al público), siguen sobrando algunos bancos de algunos estados importantes (Sao
Paulo, con el Nossa Caixa Nosso Banco) que pueden, e ciertas situaciones políticas de debilitamiento del gobierno federal,
n
volver a realizar operaciones bancarias en favor de su accionista mayoritario.
^En el futuro los estados podrían tener derecho a emitir obligaciones provocando la desestabilización del mercado de títulos de
deuda pública, aun si en los recientes contratos de renegociación los estados se han comprometido a no emitir deuda en
obligaciones durante un cierto período. Esta regla se vuelve m s plausible cuando así lo determina una ley superior o una
á
resolución del Senado, ya que este ultimo tiene la prerrogativa de aprobar, en ciertas circunstancias, el incremento de los límites
del endeudamiento público.
45E1 problema se resolvió con la Ley Complementaria 8 de 1 9 . Nótese que el sistema de compensación delineado y aprobado
7
96
abre la posibilidad de revisiones en función de las opciones existentes y lleva, inexorablemente, a incrementos e los m n o
n
o ts
retransferidos a los estados.

43Pese al

Sesión 1 - 42

ejemplo la contribución social sobre las utilidades de las empresas y el impuesto sobre las transacciones
financieras. Se trata, en realidad, de impuestos en cascada, de baja eficiencia económica, pero el gobierno
federal ha basado en ellos el aumento de su recaudación tributaria dado que, en función del esquema de
transferencias, cualquier aumento en los impuestos más eficientes (impuesto sobre la renta de las personas
jurídicas y físicas, impuesto sobre los productos industrializados) están preasignados automáticamente a
los estados y municipios los cuales, a su vez, no sienten compromiso alguno con el ajustefiscal a nivel
nacional.
Cabe destacar otras dos consecuencias negativas de la estructura actual de transferencia de
impuestos de la Unión a los estados y municipios. Por un lado, este mecanismo termina por desincentivar
el esfuerzo local de tributación de fuentes propias, sobre todo en lo que respecta a los municipios46. Por
otro lado, los significativos aumentos de las transferencias de ingresos a los estados y municipios a partir
de la Constitución de 1988 han tenido un efecto multiplicador de los gastos de los gobiernos
subnacionales sin mejoras en su desempeñofiscal—es decir, reducción del déficit y de la deuda a niveles
prudentes. Además, podría aducirse que los gobiernos subnacionales maximizan el gasto debido a que
parte de su cobertura queda o quedará a cargo del gobierno federal47.
Por último, cabe señalar elriesgomoral como consecuencia del papelfinanciero desempeñado por
el gobierno central hacia los estados y municipios—además del mecanismo normal de transferencia de
recursos vía el presupuesto federal, con una mayor concentración en las regiones más pobres—mediante
operaciones sucesivas de rescate y renegociación de las deudas de los estados y municipios, y de rescates
de los bancos estatales con activos incobrables frente a los gobiernos estatales, especialmente a través del
Banco Central. Es decir, el procedimiento repetitivo de transferencia constitucional y no constitucional de
recursos tributarios, vinculados a asistencia en la intermediación entre los estados, así como la
implementación de programas de rescate de deudas de los gobiernos subnacionales y de sus instituciones
financieras, han fomentado un comportamientofiscal deficiente a nivel subnacional48.
Es evidente, por lo tanto, la necesidad de mejoras sustanciales en la imposición de responsabilidad
fiscal, ya sea con respecto al control de gastos, al esfuerzo de recaudación de fuentes propias, y a la
limitación del déficit, como del endeudamiento al nivel de estados y municipios. Puede ser criticable el
hecho de que hasta la fecha no se hayan dado pasos hacia estas mejoras en tanto que los gobiernos
subnacionales han tenido un amplio espacio de maniobra con respecto a sus operaciones de crédito y, al
mismo tiempo, el gobierno federal ha sido susceptible a las presiones políticas para implementar
programas de rescate y asistencia a los gobiernos subnacionales49.

^Pese a que la Constitución de 1 8 definió claramente los impuestos a nivel municipal (impuesto sobre los bienes inmuebles,
98
impuesto sobre los servicios, tasas de servicios públicos, etc.) no hay grandes incentivos para aumentar y ni siquiera para aplicar
estos tributos (en los casos en que no existan). En general, los municipios m s pobres del nordeste, por el hecho de que reciben
á
u gran volumen de recursos a través del gobierno federal, tienden a disminuir sus esfuerzos de recaudar impuestos locales. Lo
n
ms o ocurre, por ejemplo, con los municipios m sricosde Sao Paulo, que reciben cuantiosas transferencias estatales por s
im
á
u
porción alícuota del ICMS.
47Wemeck (1995) y Rangel (1999) proponen esta explicación.
48Rangel (1999) trata esta hipótesis formalmente para el caso brasileño.
49Véase Rangel (1999, cap. II) al respecto.

Sesión 1 - 43

IV.

Perspectivas para el futuro

Una manera de resumir la evaluación del desempeño del sector público a nivel subnacional en
Argentina y Brasil es con un fuerte sesgo al desequilibrio vertical, que se puede atribuir en general al
proceso de descentralización y al mecanismo de relaciones intergubernamentales. La devolución de
funciones de gasto en las últimas décadas ha sido algo ambigua—aunque siguiendo más o menos el
principio de subsidiariedad—50 y no siempre fue acompañada de suficientes incentivos a la recaudación de
ingresos tributarios al nivel subnacional. De ahí que los gobiernos subnacionales en lo posible hayan
tratado de interpretar de manera limitada su responsabilidad de gasto, hayan otorgado diversas exenciones
tributarias —que en el caso brasileño se degeneraron en guerras tributarias entre algunos estados —y se
hayan apoyado en la coparticipación de ingresos y transferencias financieras del gobierno central. Todo
esto ha conllevado a que, en varias ocasiones, las autoridades federales se hayan visto obligadas a rescatar
de manera ad hoc—mediante medidas mas sutiles y menos voluminosos en la Argentina— provinciales o
estatales. En resumen, este comportamiento ha contribuido directa o indirectamente a los déficits y al
endeudamiento del sector público.
Para remediar esta situación, ambos países han emprendido programas globales de ajustefiscal con
un horizonte a mediano plazo. Dentro del marco de este esfuerzo, el reordenamiento del mecanismo de
reparto de los ingresos tributarios entre varios niveles de gobierno, en combinación con una reforma del
sistema tributario indirecto, ocupa un rol central. Al respecto, en Argentina, la aprobación de la pendiente
ley de coparticipación nacional resultaría en una excelente posibilidad para hacer más transparente, sólida
y permanente las relacionesfiscales entre el gobierno central y las provincias. Pero quizá el paso mas
importante hacia la responsabilidad fiscal es la legislación inminente sobre reglas fiscales a todos los
niveles de gobierno en Brasil y la reciente promulgación de reglas al nivel federal en Argentina.
Con los antecedentes de rescatefinanciero por el gobierno central, falta de incentivos para la
recaudación de impuestos y renuencia al control de gasto a nivel subnacional, la experiencia internacional
provee una lección clara para ambos países. Esencialmente, tomando en cuenta el peligro del desequilibrio
macroeconómico que enfrentan, no obstante el obvio atractivo del modelo autónomo en una genuina
estructura federal, el modelo de coordinación parece ser más apropiado para asegurar los beneficios del
ajuste fiscal51.
En Brasil, el proyecto de Ley de Responsabilidad Fiscal que se está debatiendo en el Congreso
representa un reconocimiento de esta lección en cuanto reúne ciertas características de los enfoques
australiano y alemán. Además de procurar asegurar la adopción de la regla de oro a cada nivel de
gobierno (en vigencia comenzando en el año 2001)52, la Ley incorpora metas para la deuda pública en
proporción a los ingresos propios más transferencias constitucionales. Esta meta se deberáfijar para cada
nivel de gobierno—bajo la recomendación del Presidente y autorización del Senado—en base del ratio

En la UE este criterio directriz significa que aquellas funciones (educación primaria y secundaria, servicios básicos de atención
médica, servicios públicos, etc.) que se pueden desempeñar m s eficazmente a nivel subnacional, se deban ejercer a ese nivel,
á
dejando u mínimo de funciones (defensa, relaciones exteriores, etc.) al gobierno nacional.
n
CEPAL (1998, p. 221) parece estar de acuerdo con este punto de vista, con énfasis en el papel de la coordinación política m s
a
allá de los aspectos técnicos.
5 2 La regla de oro, aunque presente en la Constitución, n n a fue reglamentado y por lo tanto no se ha usado para contener el
uc
endeudamiento publico. El proyecto de Ley de Responsabilidad Fiscal que fue enviado al Congreso contiene la reglamentación
necesaria la aplicación de la regla.
50

Sesión 1 - 44

consolidado a ese nivel de gobierno53. En relación a las principales categorías de gasto, la Ley también
impone límites sobre salarios y pensiones a servidores públicos (la mayor categoría en los gastos
corrientes). Además, la Ley exige equilibriofinanciero y actuarial en los regímenes de seguro social (ya
con cuentas separadas de los respectivos tesoros estatales) y de que nuevos beneficios solo podrán ser
aprobados previo un aumento equivalente de ingresos o corte equivalente de gastos.
Un aspecto crucial de la Ley es la prohibición estricta de nuevas renegociaciones (inclusive
refinanciamiento o postergación) de deudas de gobiernos estatales por parte del gobierno federal. Además,
la Ley da lugar a la utilización de transferencias voluntarias como instrumentos importantes da
responsabilidadfiscal, y asimismo exige cumplimiento con los límites al endeudamiento y a los gastos de
personal para la realización de tales transferencias.
Otras dos innovaciones de procedimiento presupuestario que también afectan a los gobiernos
subacionales en Brasil merecen ser mencionadas. Por un lado, la introducción de un mecanismo de
compensación presupuestaria en caso de renuncia de ingresos tributarios y en caso de creación de gastos
con carácter continuado, para evitar gastos de larga duración sin recursos parafinanciarlos que pueden
poner en peligro la sostenibilidadfiscal. Por otro lado, la introducción de cortes automáticos bimensuales
si son necesarias durante la ejecución del presupuesto para el cumplimiento de metas aprobadas para el
año.
En Argentina, en cambio, la Ley 25125, de Administración de los Recursos Públicos (también
conocida como Ley de Convertibilidad Fiscal), aprobada en setiembre de 1999, sigue esencialmente el
modelo autónomo, en concordancia con las características institucionales del país. Mientras que establece
una regla de saldo presupuestario equilibrado a nivel federal (con vigencia desde 2003), la Ley solamente
propone invitar a los gobiernos provinciales a adoptar una regla similar54. Bajo la necesidad de un ajuste
fiscal global, inclusive a nivel provincial, en los últimos meses" se han lanzado iniciativas que se
aproximan al modelo de coordinación. En efecto, en el Compromiso Federal (firmado afines de 1999 por
los gobernadores de las provincias) se expresa la intención de tomar medidas legislativas a nivel
subnacional con elfin de adoptar reglas similares a las que contiene la Ley de Convertibilidad Fiscal55. En
las provincias de Salta, San Luis y Tucumán ya hay leyes vigentes que cumplen con el Compromiso56.
Aunque el enfoque argentino está menos definido en el trato de gobiernos subnacionales, quedan
numerosos interrogantes también en la implementación de la Ley de Responsabilidad Fiscal en Brasil. En
especial, esta ley impone un desafío a la capacidad de gestión de los estados y los municipios brasileños
no solamente en los aspectos operacionales de las reglas sino también en cuanto al diseño y ejecución de
reformas—sobre todo en cuanto al sistema de pensiones para empleados públicos y al mecanismo de
impuestos indirectos—que son indispensables para su cumplimiento. Sin embargo, las provincias
Es decir, si el ratio de la deuda de u determinado estado se encuentra por encima del ratiofijado para todos los estados,
n
entonces este estado debe apuntar a u superávit primario que permita reducir s deuda a ese nivel. En cambio, si su ratio está
n
u
por debajo del ratio que se ha designado c m m t para todos los estados, entonces debe seguir simplemente la regla de oro.
o o ea
Lamentablemente, existe elriesgode que este m c ns o en el proyecto de ley original se modifique en la versión del Relator
e a im
e la Cámara de Diputados permitiendo que el ratio de la deudafijado c m m t podría ser alteradas durante el ejercicio.
n
o o ea
54 Según la Ley 25125, "El Poder Ejecutivo nacional invitará a los gobiernos provinciales y al Gobierno Autónomo de la Ciudad
de Buenos Aires al dictado de normas legales e concordancia con lo dispuesto en la presente ley" (Art. 10).
n
55E1 Compromiso Federal (Art. 5) establece que "las provincias y el Gobierno de la Ciudad Autónoma de Buenos Aires se
c m r m t n a impulsar en sus respectivas jurisdicciones, durante el año 2000, el dictado de una legislación que de acuerdo a
o po ee
sus propias normas constitucionales adopte principios o parámetros similares a los establecidos para la Nación por la ley 2 1 2
55
de Administración de los Recursos Públicos o Solvencia Fiscal, con el objeto de disminuir el déficit fiscal, contener el gasto
público y asegurar la transparencia fiscal."
S6Véase Anexo 2. En el m s de enero de 2 0 se aprobaron leyes de este tipo e Catamarca y Formosa, que no se pudieron incluir
e
00
n
e el anexo.
n
53

Sesión 1 - 4 5

argentinas enfrentan un desafío aun más fundamental: el diseño de reglas con suficiente flexibilidad
institucional como para que mantengan el grado deseado de autonomía y que al mismo tiempo ofrezcan de
manera uniforme suficientes incentivos para el ejercicio de disciplina fiscal.
En la Argentina queda pendiente el diseño de las reglas fiscales por parte de los gobiernos
provinciales (que según el Compromiso Federal debería llevarse a cabo en el transcurso del año) y su
posterior cumplimiento. Algo que será necesario comprobar en la práctica también en Brasil. Al respecto,
dada la magnitud de ambos países y la diversidad regional de su estructura económica, sería conveniente
diseñar un mecanismo automático de compensación o suspensión de las reglas para amortiguar el efecto
de choques exógenos asimétricos a nivel subnacional. El Pacto de Estabilización y Crecimiento, bajo la
UEM en la Unión Europea, ofrece un ejemplo de limitada utilidad para los casos argentino y brasileño57.
En el futuro más lejano, también será necesario modificar las reglasfiscales en el contexto de una
política macroeconómica coordinada dentro del Mercosur—un tema que últimamente ha recibido
creciente atención de parte de las autoridades en ambos países—quizá siguiendo los criterios de la UEM.
Idealmente, esto requiere unificación de la política monetaria e introducción de reglasfiscales uniformes
en todos los países miembros del Mercosur a todos niveles de gobierno58, dentro de un espacio económico
integrado. A su vez, este último involucraría la armonización concertada de impuestos indirectos entre los
países miembros59.

Para análisis y simulaciones con respecto a los países G7, véase Kopits y S m n k (1998).
y asy
Vale la pena destacar que los valores de referencia de la UEM se aplican solamente al nivel nacional, lo que transfiere a países
miembros la tarea de distribuir el ajustefiscal entre diferentes niveles de gobierno. Este es u problema potencial para algunos
n
países con una estructura federal. Por ejemplo, en Alemania el problema esta resuelto e cierto grado m da t regla de oro a
n
e i ne
nivel subnacional. En Italia, hasta ahora los gobiernos regionales no h n cumplido c n el pacto de estabilidad interno e
a
o
n
vigencia durante el periodo 1999-2001, que los obliga a u ajuste marginal en su déficit. En el futuro las tensiones internas se
n
van a agudizar, como lo ilustra Reviglio (1999) e referencia a los gastos por servicios de salud, e parte por presiones
n
n
demográficas.
5 9 Esto significa el establecimiento de bases impositivas y tasas mnm s uniformes. Véase Kopits (1999b).
íi a
57
58

Sesión 1 - 4 6

V.

Referencias

Almeida, A. O. (1996), "Evolugáo e crise da divida pública estadual," Texto para Discussáo No. 448.
Rio de Janeiro: EPEA.
Banco Mundial (1996), Argentina: Provincial Finances Study. Washington.
Banco Mundial (1998), Argentina: The Fiscal Dimensions ofthe Convertibility Plan. Washington.
BED (1997), Latin America After a Decade ofReforms: Economic and Social Progress in Latin America.
Washington.
Bisang, R., y O. Cetrángolo (1997), "Descentralización de los servicios de salud en la Argentina," Serie
Reforma de Política Pública, N°47. Santiago de Chile: CEPAL.
Carciofi, R., O. Cetrángolo, y O. Larrañaga (1996), "Desafíos de la descentralización, educación y salud
en Argentina y Chile." Santiago de Chile: CEPAL.
CEPAL (1998), The Fiscal Covenant: Strengths, Weaknesses, and Challenges. Santiago de Chile.
Cetrángolo O., y J.P. Jiménez (1995), "El conflicto en torno a las relacionesfinancieras entre Nación y
provincias", Serie Estudios 9 y 10. Buenos Aires: CECE.
Cetrángolo O., y J.P. Jiménez (1996), "Apuntes para un nuevo régimen de Coparticipación Federal de
Impuestos", Serie Estudios 13. Buenos Aires: CECE.
Cetrángolo O., y J.P. Jiménez (1997), "Aportes del Tesoro Nacional: discrecionalidad en la relación
financiera entre la Nación y las provincias", Serie Estudios 21. Buenos Aires:
CECE.
Cetrángolo, O., M. Damill, R. Frenkel, y J.P. Jiménez (1997) "La sostenibilidad de la políticafiscal en
America Latina: el caso argentino", Serie de Documentos de Trabajo R-315.
Washington: BID.
Craig, J. (1997), "Australia", en T. Ter-Minassian (ed.), Fiscal Federalism in Theory and Practice.
Washington: FMI.
Damill, M. (1997),"Análisis de la situaciónfiscal de la Provincia de Mendoza," inédito.
Dillinger, W., y S.B. Webb. (1998), Fiscal Management in Federal Democracies: Argentina and BrazilWashington: Banco Mundial, marzo.
Eichengreen, B., y J. von Hagen (1995), "Fiscal Policy and Monetaiy Union: Federalism, Fiscal
Restrictions and the No-Bailout Rule," CEPR Discussion Paper No. 1247. London: Centre for Economic
Policy Research.
Kopits, G. (1999a), "Experiencia internacional con reglas macro-fiscales: lecciones para la
Argentina," presentado en el Seminario sobre Reglas Macro-Fiscales, Instituciones e Instrumentos
Sesión 1 - 47

Presupuestarios Plurianuales, Ministerio de Economía y Obras y Servicios Públicos, Buenos Aires: 10 12 de mayo.
Kopits, G. (1999b), "Política tributaria en una unión económica: perspectivas para el Mercosur,"
presentado en el VII Congreso Tributario del Consejo Profesional de Ciencias Económicas de la Capital
Federal. San Martín de los Andes, 4-8 de septiembre.
Kopits, G., y D. Mihajlek (1993), "Fiscal Federalism in the New Independent States" en V. Tanzi (ed.),
Transition to Market: Studies in Fiscal Reform. Washington: FMI.
Kopits, G., y S. Symansky (1998), Fiscal Policy Rules, Occasional Paper 162. Washington: FMI.
McKinnon, R.I. (1996), "Monetary Regimes, Government Borrowing Constraints, and MarketPreserving Federalism: Implications for EMU," paper presented at the Conference on The
Nation State in a Global Information Era: Policy Changes, Queen's University. Kingston,
Ontario, 13-15 de noviembre.
Mendes, M. J. (1999), Incentivos eleitorais e desequilibriofiscal de estados e municipios. Sao Paulo:
Instituto Fernand Braudel de Economía Mundial.
Millar, J. (1997), 'The Effects of Budget Rules on Fiscal Performance and Macroeconomic
Stabilization," Bank of Cañada Working Paper 97-15. Ottawa, junio.
Nicolini, J.P., L. Posadas, J. Sanguinetti, P. Sanguinetti, y M. Tommasi (1999), "Decentralization, Fiscal
Discipline in Subnational Governments, and the Bailout
Problem: the Case of Argentina." Washington: BID.
Rangel, M. A. (1999), "Resgates financeiros, restrigáo orgamentária fraca e posturafiscal nos estados
brasileiros," Disserta?ao de Mestrado. Rio de Janeiro: Pontificia
Universidade Católica, julio.
Reviglio, F. (1999), "A System in Need of Reform: Health Carew in Italy," Washington, 3 de diciembre
(inédito).
Scott, R.H. (1983), The Austrálian Loan Council and Public Investment, Occasional Paper No. 31,
Center for Research on Federal Financial Relations (Camberra: Australian National University).
Secretaría de Programación Económica y Regional (1999), Informe Económico Regional. Buenos Aires,
varios números.
Serra, J., y J.R.R. Afonso (1999), "Federalismofiscal á brasileña," Revista do BNDES. Rio de Janeiro,
diciembre.
Spahn, P.B., y W. Fottinger (1997), "Germany," en T. Ter-Minassian (ed.), Fiscal Federalism in Theory
and Practice. Washington: FMI.
Tanzi, V. (1996), "Fiscal Federalism and Decentralization: A Review of Some Efficiency and
Macroeconomic Aspects," Annual World Bank Conference on Development Economics, 1995.
Washington: Banco Mundial.
Tavares, M.A.R., A. Manoel, J.R.R. Afonso, y S.P.P. Nunes (1999), "Principios y reglas para las
Sesión 1 - 4 8

finanzas públicas: la propuesta de Ley de Responsabilidad Fiscal del Brasil," presentado en el XI
Seminario Regional de Política Fiscal de CEPAL. Brasilia, enero.
Ter-Minassian, T. (1997a), "Intergovernmental Fiscal Relations in a Macroeconomic Perspective: an
Overview," en T. Ter-Minassian (ed.), Fiscal Federalism in Theory and Practice. Washington: FMI.
Ter-Minassian, T. (1997b), "Brazil", en Teresa Ter-Minassian Fiscal Federalism in Theory and
Practice. Washington: FMI.
Ter-Minassian, T., y J. Craig, (1997), "Control of Subnational Government Borrowing" en
Minassian (ed.), Fiscal Federalism in Theory and Practice. Washington: FMI.

Ter-

Werneck, R. L. F. (1995), "Federalismofiscal e política de estabilizagáo no Brasil," Revista Brasileira
de Economía, abril/junio.
Zapata, J. A. (1999), Argentina: el BID y los gobiernos subnacionales. Washington: BID.

Sesión 1 - 49

Anexo 1
Argentina: Reglamentos sobre el crédito público en las provincias
El siguiente cuadro ofrece un resumen de las exigencias más importantes de las constituciones
provinciales para la aprobación y el uso del crédito público60.
Jurisdicción

Autorización Necesaria.

Bs. Aires

Ley sancionada por 2/3 Los recursos se aplicarán a objetos
miembros presentes de cada determinados.
Cámara.
Especificar recursos especiales
con que se hará el servicio de la
deuda y su amortización.
Debe especificarlos la ley que los
autoriza.
Responsabilidad de la autoridad
que los invierta o destine a otros
objetos.

C.A.B.A.

Mayoría absoluta del total de
los
miembros
de
la
Legislatura.
Crédito público interno o
extemo. Autorizada por ley.
Determinación concreta de su
objetivo.

Catamarca

La legislatura autoriza al
Poder Ejecutivo con el voto
de 2/3 miembros presentes
de cada cámara.

Prohibido
cubrir
los
gastos
ordinarios de la administración.
No pueden aplicar los recursos a
otro objeto.

Servicio de la deuda y su
amortización
no
podrán
exceder el 20% de las rentas
efectivas de la provincia en el
quinquenio anterior.

Córdoba

Autoriza al Poder Ejecutivo Para el financiamiento de obras
con los 2/3 de los miembros públicas,
promoción
del
presentes en cada cámara.
crecimiento económico y social,
modernización del Estado y otras
necesidades excepcionales o de
extrema urgencia.

Amortización e intereses: no
pueden superar el 20% de la
renta provincial.
Base de cálculo: el menor de
los
ingresos
anuales
ordinarios de los tres últimos
ejercicios, considerados a
valores constantes.

Corrientes

Ley sancionada por 2/3 de Se deben aplicar a los objetos En ningún caso se podrá
votos del total de miembros determinados en la ley de su comprometer más del 25% de
de cada cámara.
creación.
la renta anual de la Provincia.
El Poder Ejecutivo debe ser
facultado
por
el
Poder
Legislativo.

60

Utilización Endeudamiento

Limite en el crédito público

Elaboración propia de acuerdo con información recopilada por Zapata (1999) de constituciones provinciales.

Sesión 1 - 50

Limite en el crédito público

Jurisdicción

Autorización Necesaria.

Utilización Endeudamiento

Chaco

Ley sancionada con el voto
de los dos tercios de los
miembros que componen la
Legislatura.
Corresponde a la Cámara de
Diputados autorizar al Poder
Ejecutivo
a
contraer
empréstitos, emitir fondos
públicos y celebrar cualquiera
otra operación de crédito.

Determinación
concreta
del No excederán anualmente del
objetivo de la operación y de los veinticinco por ciento de las
recursos
afectados
para
su rentas de la provincia.
servicio.

Chubut

Facultar al Poder Ejecutivo, Bases y objetos determinados.
con la mitad más uno de sus No pudiendo ser autorizados para
equilibrar
los gastos ordinarios
de
miembros.
la administración.
Facultar al Poder Ejecutivo Bases y objetos determinados. No En ningún caso la totalidad de
servicios
de
los
con la mitad más uno de los pueden ser autorizados para ios
equilibrar los gastos ordinarios de empréstitos comprometerán
miembros de cada Cámara.
la administración.
más de la cuarta parte de las
Ni el numerario obtenido de los rentas de la provincia .
mismos, ni los fondos públicos que
se emitan, podrán ser aplicados a
otros objetos que los determinados
por la ley de su creación.

Entre
Ríos

Formosa

Jujuy

Corresponde
al
Poder
Legislativo
las
siguientes
atribuciones:...
Autorizar al Poder Ejecutivo a
contraer empréstitos basados
en el crédito de la Provincia.
Dos tercios de votos bajo
pena de nulidad.
La Legislatura podrá autorizar
mediante ley especial por el
voto de los dos tercios de la
totalidad de sus miembros.
Corresponde a la Legislatura
aprobar o rechazar el arreglo
de pago de la deuda interna y
extema de la Provincia.

Financiamiento de obras públicas,
reformas
agrarias,
crecimiento
económico y social o para atender
gastos originados por catástrofes y
otras necesidades excepcionales o
de extrema urgencia.
Con fines de promoción económica
la
Provincia
podrá
contraer
empréstitos destinados a financiar
obras productivas específicamente
planificadas y cuyos servicios
financieros deberán ser cubiertos
por los rendimientos de la obra.

Sesión 1 - 5 1

El servicio de la totalidad de
las deudas autorizadas y que
hubieren
de
autorizarse,
podrá ???comprometer más
de la cuarta parte del
promedio de los cálculos de
recursos ordinarios anuales.

Jurisdicción

Autorización Necesaria.

Utilización Endeudamiento

Limite en el crédito público

La Pampa

Deberá sancionársela ley con
dos tercios de votos de los
miembros que componen la
Cámara de Diputados.

Especificar el objeto al que los
fondos se destinan y los recursos
asignados para su servicio.
Podrá
autorizarse
únicamente
cuando
su
producido
sea
destinado a la ejecución de obras
públicas, para hacer efectivos
planes de colonización agraria o
para atender gastos originados por
catástrofes, calamidades públicas
u otras necesidades excepcionales
o impostergables del Estado,
calificadas por ley. En ningún caso
podrá aplicarse a enjugar déficit de
administración.

No podrán exceder del
veinticinco por ciento de la
renta ordinaria anual de la
Provincia.

La Rioja

Con el voto de los dos tercios
de los miembros en funciones
de la Cámara de Diputados.
Corresponde a la Cámara de
Diputados:...
Autorizar
al
gobernador
a
contraer
empréstitos, emitir
títulos
públicos y celebrar cualquier
otra operación de crédito con
arreglo a lo dispuesto por
esta Constitución.
Ley sancionada por 2/3 de los
miembros presentes de cada
cámara.
Corresponde
al
Poder
Legislativo:
Reglamentar
la
administración del Crédito
Público. Facultar al Poder
Ejecutivo
para
contratar
empréstitos o emitir fondos
públicos como lo determina la
Constitución.

Ningún compromiso de esta clase
podrá contraerse sino para obras
públicas.
Los recursos provenientes de ese
tipo de operaciones no podrán ser
distraídos ni interinamente de su
objeto, bajo la responsabilidad de
la autoridad que los invierta o
destine a otros objetos.

El servicio de la totalidad de
las deudas autorizadas, no
podrá comprometer más del
veinticinco por ciento de la
renta provincial, a cuyo efecto
se tomará como base el
cálculo de recursos menor de
los tres últimos años.

Corresponde a la Cámara de
Representantes:
Autorizar al Poder Ejecutivo,
con los dos tercios de votos
de los miembros presentes.

Bases y objetos determinados, no
pudiendo ser autorizados para
equilibrar los gastos ordinarios de
la administración.
Ni el numerario obtenido de los
mismos ni los fondos públicos que
se emitan, podrán ser aplicados a
otros objetos.

Mendoza

Misiones

Deberá especificar los recursos
especiales con que no deba
hacerse el servicio de la deuda y
su amortización.
No podrán aplicarse los recursos
que se obtengan por empréstito,
sino a los objetos determinados,
que debe especificar la ley que lo
autorice, bajo responsabilidad de la
autoridad que los invierta o destine
a otros objetos.

Sesión 1 - 52

En ningún caso la totalidad de
los
servicios
de
los
empréstitos comprometerán
más de la cuarta parte de las
rentas de la provincia.

Utilización Endeudamiento

Jurisdicción

Autorización Necesaria.

Neuquén

Corresponde
al
Poder
Legislativo:
Facultar al Poder Ejecutivo,
con el voto favorable de la
mayoría absoluta de todos
sus miembros, a contraer
empréstitos.
Las leyes que autoricen la
contratación de empréstitos
serán dictadas en sesión
especial
de la
Cámara,
convocada al efecto con tres
días de anticipación. Dictar la
ley orgánica del crédito
público.
Por ley especial de la
Legislatura podrá autorizarse
la emisión de empréstitos o
emitir fondos públicos con
base y objeto determinado,
no pudiendo ser autorizados
para equilibrar los gastos
ordinarios
de
la
administración.
Con fines de promoción
económica la provincia con el
acuerdo
de
la mayoría
absoluta
de
todos
los
miembros de la Cámara de
Diputados podrá suscribir
empréstitos
destinados
a
financiar obras productivas
específicamente
determinadas por el Consejo
de
Planificación
cuyos
servicios
financieros
quedarán
cubiertos
por los
rendimientos de la obra.
Ley sancionada con los dos No pueden aplicarse los recursos
tercios de votos de los sino los fines determinados, que
miembros de la Legislatura.
debe especificarse en la ley que
La Legislatura tiene las los autoriza, bajo responsabilidad
siguientes
facultades
y de la autoridad que los invierta o
deberes: Considera el pago destine a otros objetos.
de la deuda interna y externa Pueden autorizarse únicamente
de la Provincia.
cuando sea destinado a la
ejecución de obras publicas, para
hacer efectiva la reforma agraria o
para atender gastos originados por
catástrofes, calamidades públicas
y otras necesidades excepcionales
e impostergables
del estado
calificadas por ley, sin poder
aplicarse en ningún caso a enjugar
déficit de administración.

Río Negro

Sesión 1 - 53

Limite en el crédito público
En ningún caso la totalidad de
los servicios del empréstito
comprometerán más de la
cuarta parte de las rentas
generales de la provincia. El
numerario obtenido de los
mismos podrá ser aplicado a
otros
destinos
que
ios
establecidos en la ley de su
creación.

Deben
especificar
los
recursos con que debe
afrontar los servicios de la
deuda y su amortización, los
que en ningún caso podrán
exceder el 25% de la renta
ordinaria anual.

Jurisdicción

Autorización Necesaria.

Utilización Endeudamiento

Limite en el crédito público

Salta

La
Legislatura
puede
autorizar empréstitos sobre el
crédito
general
de
la
Provincia o emisión de fondos
públicos.
Sancionada por los dos
tercios de votos de los
miembros presentes de cada
Cámara,
actuando
como
originaria la de Diputados.

Toda ley que sancione empréstitos
especifica los recursos especiales
con que deben atenderse los
servicios de la deuda y su
amortización, así como los objetos
a que se destina el monto del
empréstito.

La totalidad de los servicios
de intereses y amortización
no puede exceder la cuarta
parte
de
los
recursos
ordinarios
del
tesoro
provincial.

San Juan

Son atribuciones
de la
Cámara de Diputados:
Autorizar al Poder Ejecutivo a
contraer
empréstitos,
determinando sus intereses y
las bases y condiciones para
su amortización; emitir títulos
públicos y cualquier otra
operación de crédito con
arreglo a lo dispuesto por la
Constitución.
Ley sancionada con el voto
de los dos tercios de los
miembros de la Legislatura.
Corresponde a la Legislatura:
Autorizar la celebración de
contratos sobre empréstitos
de dinero basados en el
crédito de la Provincia u otros
de utilidad pública..

Los recursos no pueden ser
distraídos ni provisoriamente de
sus fines. La ley que provea a
otros compromisos extraordinarios,
debe establecer los recursos
especiales
con
que
deben
atenderse los servicios de la deuda
y su amortización.

En ningún caso el sen/icio de
la totalidad de las deudas
autorizadas,
puede
comprometer más del 25% de
la renta provincial, a cuyo
efecto se toma como base al
cálculo de recurso menor de
los últimos tres años.

San Luis

Santa Cruz

Santa Fe

Corresponde
al
Poder
Legislativo:...
Autorizar al Poder Ejecutivo,
con dos tercios de votos de
los miembros presentes, para
contraer empréstitos.
Reglamentar
la
administración del crédito
público.
Corresponde a la Legislatura:
Autorizar al Poder Ejecutivo
para celebrar contratos y
aprobar o desechar los
concluidos
"ad-referendum"
de la Legislatura.

El servicio de la totalidad de
las deudas provenientes de
empréstitos
no
puede
comprometer más de la
cuarta parte de la renta
provincial.

Sesión 1 - 5 4

Jurisdicción

Autorización Necesaria.

Utilización Endeudamiento

Limite en el crédito público

Sgo. Del
Estero

La Cámara de Diputados
puede autorizar empréstitos
sobre el crédito general de la
provincia o emisión de fondos
públicos. La ley que lo
autorice debe ser sancionada
por los 2/3 de votos de la
totalidad de los miembros.
Corresponde
al
Poder
Legislativo:
Autorizar al Poder Ejecutivo a
la emisión de fondos públicos
o de empréstitos sobre el
crédito
general
de
la
Provincia.
Debe tener dos tercios de
votos de la totalidad de los
miembros, debiendo prever
recursos especiales con que
haya de hacerse efectivo el
servicio de la deuda.

Debe especificar los recursos
especiales
con
que
deben
atenderse los servicios de la deuda
y su amortización, así como los
objetos a que se destina el monto
del
empréstito.
Las
sumas
obtenidas por empréstitos no
pueden aplicarse a otros objetos
que los especificados en la ley que
los autorice.

La totalidad de los servicios
de intereses y amortización
no pueden exceder la cuarta
parte
de
los
recursos
ordinarios
del
Tesoro
Provincial.

T.del
Fuego

La
Legislatura
podrá
autorizar, mediante leyes
especiales sancionadas con
el voto de los dos tercios de
sus miembros, la captación
de empréstitos o la emisión
de títulos públicos.

Base y objeto determinados.
No podrán ser utilizados para
equilibrar
los
gastos
de
funcionamiento y servicios de la
administración.

Tucumán

Necesita de la sanción de los
dos tercios de votos de la
totalidad de los miembros de
la Legislatura, entendiéndose
por la totalidad de los
miembros
a
los
que
estuvieren en ejercicio de sus
funciones en el momento de
la sanción. Corresponde al
Poder Legislativo: Autorizar al
Poder Ejecutivo para contraer
empréstitos basados en el
crédito de la Provincia-

Deben especificarse los recursos
especiales con que debe hacerse
el servicio de la deuda.
Los fondos públicos que se emitan
y el numerario obtenido por el
empréstito no podrán ser aplicados
a otros objetos determinados por la
ley de su creación.

Sesión 1 - 5 5

Anexo 2
Argentina: Leyes de solvencia fiscal en las provincias
A diciembre de 1999 se encuentran vigentes leyes de este tipo en Salta, San Luis y Tucumán. En el
siguiente cuadro se resumen en forma comparada las tres leyes.

Déficit

Uso del Crédito

Deuda

Gasto corriente

61

Salta
San Luis
Tucumán
Equilibrio presupuestario.
No podrá existir déficit
Se admitirá déficit
corriente, los gastos
presupuestario cuando sea
corrientes serán atendidos
generado por acontecimientos
exclusivamente con los
extraordinarios, calificados
ingresos corrientes. Se
por el Fiscal de Estado a
exceptúan de este cálculo las
requerimiento del
erogaciones de carácter
Gobernador.
extraordinario61.
El uso del crédito debe
El Estado Provincial solo
responder a las siguientes podrá hacer uso del crédito
pautas: deberá destinarse a con el objeto de financiar
gastos de capital; deberá inversiones o fortalecer
tomarse a largo plazo; y institucionalmente su
podrá comprometerse,
organización o refinanciar
como máximo a la tasa de pasivos. Ningún gasto
interés que resulte del
corriente podrá ser financiado
promedio de tasas de
con préstamos de largo plazo.
mercado financiero
nacional o internacional.
No deberá exceder el 70% de La deuda total no podrá
los recursos corrientes
superar el 3% del PBG, ni
anuales. Para este cálculo no el equivalente de
se computaran los prestamos recaudación de tributos
que contratare el Estado
provinciales
provincial para ser
correspondientes a 1,5
transferidos a los municipios. años.
La amortización de la
deuda pública deberá ser
financiada exclusivamente
por: recursos provenientes
de la venta de activos
públicos; superávit fiscal
corriente; y uso del crédito,
en ese orden.
La tasa de crecimiento
La tasa de incremento de los
anual del gasto corriente no gastos corrientes no podrá
deberá superar la tasa de superar la tasa de incremento
crecimiento anual del PBG. de los ingresos corrientes.
Si la tasa de variación del Cuando la tasa de variación
PBG resultara negativa, el de estos ingresos sea
gasto corriente, a lo sumo, negativa, el gasto real podrá a
podrá mantenerse
lo sumo mantenerse

La ley no especifica quien determina el carácter de extraordinario.

Sesión 1 - 5 6

Tucumán
San Luis
constante.
constante.
El gasto corriente no deberá Se exceptúan las erogaciones
superar el 50% del
de carácter extraordinario.
El producido de la venta de
Presupuesto total.
activos no podrá utilizarse
Los gastos corrientes
para atender gastos
deberán financiarse
corrientes.
exclusivamente con
recursos tributarios
corrientes.
El producido de la venta de
activos y de otras
transferencias de capital no
deberán destinarse a
financiar gastos corrientes.
No podrá superar el 65% de
No podrá superar el 65% de
los recursos corrientes netos,
los recursos corrientes netos.
ingresos netos: total de
ingresos corrientes menos
transferencias a
Municipalidades y Comunas
Rurales.
El Poder Ejecutivo elaborará
El Poder Ejecutivo
elaborará un presupuesto un presupuesto plurianual de
plurianual de al menos tres al menos tres años.
años.
Invitase a los Municipios a El Poder Ejecutivo
adherirse a la presente Ley. acordará con los estados
municipales el dictado de
normas legales en
concordancia con lo
dispuesto en la presente
Ley.
Será obligación del Poder
Ejecutivo la publicación
semestral de la ejecución
presupuestaria con detalle
de gastos, recursos, déficit
y endeudamiento en forma
desagregada.
Salta

Gasto en personal

Presupuesto
Plurianual
Municipios

Difusión

Sesión 1

Sesión 1

SUB-NATIONAL GOVERNMENTS AND CREDIT MARKETS:
SELECTED CASES

By
Maria E. Freire, Marcela Huertas, Benjamín Darche

* Fifth Annual Bank Conference on Development, Chile, June 20-22, 1999. This report draws from a
paper by the same authors presented in the First World Bank on Sub-national Capital Markets, held in
Santander in October 1998. The authors acknowledge the comments received from Pilar Solans, Michael
Shaeffer, Augusto de La Torre, Guillermo Perry, Clemente del Valle, Steven Webb. The contents of this
report are solely the responsibility of the authors; the study recommendations do not reflect the official
position of the World Bank.

Sesión 1 - 59

Table of Contente
Executive Summary

61

Introduction
I.

61

Decentralization, Local Government Borrowing and National Fiscal Stábilüy

A.
B.

62

Overview of the Latin American Región
62
Three Cases of Financial Autonomy and Fiscal Adjustment................................................63
1)
2)
3)

Argentina
Brazil
Colombia

64
64
65

C. Why do Local Governments Borrow?.....................................................................................66
D. The Need for Controls
67
II. Summary and Lessons ofSix Bond Offerings

A.
B.

The Cases
70
Analysis, Lessons Learned and Recommendations for the Further Development of Subnational Capital Markets.....................................................................
1)
2)
3)

III.

69

International Bond Structures
Domestic Bond Structures
Constraints to the Development of Sub-national Credit Markets

72
73
74

A Road Map to Understanding Sub-national Capital Markets

References

79
83

Appendices
Appendix 1: Ciudad de Buenos Aires, Argentina
Appendix 2: Province of Mendoza, Argentina
Appendix 3: Municipality of Guaymallen, Argentina
Appendix 4: City of Rio de Janiero
Appendix 5: City of Ibague, Colombia
Appendix 6: Department of Valle del Cauca, Colombia
Appendix 7: Latin America Sub-national Bond Issues

Sesión 1 - 6 0

Executive Summary
Within the framework of increasing decentralization, the need for many local governments to
access international and domestic capital markets is growing. As urban migration increases,
local authorities are under pressure to provide more services with fewer resources from the
central government, and an inadequate tax base to provide them independently. Sub-national
borrowing is one method to finance such investments, and makes sense in cases where the
useful Ufe ofthe service is long (such as schools and roads). However, the risks of sub-national
borrowing are very high for central governments, especially in new democracies where the
boundary between national and sub-national governments is vague. A central government can
have its credit ruined in international markets by defaults of local governments. This report
examines sub-national access to domestic and international capital markets, using case studies
from Argentina, Brazil and Colombia. These case studies highlight both the challenges raised
by allowing sub-national access to dijferent types of debtfinancing, and the need for a
regulatory framework that protects all interested parties: local governments, central
governments and the bond issuers. When the necessary precautions are in place, however,
local capital markets can be a viable and useful source of finance. At the end ofthe report, a
roadmap is provided to outline the necessary conditions that allow sub-national authorities to
access this financing without jeopardizing the national fiscal stability.

Introduction
The growing interest of sub-national governments in accessing domestic and international capital
markets has raised ambivalent reactions among policy makers. On the one hand, mayors and governors
believe that capital markets will bring cheaper and longer-term capital to finance infrastructure projects as
well as opportunity for good business. On the other hand, central governments are concemed with the
effect it may have on the overall national fiscal situation. The reality show increasing numbers of cities
which have been issuing bonds: during 1991-98, there were 120 cases in Latin America, 150 in ECA, 11
in East Asia and near 500 in South Africa. Asia's local bond market is estimated at $477 billion; Czech
municipalities have issued municipal bonds to offset the negative impact of declining transfers into
capital outlays. Poland, Russia, South Africa, and Zimbabwe have active bond markets.
Worries thatfiscal decentralization may contribute to structural déficits andfiscal imbalances are
common in Latin American and other countries. In countries where the main policy priority is to control
aggregate public sector borrowing, there is a genuine concern thatfiscal decentralization may accentuate
this problem. Whenever traditions of fiscal responsibility are weak, accountability is immature and
administrative discipline is poorly developed, there is ariskthat lower level governments may abuse their
borrowing authority, contributing to aggregate fiscal imbalance with accompanying adverse
macroeconomic consequences (Bird, 1995; Prud'homme, 1995; Tanzi 1996). There is a generalized
consensus that without a proper regulatoryframework, easy access to borrowing (e.g. to debt financing)
may lead to aggregatefiscal imbalances.
However, many would argüe that with adequate legalframework and with the macro fundamentáis
in place, local access to capital markets will be compatible with macrofiscal instability while promoting
the development of an important segment of the financial markets. Necessary conditions include the
existence of supervisory authorities; judicial enforceability of contracts and agreements, adequate
accounting, disclosure, and reporting standards and availability of skilled staff.

Sesión 1 - 6 1

This paper reviews the experiences of some recent bond issues (including Brazil, Argentina and
Colombia) and illustrates the dynamics of the capital markets and the regulatory andfiscal impact issues
these transactions raise. Part I provides background on the macroeconomic framework, Part II summarizes
the bond issues, and Part IH identifies the main issues and makes recommendations on the framework that
needs to be put in place to guarantee a healthy function of sub-national markets.

I. Decentralization, Local Government Borrowing and
National Fiscal Stability
A.

Overview of the Latin American Región

Decentralization has been a reality in Latin America in the last decade. Nations throughout the
región have shared responsibilities with local governments, backed by automatic revenue transfers that
more than doubled local govemment revenues, while achieving a democratic transition. Virtually every
mayor and council member in about 13,000 units of state and local govemment is now chosen by
democratic vote.
By international standards, however, LAC countries are relatively centralized. For the average
OECD country, 35% of the public spending is the responsibility of local governments; in the case of
LAC, local governments account for only 14% of public spending. The situation varíes significantly
among large countries (where the share of sub-national spending is above 45%) and smaller ones.
Decentralization has a significant impact on publicfinances. In some countries, decentralization
has been adopted as a mechanism for reducing central govemment déficits and cutting aggregate publicsector expenditures. For example, central governments in Central and Eastem Europe have tumed over
highly subsidized activities like housing supply and water provisión to local authorities to run at their own
expense.
In Latin America, local authorities have secured sufficient transfers or shares of centrally collected
taxes tofinance services at their current or mandated level. In Argentina, Bolivia, Brazil, Venezuela, and
other countries, revenue-sharing commitments for several years into the future were spelled out in
national laws so as to reduce thefinancial risksof decentralization for local governments.
Actually, decentralization in Latin America has favored local govemment budgets. In general, the
process has started by transferring a larger share of centrally collected revenues to localities, with only
few spending obligations being shifted to local authorities. Only after the transfer of revenues is well
under way has the devolution of services begun. For example, in Brazil, the 1988 Constitution shifted 6%
of total public revenues from the central govemment to the sub-national governments. The new
constitution did not, however, transfer any expenditure responsibility to states and municipalities. The
result: a windfall gain for municipalities but a tightening of the central govemment's fiscal situation. In
countries like Argentina and Colombia the process was similar: subnational governments ended up with
relatively more revenues than additional responsibilities (see Stein, 1998 and Terimassiani, 1997).
Why decentralize? Decentralization seems to be primarily a politically driven process supported by
increasingly influential local authorities. From an economist point of view, the most common arguments
are those of Tiebout (1961), Musgrave (1969) and Oates (1972) who argüe that decentralization would
promote a more efficient allocation and delivery of local public services, as well as improving the
relationship between local preferences and the provisión of local services. Ongoing Bank work tries to
establish the nature and extent of such benefits.

Sesión 1 - 6 2

However, decentralization can also entail costs in terms of distribution equity and macroeconomic
management. The most seriousriskis the fact that local governments may have no incentives to behave in
a fiscally responsible manner. On the one hand, expenditures are more easily decentralized than revenues
(there are few good local taxes). This difference between assumed responsibilities and own revenues (or
vertical fiscal gap) is generally financed by transfers from the central government. Heavy reliance on
revenue transfers may weaken the budget constraints of sub-national governments. This is especially the
case where the sub-national governments have borrowing autonomy and the central government cannot
avoid bailing them out in case of financial trouble. In some cases, foreign borrowing by sub-national
government can ruin a national government's credit in international markets, if the market fails to
distinguish between local government defaults and central government performance.

B.

Three Cases of Financial Autonomy and Fiscal Adjustment

Argentina, Brazil and Colombia have the three most decentralized public sectors in Latin America
(see Table 1) and along with India are the most decentralized democracies in the world. Sub-national
governments account for about half of public spending in Brazil and Argentina and almost 40% in
Colombia. In Argentina and Brazil, the 1980s saw the retum of democracy, the strengthening of federal
practices and also poor macroeconomic performance with unsustainablefiscal imbalances. While national
issues were at the core of thefiscal crises, inter-governmentalfiscal relations andfiscal management of subnational governments also played crucial roles (Dillinger and Webb, 1998). Sub-national déficits made a
significant contribution to the aggregate public sector déficit and stabilization was not possible without a
strict control of sub-national déficits.
Table 1
Decentralization in Lac, 1995
Sub-national / Total
Country
Govt. Spending
Argentina
49.3
Brazil
45.6
Colombia
39.0
Bolivia
26.7
México
25.4
Venezuela
19.6
Uruguay
14.2
Chile
13.6
Honduras
12.3
Perú
10.5
Guatemala
10.3
Ecuador
7.5
Trinidad
7.2
Paraguay
6.2
El Salvador
6.0
Nicaragua
5.2
Panama
3.2
Dominican Rep
2.9
Costa Rica
2.3
Avg. LAC
14.6
Avg. OECD
35..3
Source: IDB, Latin America, a Decade of Reforms, 1999

Sesión 1 - 6 3

1) Argentina
In April 1991 Argentina introduced a highly successful stabilization plan (the Convertibility Plan).
Axnong other measures, the planfixed the exchange rate to the dollar, required that the monetary base not
exceed the dollar valué of international reserves (which transformed the Central Bank into a Currency
Board) and removed the power to devalue the currency from the Ministry of Economy. A difficult fiscal
adjustment was necessary to bring down inflation and real interest rates. This was achieved with massive
privatization, govemment lay-offs, imposition of a severe budget constraint on provinces as they were
prohibited from borrowingfrom the provincial Banks and the Central Bank.
The effect of this budgetary constraint was not felt until 1994 The Tequila crisis in December 1994
hit the Argentine economy especially hard, leading to an important capital outflow, economic slow down
and substantial reduction of provincial revenues . The crisis was a test for the hard budget constraint put in
place by the Convertibility Law. Atfirst, provincial govemments faced the fall in revenues by trying to
borrow from their own banks and the Treasury. Since the Central Bank was unable to print money or
rediscount provincial banks, and the national treasury refused to bail out the provincial govemments,
provinces ran into arrears, paid staff and suppliers with bonds, and borrowedfrom the banks pledging their
shared revenues (coparticipación)'. By mid-1996, provincial debt had reached $17 billion, compared with $5
million in 1991.
By the end of 1996, it was clear that the budget constraints were there to stay. As the central
govemment refused automatic bailouts2, most of the provinces began adjusting their expenditures. First they
cut the wage bilí 5% in real terms in 1995 and another 5.5% in 1996. A small recovery in revenues (4%)
associated with GDP recovery was enough to cut the provincial déficit in half.
The impact of sub-sovereign borrowing on the nationalfiscal condition in Argentina was offset by the
macroeconomic constraints placed by the Convertibility Law. Provinces had no choice but to honor their
debt service payments to commercial banks and international creditors because direct transfers of coparticipation revenues secured these payments to the provincial creditors. The provinces did not receive their
revenue transfers until after creditors were paid. This created a strain on provincial operating budgets, which
led to substantial adjustments in less than two years.3
2) Brazil
In 1997, Brazil per capita sub-national debt reached $800 per inhabitant. This was the highest in Latin
America and one of the highest in the world: (Argentina averaged $500). This situation resultedfrom a
series of debt crisis at provincial level which were absorbed by the central govemment, without taking
measures to promote/impose stricter rules in subnational borrowing. Thefirst crisis occurred in the late
1980s with the international debt crisis. Brazilian states - which had borrowed heavily to finance
infrastructure projects during the 1980s - defaulted on foreign debt and the federal govemment (which had
guaranteed much of the debt) had to absorb it. The second crisis came with the default of states to federal
financing institutions (principally CEF). In both cases, the states' debt was rescheduled over a 20 year period
with a generous grace periods on payment of principal and a cap on debt service to current revenues

1
2
3

These transfers are appropriated at the source by the public bank (Banco de la Nación)..
There were very few cases in which the central govemment assisted s m provinces. The two or three cases took the f r of grants
o e
om
and to pay salaries in arrears.
The Provincial Structural Adjustment Loans negotiated by the Bank with selected provinces aimed at supporting the provincial
govemments to implement these adjustment programs.

Sesión 1 - 6 4

The third crisis was a consequence of the 1994 stabilization plan which brought down annual inflation
rates but led to high real interest rates. The major default was of state bonds. States were unable to liquidate
the bond debt (R$31 billion) and the federal government authorized the exchange of state bonds for federal
or central bank bonds. All state debt (principal and interest) was rolled over and debt service above the cap
was automatically capitalized. By 1996, the bond debt grew had reached $51.7 billion (5.4% of GDP).
Combined with the debt assumed by the failed state banks, at the end of 1996, the stock of state debt
assumed by the federal government totaled $123 billion (Dillinger, 1997).
During 1997-99, a more comprehensive approach to the restructuring of state debt was undertaken,
which involved sale of state assets, and commitments by states to adjustment programs, but the burden of the
ever growing state debt remained with the federal government. Dillinger (1998) estimates that Brazil's subnational government déficit due to unpaid interest obligations is more than half the overall public sector
déficit.
Since early 1999 — in the aftermath of the majorfinancial crisis which led to a major currency
devaluation — significant progress has been achieved towards strengthening the sub-national budget
constraint and improving commitment to fiscal targets. The government program includes (i) guaranteed
debt collection ; (ii) establishment of financing constraints; and (iii) a new fiscal code (Law of Fiscal
Responsibility recently introduced in the Senate) that will seek to further strengthen fiscal discipline at the
federal and sub-national level, including the mandatory establishment of fiscal targets and limits on debt. In
addition, a constitutional amendment limits personnel expenditures and introduces penalties on noncompliance. The Federal Government has cut most sources of financing to the states and has completed debt
rescheduling agreements with most states. These agreements require that the states service their debt
obligations with the federal government, which can now be enforced with recourse to constitutionally
mandated transfers as well as to the governments' own revenues.
3) Colombia
The 1991 Constitution devolved substantial expenditure responsibilities to Colombian municipalities
and departments. However, the decentralization process is rather limited. On the one hand, most of the
departments and municipalities lack the institutional capacity to effectively perform the assigned functions
(Terminassian, 1998). On the other hand, the central government determines norms for expenditures such as
education and health, further limiting decentralization. In addition, the role of own source revenues in subnationalfinancing is very small, and most sub-national revenues come from earmarked central government
transfers. In these circumstances theflexibility of the local governments to allocate expenditures is very
limited, a fact which constrains the accountability of these governments.
Colombia's case exemplifies how decentralization can contribute to macroeconomic imbalances.
Heavily dependent on transfers from the central governments, sub-national governments have contributed to
a deterioration of the overallfiscal position between 1991 and 1996. The largest source of growth in central
government administration expenditures has been current transfers, mainly to the local governments. This
reflects the importance that transfers have as a major source of revenue for the local governments and the
need to reform the system and improve the revenue raising ability of sub-national governments to increase
own-source revenues which will improve elected officials' accountability to their constituents.
In addition, given the lack of flexibility of the earmarked transfers, and the nationally determined use
of transfers, these transfers limit the effective decentralization of expenditures and encourage the use of
credit. On the one hand, banks have an incentive to lend based on the strength of constitutionally increasing
transfers from the nation and on the fact that local governments have earmarked resources that they are
unable to use. On the other hand, local government have inadequate incentives to raise own source revenues

Sesión 1 - 6 5

and have easy access to debtfinance. In fact, sub-national borrowing has increased rapidly - at more than
50% per year in nominal terms. During the same period the stock of local govemment debt doubled from 1
to 2 percent of GDP between 1993-1996.
The Ministry of Finance has taken several measures to curb the growth of sub-national debt:
•
•
•

Higher collateral requirements - in 1995, the amount of revenues that might be used as collateral
increased from 130% to 150% of the guaranteed loans;
Credit institution have to increase their ratio of net worth to loans to sub-national governments;
In 1997, the Law of Sub-national Debt established a system of waming signs based on stock and
flow indicators of sub-national governments and specified the central govemment entity that
must approve the requests. The indicators are: (i) interest payments as a percentage of operating
surplus; (ii) the stock of debt as a percentage of current revenue.

C.

Why do Local Governments Borrow?

Part of the responsibilities of local governments is to provide basic services. For that they will need to
invest in infrastructure. Pressure for investment are particularly heavy in urban areas undergoing rapid
migration. It is estimated that Asia alone will need to invest $280 billion a year over the next 30 years.4
Some of the investment will come from prívate sources; however publiclyfinanced infrastructure will
continué to be necessary, especially tofinance public goods. Traditionally, central governments would take
care of this type of investment. However, the move towards decentralization imply that subnational
governments will need to raise funds andfinance those outlays.
Tofinance these investments, local governments can use local taxation (e.g. property taxes) and/or
current savings. But these may be insufficient tofinance all infrastructure investments. A supplementary
source is debt. In principie, there is much to be said for financing infrastructure through debt. Roads,
schools, and pipelines have long useful lives, and debt spreads out costs of infrastructure to match its useful
life.
But where will they borrow? The experience of industrial countries suggests two basic approaches:
municipal bonds and municipal funds. Municipal bonds are often used in the USA and Cañada (sub-national
bond debt is about $7 billion in these two countries). Here, the markets are stable, the macroeconomic
situation has created conditions favoring long-run investment, capital markets are developed and the rules
and laws and known and enforced. In most developing countries, the conditions are quite the opposite. Long
periods of macroeconomic instability discourage long-run commitments; the legalframework is often faulty
in its capacity to protect investors, and information is not transparent. However, and in spite of these
problems, the markets for sub-national debt are expanding quickly. Some experiences have gone wrong (e.g.
Brazil and Istanbul) but in other countries are genuinely interested in building upfirm legislation to protect
investors on sub-national debt and favor a healthy subnational capital market.
Municipal funds and banks have been used in Europe. They were initially financed by the central
govemment and tried to substitute for the reluctance of prívate capital tofinance small municipalities5. More
recently, MDFs have developed in many developing countries. Under an MDF, the central govemment bears
the ultímate risk of municipal default. Some governments have insisted on prudent lending standards and
4
5

Brockman a d Williams 1996.
n
In their initial years, m n of t e (including Credit Local de France, t e Spanish Banco de Crédito Local, a d t e Brítish Public
ay
hm
h
n h
Works Loans Board)financed their operations through loansfrom t e central govemment. In effect, central g v r m n s used their
h
o en e t
strong credit standing to raise m n y cheaply in capital m r es a d on-lent the proceeds to municipalities through t e MDF
oe
ak t n
h
Sesión 1 - 6 6

strict debt service enforcement. Others have suffered high levels of default. Recently MDFs have been
encouraged to include prívate participation or even to be sold out to prívate interests6.
Bonds, banks and MDFs can operate simultaneously, as they do in the U.K. In all cases, the success
of the financial instrument depends on a stable macroeconomy, a legal framework defining the rights of
lenders and borrowers, and a supply of long-run savings. Central governments, specifically, will need to
adopt a legal framework for municipal borrowing, defining the rights and obligations of lenders and
creditors and a procedure for municipal bankruptcy.

D.

The Need for Controls

Controls and regulation over sub-national borrowing and debt stock are needed to avoid excessive
indebtedness at the local and national level. Some controls apply to the design of the inter-governmental
fiscal relationship trying to reduce the cases that may favor excessive dependence on the central government
or lack of clear accountability. These examine such factors as (i) correspondence between responsibilities
and tax raising power; (ii) the need to guarantee stable and predictable transfers from the center; and (iii) the
need to avoid concurrent responsibilities so that the lack of one level of government will not transíate in
additional effort from the other part of the government. Other type of controls focus directly on the total
amount of debt or borrowing, and are discussed below.
Regulatory Framework and Direct Controls: Regulation of sub-national borrowing is especially
needed in new democracies due to the lack of a clear boundary between national and local government
liability, causing sub-national defaults to be passed onto the national government. In addition, the national
government can have their credit ruined in international markets by defaults of local governments. To
address theserisks,central govemments may want to limit annual or accumulated déficit or stock of debt;
and/or impose controls on specific debt instruments. Table 2 summarizes the main controls used in Latin
American countries to control the borrowing by their sub-national governments. Some of these controls
include:
•
•

•

Prohibit Borrowing (Chile): this has the advantage of eliminating excessive borrowing but at the
cost of severely restricting the investment capacity of state and local governments, unless funds
are made available in some other fashion.
Impose restraints on a local government's déficit, or on total debt stock. Only Colombia has a
limit on total debt7. Most of the others either have no quantitative control on total debt or use the
debt service/revenue ratio to control outstanding debt.
Set limits on the ability of sub-national governments to use their future claim on
intergovemmental transfers as a guarantee for new loans. While these arrangements reduce the
cost of financing by transferring theriskto the national government, it limits the ability of the
market to signal an unsustainablefiscal position until the guarantee is exhausted.

Under Colombia's FINDETER program, prívate banks originate all municipal loans a d bear t e full riskof default, with t e
n
h
h
g v r m n acting merely as a second-tier b n to provide liquidity. The Czech Republic's MUFIS acts along similar lines6.
o en e t
ak
(Indeed, m n of t e European MDFs have shifted to maiket sources tofinance their operaüons and are in the process of
ay
h
privatization.)
7 Colombia is t e only c u t y in Latin America t a h s a regulatoryframework in place to control t e indebtedness capacity of t e
h
o nr
ht a
h
h
municipalities in accordance with their operating surplus. In 1 9 a law was passed to identify three levels of indebtedness k o n
97
nw
as t e street light s s e . This s s e establishes three levels of e e g n y by following two economic indicators: thefirst o e
h
yt m
yt m
mre c
n
m a u e liquidity a d t e second one m a u e solvency. The latter is only used for the "red level," or for highly indebted
es rs
n h
es rs
municipalities. The "red level" represents a danger in t e capacity to repay t e debt (Peiry a d Huertas, 1997).
h
h
n
6

Sesión 1 - 6 7

Install a proper framework to allow for project finance. Sub-national govemments should be able
to pledge the cash flow of a development project to secure finance. However, clear default rules
should be enacted so as to clarify propertyrightsand restrict the possibility of bailouts.
Refrain from bailing out sub-national govemments to convey the image of independence and the
fact that the market should be responsible for its own decisions when lending non-guaranteed
money to sub-national entities.
Table 2
LAC Countries - Controls on Sub-National Borrowing
Can use tax
Constraints
Borrowing must be
Debt numerical
sharing as
to use the
Country
authorized
constraints
guarantees?
debt
Argentina

Yes; by central Gov. in case
of foreign debt

Bolivia

By central and local

Brazil

by national and state

Chile

It is prohibited

Colombia

Domestic: by local legislation.
External: By national
legislature

Costa Rica

Domestic: by local legislation.
External: By national
legislature

Dominican
Rep
Ecuador
El Salvador
Guatemala
Honduras

México

Nicaragua
Panama

Domestic: by local legislation.
External: By national
legislature
No authorization required
No formal authorization is
required

Debt service <2025% revenues

Debt service <40%
current revenues;
debt/current rev
<80%
Debt service <10%
of municipal
revenues

Venezuela

Special law
Not authorized to borrow
abroad

For
investment

Yes

No

For
investment
External: for
investment

No

No
No

No

By national legislature

No

Debt service<20% of
revenues

Yes

Uruguay

Yes

No

Some states limit
debt service of
municipalities to 35%
of revenues

Perú

Yes

No

States and Municipalities are
not allowed to contract
external debt

No authorization for domestic
or external unless the
guarantee is required

Yes

No

No

National legislature
Domestic: by local legislation.
External: By national
legislature

For invest.
And reforms
for
investment
For project

None
Established annually

For

No
Yes for
municipalities

investment
For
investment
Investment

Determined in the
annual budget
No

No

No
No
No
No

No

No

Source: Latin America After a Decade of Reforms, Economic and Social Progress, 1997 Report, IDB

Sesión 1 - 6 8

Regulations are not always effective and can easily be violated. However, they are especially
important in the case of foreign borrowing. Foreign markets may not distinguish between the default of subnational governments and default of the central government. Control in this case would be required to
prevent local government defaults from undermining the central government's own external borrowing. In
addition, allowing local governments to have their own foreign debt can only complicate foreign exchange
management. For most of the other cases, the best regulation would be for the central government to see
local government borrowing as a prívate transaction between local government and the lender and refuse to
bail local government defaults.

II.

Summary and Lessons of Six Bond Offerings

This section briefly summarizes the six domestic and international bonds issued by sub-national
governments in Argentina, Brazil and Colombia. The cases were drawn from the universe of sub-national
bond issues sold in the Latín American domestic and international capital markets between 1991 and 1997,
as indicated in Appendix 7.
We selected the cases to represent large and small bond offerings in domestic and international capital
markets under different macroeconomic situations as described in Part I. The cases also form the basis of the
lessons learned about sub-national capital markets presented at the end of the report. Table 3 provides a
summary of the cases, which is followed by a short discussion of their salient features. Appendices 1 through
6 contain a more detailed description of each case.

Province/ City

Amount

Table 3
Summary Description of Sub-National Bond Issues
Purpose of the Issue
Market Issue Date Maturity Date

Buenos Aires
US 500 million in Refinance Citys Debt Stock
(City)
series
II
First
Issue US 250 million
»
Second
Issue LIT o.1 trillion
II
Third
Issue USD 150 million
t
i
Fourth
Issue LIT 69,000 million
Guaymallen
US 2 million
Public Works
(City)
Seríes
One
US 1 million
Gutters
Series
Two
US 1 million
Street pavement
Ibague (City)
Pesos 8 million
Civil Works
Seríes
One Pesos 4.8 million Road and Transport Infrast.
Series
Two Pesos 3.2 million Road and Transport Infrast.
Mendoza (Prv)
US 125 million
Refinance Province's Debt
Rio de Janeiro
US 125 million
Refinance Existing Debt
(City)
Valle del Cauca
Pesos 50 billion
Debt Restructuring
(Prv)

Sesión 1 - 69

Eurobond
4/11/97
5/23/97
5/28/97
6/10/97

4/11/07
5/23/07
5/28/04
1/6/00

12/1/96
7/1/98

8/1/99
7/1/99

Eurobond
Eurobond

7/1/96
7/1/97
8/2/96
7/12/96

7/1/01
7/1/02
7/25/02
7/12/99

Domestic

12/26/96

12/26/03

II

»
II

»

Domestic
II

M

Domestic
i
t
II

Collateral

Rating
B1 MoodVs / S&P BB-

Treasury Bonds
Gasoline Taxes

AAA S&P and domestic rating agency
A+ Duff & Phelps, Colombia

88% Oil Royalty

Interest Rate
Issue Price
Buenos Aires: Par, discount or premium
semi-annual at
11.25%
annual at 10%
annual at 10.5%
annual at 9.5%
Guaymallen: Par
annual at 10%
Ibague: Par
DTF + 3.5%
DTF + 2%
Mendoza: 100% Fixed
Fixed at 10% per
Re-Offer
annum
Rio de Janeiro: 99.96% Fixed, at 0.375% per
Fixed Re-Offer
annum
Valle del Cauca: Par

BB Duff & Phelps Int. / AA- Duff &
Phelps Ntl.
BB Duff & Phelps / B+ Standard &
Poor's
A by Duff & Phelps

Unsecured
Liquor Sales Tax

A.

The Cases

City of Buenos Aires. The City of Buenos Aires issued USD equivalent 500 million Euro Médium

Term Notes (EMTN) in the international debt markets. The EMTN program allows the issuer to offer the
notes with maturities from 30 days to 30 years in a variety of currencies. The city took advantage of good
market conditions and sold all the notes in four series within two months. The first series comprised a USD
250 million tranche with a 10-year maturity. The notes were sold via the 144a prívate placement market in
the U.S. at an interest rate of 11.25%, about 370 basis points above the 2006 U.S. Treasury Note (the
benchmark). The remaining series of notes were denominated in Argentine Pesos, USD, and Italian Lira
with maturities ranging from 3 to 10 years. The note proceeds were used to refinance the city's debt stock
and the liabilities of its bank, Banco de la Ciudad de Buenos Aires.

Municipality of Guaymallen, Argentina. The municipality sold two series of USD one million in

the local Mendoza market. The notes were sold in 1996 and matured in 1999, within the current mayor's
term. The municipality had planned to offer three series of the notes, but thefinance director did not want to
extend the note maturities beyond the term of the current mayor (1999). Therefore, the last one-million USD
series was canceled. The notes were secured by the central govemment revenue transfers (co-participation)
and were also collateralized by USD denominated Argentine treasury notes. The collateral gave the notes a
AAA rating and provided local investors with sufficient comfort to allow the underwriters to sell them
primarily to retail buyers in the Mendoza area. The note proceeds were used for public works improvements.

City of Ibague, Colombia. The City of Ibague sold two series of 5-year notes in 1996 and 1997
denominated in Colombian Pesos. The notes were secured with gasoline taxes collected by a special trust
created by the city. The trust collects the gasoline taxes and pays the debt service on the bonds. This trust
mechanism and a sufficient debt service coverage ratio (gasoline taxes to total debt service) allowed the
issue to receive an A+ domestic rating from Duff & Phelps, Colombia. The city offered thefirst series of
notes at an interest rate of 3.5% over the benchmark DTF which then declined to 2% over the DTF for the
second series. Thefirst series paid a higher rate because of the uncertainty associated with the viability of
the trust mechanism. This uncertainty was reduced for the second series as reflected in the lower interest
rate. Proceeds of the notes were used for road and infrastructure improvements.

Sesión 1 - 7 0

Mendoza Province, Argentina. Mendoza issued its inaugural international bond in August 1996 with
a USD 125 million six-year offering. Oil royalties paid by oil companies secured the bonds. Much of the due
diligence for the bond issue was focused on the legal issues for the oil royalty payments. These issues
included the terms and conditions of the oil concessions, the validity of oil company permits, the relevant
hydrocarbon laws, the province'srightsto the oil royalties and other revenues sources, as well as the validity
of the collateral documents and arrangements. The bonds were priced at 403 basis points over the 2002 U.S.
treasury at a fixed interest rate of 10% and sold to U.S. institutional investors in the 144(a) prívate market.
Proceeds from the bonds were used to refinance the province's existing debt.
The City of Rio de Janeiro. The puipose of Rio's USD 125 million three-year note issued in 1996
was to help the city refinance its existing short-term debt. The notes were a general obligation of the city and
were not secured by any dedicated revenues or collateral. Because of the general obligation structure, the
due diligence process required that the city present detailed information on its economic base, financial
operations, and management. The city'sfinancial statements had to be converted to international accounting
standards8. Investors showed much confidence in the city as they bought the bonds at a price that was only
25 basis points higher than the price of an equivalent sovereign bond. This relatively low spread to an
equivalent sovereign issue reflected the comfort investors had in the city's general financial condition. The
bonds were 2-3 times oversubscribed at an interest rate of 10.375% and a yield that was 403 basis points
over the benchmark 1999 U.S. Treasury note.
Department of Valle del Cauca, Colombia. The Valle de Cauca offering had a similar structure

as the Ibague bond issue: a dedicated tax revenue source to secure the bonds. However, in the Valle del
Cauca case the tax revenue was on liquor sales rather than gasoline sales. Similar to Ibague, a trust
mechanism was established to insure that the tax revenues were first used to pay debt service on the
bonds before passing them on to the department. Unlike Ibague, the Valle del Cauca bond received an A
rating, one notch below the rating of the Ibague issue. This might have been due to a different trust
mechanism. In the Valle del Cauca transaction, the department retained collection responsibilities
whereas the Ibague bond required the trust to collect and deposit tax revenues for debt service payments.
The 50 billion Colombian Peso seven-year Valle del Cauca bond issue was sold in part to refinance more
costly debtfrom Findeter, a govemment agency that funds municipal loans. The remaining bond proceeds
were used for capital improvements.

B.

Analysis, Lessons Learned and Recommendations for the Further
Development of Sub-national Capital Markets

This report now analyzes the six case studiesfrom several perspectives. First, it examines whether
they were unsecured general obligation bonds or bonds secured by specific revenues and why these
particular structures emerged. It then identifies the constraints for Latin American local govemments to
access to the domestic and international capital markets and provide some suggestions as how to alleviate
these constraints. The constraints include (i) lack of local govemment creditworthiness and poor institutional
capacity; (ii) lack of an adequate regulatory framework and supervisory authorities of the domestic capital
market and (iii) development of the domestic capital market. From these observations, a set of
recommendations is made.
8

If a sovereign or a sub-sovereign, as Rio de Janeiro,registert e bond issue in t e SEC, t e sub-sovereign has to fulñll schedule B.
h
h
h
This m a s t a t e sub-sovereign has not to change its accounting principies. ( Prívate corporations have to fulfill schedule A a d
en h t h
n
change their accounts to American accounting principies). The Eurobond m r e requisites are less restritives t a those applied by
ak t
ht
t e SEC. As a consequence, Rio de Janeiro had to change its a c u t to international accounting principies, because t e accounting
h
c o ns
h
plan was not good and t e financial advisors asked for t e change.
h
h

Sesión 1 - 7 1

Box 1
The importance of Marketing and Management
The importance of Marketing: When an issuer is raising funds in the capital markets,
he/she is simultaneously selling a product. In a local government the product is the economy
of the city and the financial creditworthiness of the local government. Deutsche Bank, IBM or
Walt Disney can have easier access to the capital markets than Banco de la Pequeña y
Mediana Empresa, Computadores Reunidos or Moya Films. This doesn't mean that the last
three companíes can not access the capital markets, but they need to do marketing. In the
financial markets this means road-shows, presentations, bilateral meetings and good and
timely urban economic and financial information.
Budget: Size means to be frequent in the capital markets and frequency means liquidity
in the secondary market. This is very important in public issues and, in lesser extent, in
prívate placements. To access to the American market the amount of the issue has to be, at
least, US$ 100 million. The average is above US$ 200 million. Prívate placements in the
European market can be smaller, but, al least, Euro 25 millions.
As a consequence of marketing and size only the most important cities and regional
governments of l_AC can access the international capital market through general obligation
issues.
The example of the Base Country is well-known, but for a bad political situation. In the first
half of the 90's the Base Country has raised successfully public issues in the European,
Japanese and American markets. The main goal was to introduce its ñame in the World as a
peaceful country. For this reason, these issues meant the possibility of road-shows,
presentations, bilateral meetings, media, etc., that permits the Base Government to talk
about the good situation of the Base economy, that the country has good infrastructures,
skilled personnel and so on. ( the Base Country has approximately 2.4 millions of
inhabitants)

1)

International Bond Structures

Bonds can either be secured by a dedicated revenue source or another form of collateral, or they can
be an unsecured "general obligation" with payment of debt service based on the "full faith and credit" of the
issuer. Whether or not some form of dedicated revenue source or other type of collateral to secure the bonds
was necessary for the issuers in these case studies depended on the recognition of the issuer's ñame by
investors and the perceived importance of the city within the country. It appears that many international
emerging market investors develop their view of a sub-national entity primarily by their assessment of the
sovereign's credit rating and their general knowledge of the sub-national issuer.
Ñame recognition is also directly related to investor perceptions about moral hazard - the willingness
of the sovereign to bail out the sub-sovereign when a bond defaults. To gauge the moral hazard, investors
look at the sovereign's behavior regarding bailouts of sub-sovereign domestic and international obligations.
Fear that the sub-sovereign will contaminate the ñame of the sovereign in the capital markets drives the latter
to bail out sub-sovereign deadbeats. This is detrimental to sub-nationalfiscal discipline, and is a scenario
that is much more likely to occur with larger well-known municipalities. The smaller sub-national issuers are
politically andfinancially less important and cannot always rely on bailouts. However, from a bondholders
point of view, the lack of additional security in the form of a dedicated revenue source or some other
collateral is offset by investor perception that the sovereign will bail out the municipality in case it gets into
trouble and cannot make its debt service payments. The case studies seem to demónstrate, although it is
difficult to prove, that an implied government bailout was priced into the international bond issues for the
'general obligation" bond structures.
Sesión 1 - 72

Box 2
Moral hazard or "big brother"
Baíl outs or moral hazard are known in the capital markets as the "big brother"The creditworthiness of the City of Barcelona is a consequence of the economy of
the city and the financial performance and management, but also as a result of its big
brother, the Kingdom of Spain-. This effect, more than the ñame, is related to the
importance of the local government in the economy of the country,. Size is another
time the word, but the big brother doesn't mean "the father". The investors know that
there is not an automatic bailout, but wait some kind of support from the National
Government. For example, when Spain began to be member of the Euro, Mood/s
upgrade Spain from AA2 to AAA, as Germany and France, because they think that
Spain will have some implicit bailout, if it is needed.

In addition to the issue of ñame recognition, the underwriters of the case study offerings had to take
into account the issuer's desire for future bond sales to determine the appropriate bond structure. In the case
of Buenos Aires, its goal was to diversify its borrowing sources as general liquidity in the market place
began to tighten as a result of the Tequila crisis. Restructuring its outstanding short-term debt was also an
important consideration, but the need tofind additionalftmding sources was paramount.
The primary lesson learned from the case studies is that cities of LAC can access the international
capital markets with general obligation type of bonds if:
(a) they have a credit rating of international independent agencies,
(b) they present good economic information of the city and timely financial information of the local
government through road-shows, advised by financial intermediaries with good international
reputation, and
(c) the budget size of the city permits to consider that the city will be present in the market with
some frequency.
(d) In another case, they can access the capital markets through prívate placements and revenue
bonds".
2)

Domestic Bond Structures

The changefrom general obligation to revenue type bonds has contributed to the development of the
local, but undeveloped bond market. In addition, the newly created purchasers of local bonds such as
pensión funds, mutual funds and other institutional investors are more sophisticated and are aware of the
history of local government general obligation debt and the political influences that control repayment of the
debt. They now require additional security to purchase bonds. These factors have led to the further evolution
of "dedicated revenue" types of structures as demonstrated in the Ibague and Valle del Cauca issues in
Colombia (gasoline and liquor taxes) and in Guaymallen in Argentina (government treasury bilis and coparticipation revenues). We expect that dedicated revenue funds will continué to be the primary bond
structure for local bond market issues in the near future, especially small and médium size municipalities.
This will most likely be the case as central governments increasingly shift from a paternalistic financial and
political policy to one of decentralization and localfiscal control.
We expect that the dedicated fund type of revenue bond structure will continué to evolve. As
decentralization proceeds, local government should be given new taxing powers and the ability to raise
revenues for specific purposes. These new sources, such as a gasoline tax, can be used to secure bonds

Sesión 1 - 73

issued for a specific improvement, like road pavement in the Ibague case. Taxpayers are more willing to
absorb the additional cost for gasoline if they can see the direct benefit of their tax payments. This rationale
can also be used for other types of taxes that local govemment may be authorized to levy as the process of
decentralization evolves. They will require new revenue sources to fulfill service mandates imposed by
decentralization. These revenues can then be used to secure bond issues and legal structures developed to
insure that an independent third party collects the revenue and distributes it to the bondholders.
3)

Constraints to the Development of Sub-national Credit Markets

The following analysis of the constraints in sub-national credit markets allows us to summarize the
findings of the case studies, as well as make recommendations towards their future development.

Credit Ratinas and Credit Rating Agencies
Credit ratings have become an essential element of international bond issues and are used more
frequently in domestic bond offerings. The case studies indicated that credit ratings were an integral part of
the bond offering process. We expect that credit ratings will continué to play a central role in the
development of sub-national municipal markets for both domestic and international bond issues. However,
several problems related to international and domestic credit ratings were apparent in the case studies.
International ratings required a more stringent set of rating criteria than the domestic ratings. This is
probably due to the fact that international investors have used credit ratings for several decades to
complement their own credit risk analysis and have clear expectations as to the depth and quality of the
rating process. Domestic ratings, on the other hand, were done for a variety of purposes. In Colombia
ratings are necessary for regulatory purposes and in Argentina as a disclosure gesture to inform investors
about thefinancial condition of the municipality. There are a wide variety of domestic credit rating agencies
that use different methodologies for their credit ratings. A standard has not developed, as one generally finds
in international ratings, but the latter also has significant differences in analytical approaches between rating
agency companies.
One of thefindings of the case studies that demónstrate the importance of the rating agencies in the
development of the domestic capital market was the Valle del Cauca issue that used dedicated liquor taxes to
secure the bonds. The details of the structure were developed jointly by thefinancial advisor, underwriter
and the rating agency. They defined the most appropriate mechanisms for revenue collection and distribution
to bond holders. In the industrialized economies rating agencies frequently dictate specific bond covenants
and terms and conditions for the issuer to achieve a certain rating goal. In the case of Valle del Cauca, this
method was applied and we expect to see it occur more frequently.9 However, unless relatively standard
criteria is developed among the rating agencies to achieve certain credit ratings, this process can send
confusing signáis to the domestic market. It may be worthwhile to establish a self-regulatoiy mechanism
among the prívate rating agencies to develop a set of rating guidelines for municipal bond issues, although
competitive business pressures may prevent cooperation among these companies.
Ideally, the credit rating process for international and domestic bond issues should be the same except
for the obvious issues associated with sovereign related risks such as foreign exchange, currency
convertibility and legal matters associated with particular bond structure. Analysis of the municipality's
9

The international rating agencies try to assess objectively t e local g v m e t creditworthiness, but as all process of evaluation,
h
oe mn
t e apply a lot of subjective appraisal. They try to solve t e problem through evaluation meetings t a participate a t a of
hy
h
ht
em
different experts

Sesión 1 - 74

financial operations (financial statements, budget, liquidity, operating balance, etc.), administration,
economic base, debt management and political factors should use the same general obligation debt criteria
for all issuers. Financial criteria for dedicated revenue debt should also be the same (e.g. tax forecasts, debt
service coverage ratios, etc.). However, even among the international credit rating agencies, there is a
difference in methodology in assigning a rating because of a different emphasis on the certain factors. Some
factors may have more weight in the rating than others. Nevertheless, a self-regulatory body or a government
regulatory agency could prepare credit rating guidelines for international and domestic bond sales.
The case studies indicated that credit ratings in the domestic market appear to be more of a regulatory
requirement than a service requested by investors, although this is beginning to change, as ratings become
more common. The structure of the bond issue also plays a role in whether or not a rating is necessary, as
does the issuer's perception as to whether the cost of the rating is needed to place the bonds.
One of the factors that distinguished the international credit ratings from the domestic ratings
appeared to be the quality of the information provided by the issuer to the rating agencies, especially
financial statements. The underwriters in the international bond issues were concemed about the quality of
the financial statements and in one case had them audited by an extemal accountingfirm using international
accounting standards. In another case the financial statements were converted to international standards as
part of the due diligence process. This did not appear to be the case for the domestic ratings. The due
diligence process for the international issues required that the underwriters restate thefinancial statements
using international accounting standards. This was a costly, lengthy process and could have been used by the
municipality to initiate a change in its accounting standards. But this opportunity was lost as neither the
underwriter ñor the issuer had any incentive to change its accounting standard, especially if the resulting
standard did not conform to legal auditing requirements. Developing appropriate accounting standards is a
major issue that has to be tackled at the national level so those municipalities have an incentive to change.
National Governments should develop a Public Administration Accounting Plan. The local governments that
accomplish this plan can access easier the domestic market and probably the international market as well.

Institutional Development and Capacity
An important factor contributing to the success of the case study bond issues was the level of staff
development and their experience in municipalfinancial management. Most staff members did not have any
experience with bond issues and recognized the need for external advisors and underwriters to assist in the
process. However, the way in which the external advisors were used depended on the ability of the
municipal administration to understand the details involved in preparing for the offering: gathering financial
statements and statistics; rating agency presentation and strategy; regulatory applications and permits;
preparation of the offering circular, etc. The most time consuming activity was organizing financial
statements andfinancial performance information and presenting it in a format acceptable to the rating
agencies. This was especially difñcult for the international bond issues as the staff had tofirst understand the
need for presentingfinancial statements using international accounting standards and then provide the
support to the underwriters to complete the conversión. Preparation for domestic issues was simpler, but still
required, in some cases, a reorganization of the municipal accounts to even produce afinancial statement.
The lack of adequatefinancial accounting and reporting systems is a major impediment to the efficient
development of both the international and domestic bond markets. A good supervisory authority (e.g.
Securities and Exchange Commission) that could publish guidelines of the legal andfinancial information
that the local government has to register, assuring transparency.
Another area that will require substantial effort to facilítate development of the sub-national capital
market is the need for disclosure and due diligence standards for offering circular documents. There is a

Sesión 1 - 75

huge gap between the quality of the offering circulare used in the international issues and the disclosure
provided for domestic investors. A self-regulatory body should develop some basic standards for domestic
capital market issues that follow the standards used for international issues. The most important guideline is
the need to prepare adequatefinancial statements, as discussed above. The guidelines should take into
account the cost of securing the information to insure that small and médium size municipalities are not
prohibited from entering the market because of excessive bond preparation costs. The World Bank may
consider a grant or loan fiind for a country to support the organization of a self-regulatory body and the
preparation of disclosure guidelines for the municipal bond market. It may also consider a special fund to
provide assistance to municipalities to prepare disclosure documents for municipal bond issues. This effort
can be started with a pilot program such as the "city assistance" strategy.

Development of a Professional "Infrastructure" for Domestic Bond Issues
Both the international and domestic bond issues had a variety of external advisors and professionals to
assist in the preparation and placement of the domestic and international offerings. They managed the due
diligence process; prepared the rating agency strategy and presentation and incorporated changes to the bond
structure in response to rating agency comments. Bond market professionals also prepared the offering
circular; developed the pre-sales strategy; and negotiated the bond price with the issuer and investors. The
municipal staff gained valuable experience in following the details of a bond offering
Although the professionals were hired only to assist the municipalities in the bond offering process,
the staff lost the opportunity to make general improvements in theirfinancial management systems that
would accelerate the development of the municipal markets. Municipalities may consider expanding the role
of the external advisors to prepare a checklist of items the municipality should improve based on the
disclosure and accounting guidelines developed by a national bond professional association or selfregulatory body, as proposed above. Ideally, however, the checklist should be made by regulatory and
supervisory authority, who are political independent and can serve the general purpose of improving the
capacity of the sub-national govemment.
The pricing of domestic offerings has to be based on competitive bidding, with a supervisory authority
taking care of transparency of the process and that insight information has not occurred. If the domestic
market is incipient, then a self-regulatory body could provide valuable assistance is in the pricing of
domestic offerings. Due to limited bond volume, a small secondary market, lack of regulations regarding
bond underwriters' guidelines and a variety of other factors, pricing of domestic bond issues in small
markets is not very transparent. Municipalities do not know if they are receiving a fair price for their bonds
(this is also the case for international offerings). For example, in one of the domestic issues with two
tranches, the period to sell thefirst tranche was rather extended indicating a need to increase the price to
attract investors. The second tranche sold off to a single investor in a short period indicating that the price
might have been too cheap for the investor (interest rate too high and/or bond discount too great). The
second tranche probably sold much faster because the bond security (a dedicated revenue source) was
understood by the market and it was willing to purchase the bonds at a higher price the second time around.
In an international issue the initial offering price was two to three times oversubscribed indicating strong
demand and/or a low price with a spread to the equivalent U.S. treasury security of about 400 basis points.
The spread declined to 225 basis points in the secondary market demonstrating either a bullish market for
Emerging market debt and/or a high initial price; probably a combination the two factors. In both of these
cases the issuer had no way of knowing whether they received a fair price because of the lack of pricing
data. To obtain better information about pricing, a self-regulatory body could help establish rules for the
dissemination of municipal bond prices in the primary and secondary municipal bond market. Municipal

Sesión 1 - 7 6

officials can also hire independent financial advisors to negotiate the pricing of bond issues with
underwriters.
The case studies also pointed to another difficult issue regarding professional services: How do the
municipalities know what professional assistance is necessary for the bond sale and the price they should
pay for that assistance? Most of the professionals hired for the case study bond issues were retained on a
negotiated basis10. Municipalities should consider the competitive selection of bond professionals. This
makes the selection process more transparent and also provides the municipality with competitive pricing for
professional bond services. . In the selection of financial advisors, underwriters, etc., it is very important to
take into account their reputation in the capital markets. The credibility of the local government improves if
a goodfinancial advisor and/ or underwriter works with the local government. When a local government has
a good reputation, the process can be the opposite. A financial advisor wants to work with this local
government, because it increases its reputation.
Municipalities may also consider hiring afinancial advisor prior to selecting a bond underwriter, as
was the instance in some of the case studies. Thefinancial advisor should provide the city with a long-term
debt management strategy and an initial bond structure based on that strategy. The advisor can also do a
significant amount of preparatory work for the bond issue. This includes organizing thefinancial statements
and other documents to submit to the rating agencies; obtain regulatory approvals for the bond issue; and
prepare terms of reference and a tender document for the competitive selection of the underwriter.

Regulation of the Sub-national Capital Market
Part I of the paper discusses the need for greaterfiscal regulation of sub-national bond issues. The
focus of this regulation is to ensure that sub-national governments arefiscally sound and do not threaten a
country's macroeconomic stability. This type of regulation is more closely associated with improving the
financial condition of municipal government before the central government allows it to issue bonds in either
the domestic or international markets. The basic premise is that municipal borrowing should not occur if it
creates an additionalfinancial burden on the central government. The best way to avoid a negative impact on
the national government is to require limits on total outstanding debt, as is the case in Colombia. Part I of
this paper also discusses other options to maintain thefinancial health of municipalities.
Regulation can also include controls on the parties involved in bond transactions. The controls should
insure that taxpayer revenues are used properly and municipalities are paying a fair price for their bond
issues. This type of regulation was briefly discussed above in regards to guidelines for due diligence and
disclosure documents, accounting standards, and bond pricing.
There are several options for developing bond transaction regulations and regulatory bodies. One
option is to establish a self-regulatory body like the Municipal Securities Rule Making Board (MSRB) that
regulates the municipal bond market in the U.S. This non-government entity is comprised of representatives
from the municipal bond industry (primarily underwriters), public citizens and government agency
representatives. It provides guidelines for the preparation, sale, disclosure and secondary trading of
municipal bonds. Another non-government association, The Government Accounting Standards Board
(GASB), promulgates accounting standards. However, neither the MSRB ñor the GASB have the legal
authority to mandate changes in the management of the municipal bond market. This remains with the
Securities and Exchange Commission, a government agency that regulates the broader securities market.

10

S l c i n of u d r r t r f r itrainl i s e was done on a " o p t t v - e o i t d b s s , w t p i e a t e p i a y slcin
eeto
n e w i e s o nentoa s u s
c m e i i e n g t a e a i " i h r c s h r m r eeto
etn.
nea
Sesión 1 - 77

Another option is to delegate self-regulatory activities to other players in the nascent domestic
municipal market - commercial banks, non-government organizations, trust organizations - whichever
entities represent the interests of the parties involved in the market. International Emerging markets have
self-regulatory bodies, such as the International Secundes Dealer Association to maintain standards for
Eurobond issues.
It may be premature to establish a self-regulatory body for the municipal market until the market
reaches a certain volume of transactions and/or the central govemment establishes debt limits for subnational governments to maintain macroeconomic stability. Nevertheless, the World Bank and other national
and multilateral organizations interested in the further development of sub-national capital markets should
begin to consider how to regúlate municipal bond transactions to avoid potential corruption, insider trading
and other negative behaviors that can severely damage an incipient municipal market.

Price Volatility. Market Demand and Investor Perceptions
The extent and duration of price volatility can sometimes be used as an indicator of the maturity of a
market (breadth and depth of the primary market and volume of secondary trading) and the overall credit
quality of the market participants. Price volatility in the international and domestic capital market for sub
sovereign debt has many causes. The underlying cause is the lower credit quality of the issuers (after
controlling for macroeconomic conditions) and narrow secondary markets. Uncertainty about the ability of
issuers to make debt service payments combined with political uncertainty about their willingness to pay is
the basis of the volatile conditions. This general condition applies to both the international and domestic
capital markets.
The case studies demonstrated the volatility of both the international and domestic markets, but they
each have their unique cultural characteristics. In the international emerging markets, bond prices of debt
from the developing countries seem to move in blocks for regions and the market as a whole; they generally
go up with good news such as progress onfiscal and monetary reforms; reduced inflation; currency stability;
declining budget déficits, etc. Prices go down as a block on bad economic and political news: currency
instability and potential for devaluation; deteriorating current account balances; anticipated budget déficits;
changes in critical cabinet ministers, etc. International investors have begun to discrimínate between
countries and reward good economic policy with higher bond prices. But investors seem to move in a herd
and punish most countries with lower prices during a crisis, such as the recent turmoil created by the
problems in Asia. Emerging market investors have been recently criticized by officials in Latin America for
failing to distinguish between emerging market regions, such as Asia and Latin America, not to mention the
lack of discrimination within a región.
Extreme volatility in the international capital markets will continué for some time until the developing
countries improve their credit condition and achieve solid investment grade ratings, in the A to A+ range.
Until that time the markets will continué to move more on changes in political uncertainty, such as the
election or appointment of govemment officials that change or reverse appropriate economic policy. They
will also remain volatile until investors feel confident that economic policy will improve credit conditions
measured by indicators such as General Government Balance/GDP, General Government Debt/GDP, Net
Public Debt/Exports, etc.
As mentioned earlier, ñame recognition and the perceived credit condition of the sovereign issuers
plays a significant role in the pricing of sub-national debt. Sub-sovereign debt will follow the lead of the
sovereign in the capital markets, as demonstrated by the case studies. When the volatility of sovereign bonds
subsides, investors can begin to discrimínate between countries with higher or lower credit ratings rather
than indiscriminately punishing all countries when a severe problem arises in one country. Once this

Sesión 1 - 78

sovereign discrimination takes place, investors can then differentiate between sub-sovereign entities with
different credit ratings. However, this will most likely not occur for some time. High volatility in the
sovereign market will remain until the major economies in a región achieve stable investment grade ratings.
In the meanwhile only internationally recognized cities and regions will be able to access the capital markets
on an unsecured or general obligation basis, with some occasional exceptions.
The case studies demonstrated that ñame recognition, size and management were the most important
factors in determining the price of the bond issues (after taking into account the sovereign credit rating).
Even with good financial indicators, a sub-sovereign entity that is not recognized by investors will pay a
higher price and may not be able to access the markets at all unless it uses a dedicated revenue source to
secure the bonds. These factors will act as natural obstacles for sub-sovereign govemments seeking capital
in the international capital markets and will most likely cióse the market for médium and small size issuers
in Latin America.
Market demand and pricing in the domestic markets has a different cultural base than the international
capital markets. The case studies illustrate that investors are skeptical about the willingness of elected
officials to repay debt obligations, not to mention the ability of these entities to service their debt. They are
also skeptical of newly elected officials to honor the obligations of a previous administration. In one case,
the third tranche of a bond issue was canceled because the current administration's term was completed prior
to the repayment of bond principal. Thefinance officer was concemed that the new administration would not
honor the obligation of the previous administration and he decided to cancel the third tranche. This occurred
in a municipality with a good financial record.
Most of the domestic investors who are familiar with the operations of sub-sovereign govemment rely
more on their feelings of trust in the entity's willingness to pay rather than their ability to pay, especially in
the médium and small size municipalities. This occurs for several reasons. First, it is very difficult to
accurately determine thefinancial status of small and médium size municipalities. The accounting systems
used by these local govemments usually are not capable of forecasting the municipality'sfiiture cash
posilion. The entities are managed by crisis asfinancial officers scramble tofind the funds to meet critical
payments such as payroll and essential vendors. This circumstance requires investors to demand bond
structures that isolate the revenues used for debt service from the local govemment administration. This has
resulted in the "dedicated revenue" bond structure described in thefirst part of this section. We expect that
dedicated bond structures will be the most common form of bond issued in the domestic capital markets until
investors feel secure that the culture of municipal political andfinancial management changes. These
changes are slowly occurring as local officials are held more accountable to voters and financial
management becomes an important political issue.

III.

A Road Map to Understanding Sub-national Capital Markets

The case studies, although not representative of the general capital market universe, provide some
interesting lessons and confirm some of the key factors that have to be in place before one thinks about local
access to capital markets. Sub-national bond markets can be an important source of credit as is the case in
the US and Cañada. However, some basic conditions are required as a pre-requisite for successful
development of a local govemment capital market:

>• Macroeconomic balances and Regulatory Framework - The macroeconomic situation should
be stable - macroeconomic instability increases theriskof long-term commitments. One of the reasons for

Sesión 1 - 79

the long established tradition of municipal bonds in the USA and Cañada is the stable macroeconomic
situation which has led to an important component of long-term savings and investments.
•
•
•
•

Subnational borrowing should be clearly inline with the macroeconomic requirements. This may
require strict limits in sub-national debt, including banks, bond issues, municipal development
funds, etc..
The regulatory framework should be transparent and enforceable and adapted to the
circumstances of the local government. Access to international markets should be disciplined and
the impact of exchange rateriskclearly evaluated.
Access to domestic market should be promoted and regulated in parallel with the development of
long-run sources of savings.
Inter-govemmentalfiscal relations should be predictable, stable and leading to a stable flow of
revenues to the subnational governments. (The case of Vietnam exemplifies the problems
associated with lack of a clear inter-governmentalfiscal framework: see Box 3).

>

Developed Internal Capital Markets: Before any sub-national capital markets can be created,

•

Long-term funds should be available; financial markets should be deep to facilítate the supply of
savings through formal networks

the country capital market needs to be in place and fúnctioning. This requires six conditions

Market regulatory and supervisory system: The regulations and the authority that supervises all the
participants of the market must be implemented. Two well known and successful models are the American
system , based on a complete and exhaustive regulation and, the other hand, the London system, based on
contracts and agreements, supported by a group of good lawyers in the City, but both system has in common
the judicial enforceability of the contracts
Box 3
Intergovernmental Relationships in Vietnam
The Vietnamese system of intergovernmental fiscal relationships has been in flux over
the past 10 years. A new budget law has been established by the central government on
March, 1996 and amended on May, 1998. Before the law, assignment of revenues to the
various levels of government changed from year to year; expenditures were guided by
negotiation and political influences. The 1996 law established some structure. It defined
the revenue authority and expenditure responsibilities of the various levels of government
administration. A substantial vertical gap remains to be filled but no decisions have been
taken on how to transfer new tax raising power to local governments. In addition, the
implementation of the new budget law appears to exhibit some of the discretionary
elements of the previous arrangements, at least in some of the "provincial cities. Data are
poor and indicators vary according to the source.
Ho Chi Minh issued a city bond in 1996. It was the first and the last of sub-national
access to markets in Vietnam. In 1995, Ho Chi Minh was the first city to access the capital
markets in Vietnam. It placed a revenue bond to finance a toll road. The bond had a
maturity period of 3 years and was placed with an interest rate of 17% p.a. The Bank for
Investment and Development of Vietnam (BIDV) in HCMC was part of the team
preparing the term of the bond and its issuance procedures. The BIDV was also the
biggest buyer of the bond. The bond had some particular characteristics: Coupon fixed at
inflation adjusted figure. The buyers were banks and public companies. There was no
credit rating and no underwriters - the bond was bought at the Peoples' Committee
Agency

Sesión 1 -

0

•

Skilled personnel: Skilled and pragmatic personnel is needed in the supervisory authority,
issuers, brokers and portfolio managers. A special consideration has to be made in the training of
financial advisers, owing to thefinancial knowledge is not broadly diffiised among prívate
investors. Investors should be familiar with the system; there is a need for a secondaiy market so
that investors can sell their assets prior to maturity (increasing the liquidity of otherwise illiquid
assets

•

Market organization: The market can be, physically, a place where brokers meet to buy and sell
securities, according to the regulations of this market. The future is an information system that
integrates the clearing system, the spot market and the derivative market. The software has not to
be developed. For example, the Swedish and the Australian stock exchanges use a software
developed by a Swedish company, the Spanish stock exchange use the software of the Toronto
stock exchange. The main problem is the Communications network, and the implementation of
the information system depends of this infrastructure.

•

Rating agencies and bond insurers should be encouraged to play their role informing the public
of theriskof the operations.

•

Market size: The most important factor to develop a sub-national market is to develop the
national market, helped by regulations, and also supervisión, that permits the development of
insurance companies, mutual funds, pensión funds, etc., and in general, institutional investors as
a way that permits the transformation of savings in long-term funds. The critical points in this
process of creating long-run savings structures are the following:
a) When the mutual funds, the pensión funds, or other institutional entity begins to be
implemented successfully, one may face ariskof excess demand for securities, price increase
and more incentives for additional offerings. This situation together with lack of skilled
personnel may bring serious problems for the market.
b) The development of a capital market needs the support and collaboration of the domestic
banking system. The international banks, that know this business and compete with the domestic
banks, can help accelerate the development of the capital market and, in particular, the process of
collaboration of the domestic banks.
c) The National Treasury has a leading role in the development of the domestic capital market.
Not only it sets the regulatory framework, but is own behavior in capital markets sets the
benchmarks against which the prices of other issues are fixed.
d) The national government may be in a situation of competing with local governments to issue
bonds and finance their déficits. This may lead to a conflict of interest, especially, if the market is
narrow. A regulatory framework well defined and transparent is essential to prevent
discriminatory positions of the central government.

• Liquidity: A sub-sovereign can rise founds a good prices in the capital markets if its issues offer
liquidity. The problem can not be resolved only with the implementation of market makers, t
depends of the budget size, size means to be frequent in the market and this helps that the issues
have liquidity. In another case, the altemative , usually more expensive, because a premium is paid
for not being liquid, is prívate placements. The prívate placements, on the other hand, introduce
the sub-sovereign in the market and prepare it for future public issues.
> Local Governments should have a goodfiscal and management system and capacity to generate
consistent and strong revenues. They can. improve their attractiveness to borrowers by
• instituting accounting, auditing, and disclosure standards that are compatible with international
standards.

Sesión 1 - 1

• improving the quality of their collateral by allowing central governments to deduct debt service
directly from intergovernmental transfers or pledging a specific tax or other revenue source to
service debt on a specific loan.
• Making sure that loan contracts can specify that the payment of debt service on a specific loan will
be given contractual priority—or prohibit a municipalityfrom issuing additional bonds backed by
the same revenue source until a specific debt is retired.
In many countries with recent/shallow domestic capital markets, those conditions may not be present.
Thefirst problem is the quality of municipal or sub-national management and accounting, which is often
quite poor and incomplete. The second problem is the shortage of long term savings, as decades of
hyperinflation have destroyed any habits of regular savings. Regulation and economic policy can also be an
impediment. Examples are controls over interest rates or impediments against the development of
institutional sources of savings such as pensión funds, and insurance companies. Finally the legal system
may not be prepared to deal with default, and local governments may not have the ability or the authority to
raise taxes, as these fall under the central govemment control.
Over the long term, national governments should address constraints on the domestic capital market.
To mobilize prívate long-term savings, there is a need to eliminate interest rate caps; remove regulations that
restrict the growth of insurance companies, mutual funds, and pensión funds; improve the quality of
information in the market by enabling regulatory authorities to evalúate thefinancial health of institutions;
and facilitate the development of prívate rating agencies; and strengthen legalframeworks. Finally, there is a
need to make local governments more attractive to potential lenders. Central governments can strengthen a
local govemment's revenue position by assigning broad tax bases to local governments andfreeing local
governments to determine tax rates. Local governments can, in turn, render themselves more attractive to
bondholders by using this power prudently - for example, by keeping tax rates low they leave headroom for
increasing rates if necessary.

Sesión 1 -

References
Ahmad, Ehtishaxn and Baer, Katherine (1997) "Colombia" in Teresa Ter-Minassian ed. "Fiscal
Federaliism: in Theory and Practice" IMF, Washington 1997
Bird, Richard, R. Ebel and C. Wallich (1995). "Fiscal Decentralization: From Command to market,",
in R. M. Bird, R.Ebel, and C. Wallich (eds.), Decentralization of the Socialist State:
Intergovernmental Finance in Transition Economies (Washington, DC., The World Bank), 1-68
Campbell, Tim, forthcoming, 'The Quiet Revolution", The World Bank.
Dillinger, William, 1997 "Brazil's State Debt Crisis: Lessons Learned." Economic Notes #14. The
World Bank, Latin American and the Caribbean Región, Washington DC, processed.
Dillinger, William and Steven Webb, 1998. "Fiscal Management in Federal Democracies: Argentina
and Brazil." The World Bank, Latin American and the Caribbean Región. Washington DC,
processed
Fornasari, Francesca, Steven Webb, Heng-Fu Zou (1998) "Decentralized Spending and Central
Government Déficits: International Evidence", Mimeo World Bank.
Inter American Bank, Latin America After a Decade ofReforms, Economic and Social Progress, 1997
Report
Litvak, Ahmad and R. Bird " Rethinking Decentralization at the World Bank". World Bank, 1998
Peterson, George Decentralization in Latin America, Learning Trough Experience, The World bank,
1997
Perry and Huertas "La Historia de Una Crisis Anunciada: Regulando el Endeudamiento de las
Municipalidades y los Departamentos en Colombia" In CEPAL, August 1997 Descentralización
Fiscal en América Latina: Nuevos Desafíos y Agenda de Trabajo.
Petterson and Shaman, " Building Local Credit Systems", World Bank 1997
Prud'homme, Remy (1995), 'The Dangers of Decentralization," The World Bank Research Observer
10, 201-220

Stein, Ernesto (1997) "Fiscal Federalism and Decentralization: A Review of Some Efficiency and
Macroeconomic Aspects." In Annual World Bank Conference on Development Economics, 1995.
The World Bank, Washington DC.
Tanzi, Vito (1996) "Fiscal Federalism and Decentralization: A Review of Some Efficiency and
Macroeconomic Aspects," in M. Bruno and B. Pleskovic (eds.), Annual World Bank Conference on
Development Economics (Washington DC: The World Bank) 295-316.
Teresa Ter-Minassian (1997) " Fiscal Federalism: in Theory and Practice" IMF, Washington 1997
World Bank, 1995, Brazil State Debt: Crisis and Reform Report (Washington, June)
World Bank, 1996, Argentina: Provincial Finance, Issues of Fiscal Federalism (Washington, June)
ong, unblemished credit history.11

11

Petersen a d H m a 1 9 ; Dailami a d Leipziger 1 9 .
n a mm 9 7
n
98
Sesión 1 -

Appendix 1
Ciudad de Buenos Aires, Argentina
I. Description of the Bond
Ciudad de Buenos Aires - USD 500 million Euro Medium-Term Note
Programme, Individual Notes Due From 30 days to 30 yearsfrom Date of Issue
Details

Feature
Issuer
Arranger
Dealer
Currency
Amount
Maturities
Issue Price
Method of Issue
Interest Rate
Fixed Rate Notes
Variable Rate Notes
Interest Periods
Zero Coupon Notes
Withholding Tax
Denominations
Optimal Redemption
Listing
Status
Negative Pledge
Covenants
Cross Default
Governing Law
Selling Restrictions
Transfer Restrictions
Clearing System
Pricing Supplement
Use of Proceeds

City of Buenos Aires
Chase Manhattan International/Chase Bank AG
Chase Manhattan International
Various hard currencies, including USD, Pesos, Liras,
Sterling, Francs, Yen, ECU, etc
Up to USD 500 million equivalent in series
Variable per series 1 -5 years. Variances to be
approved by the Central Bank
At par, discount or premium over par per series
Continuous basis with syndication, if need be, and
minimum offerings of USD 10 million equivalent
Each series can be at fixed, variable or zero coupon
Can be payable in arrears on agreed dates
Interest borne separately in each series by reference
to benchmarks such as LIBOR, LIBID, etc
As agreed between issuer and dealers
Do not bear interest and normally issued at a discount
Principal and interest free and clear of this tax
Definite Bearer and Registered Notes as agreed
Pricing Supplement to state redemption conditions
Luxembourg Stock Exchange (or as otherwise
specified in the applicable Pricing Supplement)
Direct, unconditional, unsecured, unsubordinated
ranking parí passu with all obligations of Issuer
Negative pledge in respect of specific factors
Interest cover charge applies. Other specific and
standard covenants included in Condition 4 of Offer
Document
In respect of present and future indebtedness, etc
English
Several restrictions included
Several restrictions on Registered Notes in reliance of
Rule 144A
Euroclear and Cedel for Bearer and Registered Note
and DTC for Registered Notes under Rule 144A
Issue price, date, maturity, principal amount, interest
rate applies to any Notes
As specified in Pricing Supplement

Sesión 1 -

A. Purpose of the Issue
The City of Buenos Aires initiated its USD 500 million Euro Médium Term Note Programme in
March 1997. The program was divided into individual series of notes, issued for 30 days to 30 years from
the date of the Programme's inception. The individual notes could be issued in a variety of currencies,
including the Argentinean Peso, USD, Italian Lira, etc. The purpose of the Programme was to refmance the
city's debt stock and restructure one of its banks - Banco de la Ciudad de Buenos Aires. The latter received
USD 100 million. Four series of notes have been issued under the Programme within a two month time span
and the entire Programme has been drawn down. Thefirst note series was issued in US dollars and targeted
primarily to the USD market. The intention was to gain credibility in the USD market and to build on this
placement in other markets.
The program is the city's first foray into the international bond market. Each of the individual
transactions sold out well and was 200% over-subscribed on average. The city has a sound administration
and reasonable international reputation. These considerations overshadowed a growing fiscal déficit in
recent years, brought about mainly by structural problems built into the current account. The city's debt at
the time of the issue represented only 1.4% of its GDP. This, coupled with a targeted reform program,
helped achieve reasonable ratings, which strengthened market perceptions.
The individual notes were placed at the higher end of the spread threshold. However, with the
exception of the Peso denominated offer, the city was quite interested in a geographic and investor
diversified placement and perhaps less with all-in costs. The role of the underwriter and its intemal financial
staff cannot be underestimated. This is indeed a common feature of such landmark issues.
B. The Underwriting and Marketing Process
The underwriting and marketing process was led by Chase Manhattan. Thefirst issue involved 10 comanagers, including:
Banco Bansud
Banco Francés del Rio de la Plata
Credit Suisse First Boston
JP Morgan Securities
SBC Warburg
Banco de Crédito Argentino
Banco Rio de la Plata
Deutsche Bank
Lehman Brothers
UBS
In the case of the Peso issue (USD 150 million equivalent), there were only four co-managers: Banco
Bansud, Banco de Crédito Argentino, Banco Francés and Banco Rio de la Plata.
The first issue under the program (USD 250 million) sold extremely well in the market, despite a the
rapidly weakening Argentine sovereign issues priced off the US Treasury benchmark (US Treasuries 2006)
during the marketing phase. The latter went from 330 basis points to 370 basis points over the benchmark in
just one morning. The city's main concern was clearly to have a successful issue at all costs. Its second
objective was to have a widely distributed placement with a strong US participation. In a tightening market
situation with the benchmark Treasuries performing poorly, the issue was a resounding success at the higher
spread spectrum. The transaction was 2 times over-subscribed with two-thirds of the notes sold to US
investors. Another important feature was that it attracted new money (rather than investors selling out of
existing portfolios). This can be considered a major achievement from the point of view of the city.

Sesión 1 -

The second USD 150 million equivalent in Pesos issue was struck in record time - with marketing
starting on a Friday and the price fixing taking place on the following Monday. The market timing was
carefully selected to coincide with a healthier market tone. The deal had first been tested in March.
However, at the time the city was concerned about the price and the relationship between the peso and the
Argentine Treasury 2006 benchmark. The spread differential during the marketing period of thefirst USD
250 million issue in April was 140 basis points. The USD 150 million equivalent in Pesos was launched in
an environment with a narrower 95 basis point differential. In this instance, the city expressed concern about
the all-in cost of Pesos over what was achievable in USD. The issue was subsequently fixed at about 70
basis points greater than where the city could do a seven year dollar tranche. As pointed out, these
considerations were overshadowed by a strategy to diversify the city's presence in as many markets as
possible. This is one of the reasons why the peso transaction was later accompanied by the lira offering. The
latter's performance was equally impressive. Chase was able to place the peso and lira denominated notes in
record time with investors pre-targeted during the preparation process.
The EMTN program as a whole was highly successful. The book or order demand was high in all
cases - about 2 times oversubscribed. The entire individual series of notes were sold out. This, coupled with
the unsecured nature of the notes, highlights the high level of investor confidence in the Administration.
Interest for the individual offers carne mainly from institutional investors, who purchased on average about
90% of the issues. The split between markets was more or less the same. The high coupons on each of the
notes, nevertheless, also attracted considerable interest from retail investors. The notes thus far seemed to
have experienced a strong secondary market demand.

II. Conclusions and Lessons Learned
Some extemal and intemal factors were important in the analysis of this case:.
•

Extemal Factors
Macroeconomic Conditions

The macroeconomic context (of the country and the city) played a major role in the issue. The
principal effect was its influence on the city's credit rating and investor perceptions. Key variables taken into
account were the Convertibility Law and its associated measures and the economic growth pattern of the
city. The city's resilience to the Tequila Crisis, for example, was a major achievement that helped solidify
the city's credit rating. The importance of the city to the overall national GDP, employment and income was
also taken into account in the credit rating and investors' risk perception. Another aspect of the economic
context was the high contribution of the city to the total Co-participation revenue and its correspondingly
low level of revenues in return. The city simply does not depend on central government transfers as is the
case with other local government authorities.
Bond Market
Bond market conditions at the time of the note sale had a major impact on the placement dates and
price arrangements. They also determined the overall underwriting and marketing process. An issue of this
size and complexity could not be sold on a "firm bid" underwriting basis. The underwriting industry does
not provide firm underwriting bids for sub-sovereign Emerging market issues. The underwriter in this
instance carried out extensive market checks and was able to forecast the placement of the issues
significantly in advance of the actual sale. There were nonetheless rapidly changing conditions at the time of
placement for thefirst and third series (USD 250 million and USD 150 million Peso series). The underwriter
faced higher bond prices because of increasing Argentine sovereign bond spreads vis-á-vis the US Treasury

Sesión 1 -

2006 benchmark. The issuer had to make difficult pricing decisions. Because the city was more interested in
a wide distribution than on all-in costs, the first series in particular sold at the higher end of the "spread to
Treasuries" spectrum. The peso series reversed this pricing process as market timing was more carefully
assessed. The lira issues were targeted at specific investor groups at high spreads.
•

Internal Factors - Issuer Related
Administration and Reputation

The administration's professional staff, and its reputation, were major factors behind the success of
the issue. It is important to bear in mind the difference between financial performance and the city's image.
The latter showed significant weaknesses. But in this instance, the reputation of the city was strong enough
to deflect at least some weaknesses.
Financial Flexibilitv and Reforms
Financial flexibility and reforms were one of the single most important factors that had an impact on
the preparation of the issue and its rating. The subject was extensively covered by the underwriter and the
rating agencies and occupied a major portion of the Offering Circular. One of the greatest concerns of the
rating agencies in particular was thefiscal gap and the reforms being put in place to reduce it. The rating
agencies view thefiscal déficit as a function of structural problems on the income and expenditure sides of
the accounts, and to some extent questioned/cautioned the city's ability to address these problems in the
immediate future.
It is essential to bear in mind that the issue itself is a function of this déficit and henee a mounting
short term debt - comprising late payments to suppliers and loans/capitalization needs of Banco de la Ciudad
de Buenos Aires. The bond was driven by the need to restructure these short-term obligations. The fact that
the notes were launched is in some way recognition of the difficulty in tackling the structural problems.

•

Internal Factors - Issue Related

It is important to note the perception and questions by investors during the road shows for the note
sale. In most cases, the concerns were about the reform program and in particular the measures aimed at
collecting a higher share of uncollected taxes, cutting back expenditures and the prospeets for the
privatization of Banco de la Ciudad de Buenos Aires. In contrast, in the case of Mendoza, the questions
concerned the legal tittle to oil royalties. These questions are all essentially about bond security. However,
none of the investors expressed interest in the likelihood or steps needed for a possible bail out. The same
issue would be treated differently in Brazil where there is a long track record of bailouts, and thus a stronger
element of moral hazard.

•

Financial Credibility

The single most important lesson from the issue is that the issuer needs to demónstrate strong
financial operations. When there is afiscal gap (as is the case here), the important consideration is to have
the mandate and reform program aimed at balancing the position. Buenos Aires would not have been able to
issue the notes with a poorly defined program of reforms. It would also have been in a more difficult position
had its fiscal gap not narrowed in the previousfinancial year. The gap had widened too quickly and would
have raised serious concerns at with the rating agencies. Thefinancial gap problem was also largely
compensated by a low debt burden. At 1.4% of GDP, it is one of the lowest in Argentina and Latin America.

•

Due Diligence and Rating

The issuer would not have been able to enter the international, or even the domestic market, on an
unsecured basis without the information base to facilítate due diligence and the rating process. These two
tasks are crucial in determining issue strategies, structures, underwriting and sales. The due diligence

Sesión 1 -

process always focuses on legal, commercial, financial and operational aspects. It is often structured around
rating agency requirements. The city was able to support both processes efficiently.

•

Role of Underwriter

There is no doubt that the role of Chase was of paramount importance in the success of the issue. A
landmark transaction of this nature normally puts greater emphasis on the role of the underwriter. The latter
can plan the preparation process, support the due diligence, manage the rating process, assess market
conditions and define thefinal strategy and structure. The role and efficiency of the underwriter is, however,
influenced by the quality of the information base and the local govemment's staff. In this instance, both
teams integrated well and played to their strengths.
The Buenos Aires issue is a landmark transaction matching issuer with issue strengths. The
transaction was well managed both from the point of view of preparation, offer and placement.

Sesión 1 -

Appendix 2
Province of Mendoza, Argentina
I. Description of the Bond
Details
US$125 million
Eurobond
Offered August 2,1996, Closed August 8
100% Fixed Re-Offer
Fixed at 10% per annum
Quarterly on 25 ,h of January, April, July and October
commencing 25 ,h of October 1996
7 years to July 25 2002
Same as Interest Payment
Direct; not subordinated
See Collateral
(i) 88% of Oil Royalty payments to the Province
(ii) Argentine Collection Account, US Collection
Account and Debt Service Reserve Account.
Guarantees:
None
Source of Principal & Oil Royalty payments to the Province by certain oil
Interest Payments:
producers under long term concessions to be
deposited in the Argentine Collection Account.
Debt Service
Minimum balance of debt service due on the next
Reserve Account:
Quarterly Payment Date.
Excess Collection
(i) Debt Service Coverage Ratio falls below 1.2 for
Trigger Events:
three consecutive collection periods, or
(ii) Proved Reserves divided by Production Level
(yearly basis) falls short of the remaining maturity.
Positive Covenants:
• Ensure validity and perfection of security and
assignment documents
• Maintain all necessary govemment approvals and
permits
• Ensure its rights under oil concessions are
maintained
Negative Covenants: • Create, transfer, etc. other liens on its rights to oil
royalties and other properties
• Modify the oil concession agreements
• Termination of dedicated oil concessions
Major Risks:
• Political risks on the regulations of the oil industry
• Fluctuation in world oil prices
• Oil reserves and future production
• Currency risks
• Country risk
Pertinent Conditions The notes were issued subject to certain international
Precedent:
and local credit ratings (see Credit Ratings)
Features
Amount:
Market:
Issue Date:
Issue Price:
Interest Rate:
Interest Payment
Period:
Maturity Date:
Amortization:
Ranking:
Security:
Collateral:

Sesión 1 -

Features
Events of Default:

Credit Ratings

Details
• Usual events of default on non-payment of debt
service and non-compliance of terms and conditions.
• Default on other outstanding indebtedness in an
aggregate principal amount of at least US$ 10m.
• Debt service reserve account falls below mínimum
requirement for six consecutive collection periods.
• Adverse changes to fiscal and constitutional laws
affecting the dedicated concessions and assignment
of royalties.
• Duff & Phelps International, BB
• Duff & Phelps Argentina, AA (Local)
• Ratto & Humphreys, AA+ (Local)

A. Purpose of the Issue
The province, as was the case with other provinces, had a short-term debt structure. As a result of
the Convertibility Plan, the province was unable to roll over existing obligations. An international bond
issue with longer-term maturity and atfixed cost of funding became an attractive alternative.
Financial considerations aside, the province viewed the issue as a good opportunity to further
expose it to the international market, gain credibility and diversify its funding sources. This was
particularly important in attracting direct foreign inward investments for the province's privatization
program.
The international bond issue was structured as a single transaction. However, the underlying
intention was to pave the way for further issues in the future as existing short-term debts mature and the
province's capital expenditure program expands.
B. Type of Bond
The Province of Mendoza issued the notes in August 1996 for a period of six years, maturing in
July 2002. The notes were the province'sfirst public debt issue, and thefirst oil royalty backed structure
for a Latin American issuer in the international market.
The notes were secured by a priority interest in 88% of all future oil royalty payments due to the
province from oil companies. A three-month debt service reserve was also funded from excess royalty
collections. A tight security package and bond structure was essential to insure the success of the
province'sfirst international bond issue.
The transaction was successfully placed despite a long marketing period and the disruption of the
Cavallo resignation. Investor interest carne mainly from US and European institutional investors attracted
to the relatively high 10% coupon
One of the primary factors that lead to the success of the issue was the pledge of oil royalties and
the flow of these funds to an off-shore collection account, as shown in Figure 1, below.

Sesión 1 - 0

Figure 1
Mendoza - Bond Funds Flow

Source: Province of Mendoza

II. Impact of the Issue
The credit rating and the due diligence process were the most important elements of the marketing
strategy.

Credit Rating
The international, and local, credit rating process required detailed and accurate financial and
economic information, which took the underwriting team a considerable amount of time to gather. Although
this effort yielded dividends, not only in the success of this issue, but also in shortening significantly the
preparation time of the second issue, there was little evidence that the actualfinancial reporting and
accounting standards used by the province were restructured. The province, however, took further steps in
tax reforms and implementing expenditure control programs following this issue. This resulted in an
improvement in its fiscal performance and eventually an upgrading of its foreign currency general
obligations rating to "BB" by Duff & Phelps.

Due Diligence Process
Legal due diligence, performed by Argentine counsel, was particularly important in the transaction
process. The key aspects of the bond issue the legal team examined included the oil concessions (terms and
conditions, validity of permits, oil royalties), the relevant hydrocarbon laws, the province'srightsto the oil
royalties and other revenues sources, as well as the validity of the collateral documents and arrangements.
Other important legal matters were the regulatory and constitutional framework goveming the province's tax
raising powers, expenditure responsibilities and debt issuance capability.
Thefinancial and economic due diligence process was by far the most difficult part of the entire
bond preparation process. The main obstacles were insufficient statistical information, and the lack of
Provincialfinancial information.

Sesión 1 - 1

Appendix 3
Guaymallen, Argentina
I. Description of the Bond

Guaymallen Municipality - Títulos De Obras Publicas (TOG2), Second Tronche of a US$ 3 millio
Global Programme, Collaterised Against Treasury Bonds (Bote 3), Enhanced with Co-participation
Revenues As Required.
The US dollar denominated notes were marketed in the local Mendoza area. The complete note
program was USD 3 million divided into three series or tranches of USD 1 million each.
Thefirst tranche (TOG1) was issued in December 1996. It had a maturity of two years and seven months
and was successfully placed. TOG1 had an interest rate of 10% andfive equal principal repayment periods. The
placement was made entirely in the Bolsa de Comercio de Mendoza. The issue was underwritten by Mercado
de Valores de Mendoza SA, with an over-subscription of 220%. The transaction was backed by Argentina
Treasury bonds (BOTE 3) in the amount of USD 1.19 million. These were deposited with afiduciaiy company
in Mendoza - Caja de Valores de Mendoza SA.
TOG 2 was the second issue in the series. It is important to note that a third series did not materialize
- principally as a result of the incumbent administration wishing to avoid transferring new debt obligations
to an incoming government in 1999. TOG 2 has a relatively short maturity of one year, with 50% of the
note's amortization taking place at the end of January 1999 and the balance at the end of July 1999. The
interest rate is 9.5% (fixed and significantly lower than the previous series), representing a 300 basis point
spread over the Argentine Treasury Note. The minimum purchase was USD 100 per lot with múltiples of
USD 100. The issue was sold at par valué, with commissions and costs charged to the issuer. The collateral
base, to cover principal and interest payments, was Argentina Treasury Bonds held by the municipality
(Series 3 or BOTE 3). However, in contrast with the previous series, the amount of pledged collateral rose
from USD 1.19 million to USD 1.26 million. This security package was further enhanced through pledges of
co-participation revenues from the Mendoza Province. In common with the previous issue, the BOTE 3
treasury bonds were placed in trust for the duration of the bond period with the samefiduciary firm - Caja
de Valores SA. Underwriting was once again the responsibility of Mercado de Valores de Mendoza SA. The
record 48 hour placement was also made in the Bolsa de Comercio de Mendoza.
A. Purpose of the Issue
The issue was to gain experience and credibility in the domestic bond market and to help raise funds
for civil works/capital projects. In contrast with many other public entities, Guaymallen did not come into
the market to restructure or rollover existing debt. The city wished to "experiment" with the bond market in
order to pave the way for an orderly entry into the emerging local capital markets. The restrictions imposed
by the Convertibility Law on local bank municipal and provincial government finance motivated
Guaymallen to explore altemative fiinding sources. The bond market provided this opportunity.
The city structured the bond issue in three series for experience and to establish contact with different
operators in the Mendoza Exchange. The administration was less concerned with price and collateral than
with a successful placement. In the longer run, its objective is to issue, as needed, on a non-recourse basis.

Sesión 1 -

B. TypeofBond
Despite the healthy financial standing of the municipality, the issue was structured around a high
collateral base, supported by treasury bonds amounting to USD 1.26 million. The collateral was pledged free
of liens and placed with a fiduciary and paying agent, Caja de Valores S.A. The higher collateral base for the
second tranche, TOG2, contributed to an interest rate that was lower than the TOG1 tranche.

• Permits/authorizations
The municipality handled all the relevant permits and authorizations:
•
•
•
•

The Executive Council - approved a certifícate within the provisions of a Decree No.
350/98 regarding municipal finance.
Provincial approvals and consents - under Decree No. 686/97, published in the Official
Bulletin of the province and order No 4429/97, the province had to approve the
transaction.
Ministry of Finance - a permit from the MOF was required in order to issue debt in hard
currency (resolution Nos. 1075/93 and 327/95).
Rating - the city selected Magister/Bank Watch and Standard & Poor' s.

The second issue was prepared and placed in six months. This compares favorably with the first
series, which took over a year. The second series benefited from the experience gained by municipality staff.
The timing in the first series was affected by lack of expertise, administrative hold ups and some internal
resistance to raise fixnds through a bond instrument.

•

Rating

The municipality engaged the services of two sepárate rating agencies to rate the issue:
• Magister/Bankwatch Calificadora de Riesgo
• Standard &Poor's
The advisor and thefinancial department staff managed the rating process. The rating agencies focused
their attention on the image and credibility of the municipality, legal due diligence (a focus on the fiduciaiy
arrangements)financial standing, debt stock obligations, information systems and intergovernmental relations.
They also assessed in more broad terms the economic base and structure of the city and the level of investment
flowing into the area. Notwithstanding these variables, however, the key focus of attention was the collateral
base pledged by the administration. This enhanced theriskprofile of the credit issue and eliminated most, if not
all, of any other concerns that might have arisen in other due diligence areas. In addition to the collateral, coparticipation revenue transfers further secured the bonds. The rating agencies assessed the central government
revenue transfer payment track record. These transfers would make good any shortfalls resultingfrom a forced
sale of treasury bonds and/or other municipality repayment income. Both agencies rated the issue AAA.

C. Subscription and Sale
The municipality mandated Mercado de Valores de Mendoza to underwrite the issue. This entity had
also handled thefirst series, with great success. The entity was also considerably cheaper than three commercial
banks that had approached the administration for the same purpose. Mercado de Valores was mandated by the
city directly without tenders. The commission charge amounted to 1% of the bond par amount, a veiy
competitive rate in the market for an issue of this nature. Mercado de Valores was thus solely responsible for
distribution of the bonds with the Mendoza Stock Exchange, primarily among personal and some institutional
investors. The underwriting contract wasfirm and required the underwriter to purchase the bonds. However, to
insure its ability to re-offer the bonds, Mercado de Valores had undertaken an extensive pre-sale contract to

Sesión 1 -

assess its ownrisks.This study involved its traditional agents/clients in the Exchange - some of whom are its
shareholders.

•

Marketing & Distribution

In common with the first issue, TOG 2 was more than twice oversubscribed. The issue was placed
entirely in the Mendoza market within a record period of 48 hours. The underwriter placed the issue entirely
through a selected list of agents contacted before the underwriting contract was signed. The brokers/agents reoffered the bonds to about 20 individual investors. The issue has not changed hands much since the placement
date. The secondary market is tight. This situation is unlikely to change between now and the maturity date. The
issue was placed during the peak of a bull market period. The actual placement dates coincided with the
inception of the Asian crisis and confidence has since been eroded quite considerably. The small amount of the
total transaction, its tight structure and collateralfiieled investor interest.

II.

Impact of the Issue

A. Benefits
The issue was targeted at institutional investors in the Mendoza Exchange. This had been the approach
taken in the previous placement - with great success. The main motivation by the municipality was to expand
its coverage/image among top investors in and beyond the province. The experience was also intended to
increase the staffs know-how and capabilities, paving the way for work on future issues. A third issue was not
considered appropriate: it would have overlapped with the administration period of the incumbent mayor and
raised the debt profíle of the city to historically high levels.

Conclusions and Lessons Learned
Guaymallen is a landmark capital markets case study for a small to médium sized municipality in
Latin America. Some of the lessons learnedfrom the Guaymallen bond issue are as follows:

•

Market Testing

The issue was not driven by short-term debtfinancing requirements (a common feature in all other cases
except Ibague), ñor by capital development expenditure programs (although the funds were specifically
assigned to complete infrastructure-related civil works). The main reason for the issue was to gain experience in
the capital markets and to prepare staff and politicians for future bond issues. The municipality understands the
difficulties experienced by other administrations with a past high dependency on short-term expensive debt. It
also understands therisksassociated with late payment delays to contractors. The latter has been another
traditional short term funding measure used by municipalities and provinces across the country. The city placed
all these considerations within the context of restricted bank lending; itself the result of Central Bank policy
reform measures. In view of these events, and given the need to maintain services, upgrade infrastructure
(whilst continuing to attract increasing prívate sector investment), the city targeted the bond market as one its
future funding options. This experimentation with the process and the players will increase its ability to react
and respond to market conditions. It will also pave the way for a future non-recourse bond issue.
•

Importanceof Financial Discipline

The city's strongfinancial discipline and disclosure is a model for other institutions, but there is still
room to improve. Guaymallen could benefit from a better information system, budget process and control. The

Sesión 1 -

concept of a supervisory commission for financial management is also an innovative element, and one that
should be replicated throughout the región.
The Guaymallen issue is one of the most interesting case examined. It highlights the aspirations of wellrun administration with a strong financial standing. It also outlines the shortfalls of market demand (tight
market) and the incompleteness of the due diligence and rating process. Guaymallen is well placed to becoming
one of thefirst such issuers to sell an unsecured general obligation bond in the domestic market.

Sesión 1

Appendix 4
Rio de Janeiro, Brazil
I. Description of the Bond
A. Purpose of the Issue
The purpose of the issue was to refinance domestic debts that had shorter maturity periods and higher
costs of funding. The municipality is prohibited under Federal Constitution to issue new debt, both
internationally and domestically, except for this purpose. Furthermore, the municipality considered that an
international debt issue would not only raise the city's profile vis-á-vis foreign investors but also pave the
way for further transactions in the international market. The issue was thefirst of its kind offered by a Latin
American municipality.
B. Type of Bond
Features
Amount:
Market:
Issue Date:
Issue Price:
Interest Rate:
Interest Payment
Period:
Maturity Date:
Principal Payment/
Amortization:
Ranking:
Depository
Source of Principal
& Interest
Payments:
Positive Covenants:
Negative
Covenants:
Major Risks:

US$125 million
Eurobond
Offered July 2,1996 Closed July 12,1996
99.96% Fixed Re-offer
Fixed at 10.375% per annum
Quarterly on 25 th of January, April, July and October
commencing on 25 th October 1996
3 years to 12 July 1999
Same as interest payment
Direct and unsubordinated, except for certain municipal
bonds that may be used to pay taxes under certain
circumstances
DTC
Municipal revenues, with Chase Manhattan as paying
agent
•
•

Standard
Standard

•
•
•
•

Currency Risks
City Credit Risk
Country Risk
Poor debt service history

The Municipality of Rio de Janeiro issued in July 1996 a 103/g% US$ 125 million three-year
unsecured Eurobond. This made Rio thefirst city in Latin America to issue an unsecured bond in the
international capital markets. The issue's registration follows standard Eurobond procedures and is listed on
the Luxembourg Stock Exchange and traded in the PORTAL Market. The notes are deposited with or on
behalf of DTC. Chase Manhattan is the Paying Agent.

Sesión 1 -

Despite the fact that this was the municipality'sfirst public international debt issue, the notes were
unsecured. The issue price was 99.96, providing a yield at launch of 10.6611% which was 25bp above the
Brazilian benchmark, or 400bp over the 1999 US Treasury Note at the time of the sale.
The transaction was considered to be highly successful as evidenced by the relatively small 25 basis
point spread to the sovereign bond. The book demand was high - 2 to 3 times oversubscribed. The
underwriters without the need to retain any bonds in their account sold out the entire issue. The successful
sale was accomplished in the last year of the Maia administration and with no collateral. Foreign investor
confidence in Rio combined with a comparatively attractive yield were the main reasons for the underwriting
success. Interest came mainly from institutional investors who purchased 95% of the issue. This was split
58% US, 25% Japan and 17% Europe. The high coupon also attracted interest among retail investors who
took up the remaining 5%.
There was a deep secondary market for the issue and the bond performance improved over time. The
spread on US Treasury narrowed to just 225 basis points one year later, reflecting a continuing rise in
investors confidence in the city brought about by consistent improvements in cityfinances and a positive
Emerging market tone (which has rapidly deteriorated since the spread of the Asian crisis).
C. Subscription and Sale
Merrill Lynch acted as the lead underwriter for the issue under afixed-price re-offer arrangement.
The underwriting discount was 0.75%, 60% of which represented selling commission and the remaining
40% underwriting/management fee. The underwriter's discount was on the low side in comparison to a
typical fee of 3% for this type of Emerging market transaction. However, the low fee was necessary in
view of Merrill Lynch's Iong-term interest in Latin American sub-sovereign debt issues. The market was
viewed by many of the leading investment banks as having high growth potential. As such, competition
was strong. Merrill Lynch competed against 19 other investment banks for the Rio underwriting mandate.
Merrill Lynch formed a syndicate consisting of 8 co-managers who were leading investment banks
targeted for their distribution networks in the US, Europe and Japan.
Figure 2
Rio de Janeiro - Syndicate Members
Underwriting
Syndicate Members
Amount
US$ million
Merrill Lynch Inc
101
Bayerische Vereinsbank AG
3
BB Securities Ltd
3
Commerz Bank AG
3
Dresdner AG
3
Nomura International Pie
3
Salomon Brothers International Ltd
3
Swiss Bank Corporation
3
Yamaichi International (America) Inc
3
Total
125
Source: Merrill Lynch

The issue was marketed in the US under Rule 144A to qualified institutional investors and outside
the US under Regulation S. The most important part of the marketing process was the road shows. Before
thefinal placement, Merrill Lynch presented the city administration and the bond transaction on road
shows in the US (New York, Boston, San Francisco, Los Angeles, the Mid West and Colorado). The
Sesión 1 -

presentations to investors were made by the city administration itself. They were well prepared and made
significant inroads into improving the city's overall credibility vis-á-vis investors.
The key pricing factor was market sentiment. This was thefirst city issue and was being offered
without collateral. Investors looked primarily to the strength of the city administration, which had the
reputation of a well-managed operation. The book order indicated that there was strong demand, but only
at a significant spread to the US Treasury yield. Issuers were driven by yield. The issue wasfinally priced
at 400 basis points above the 1999 US Treasury yield (which equated to 25 basis points above Brazilian
benchmark). At this price, the issue was 2 to 3 times oversubscribed. Following the initial launch, the
issue developed a deep secondary market. The spread on US Treasury narrowed to 225 basis points.

II. Impact of the Issue
The bond issue helped Rio establish an excellent reputation, particularly in its success in
substantially improving fiscal performance and promoting transparency. The city has a number of
supporting factors. It has a diversified economic base and plays an important role in the Brazilian
economy, accounting for 7.5% of the GDP. Against these strengths, however, the city still suffers from a
weak socio-economic profile, which will place an increasing strain on its finances.
The bond issued helped motívate the city to implement a comprehensive system of fiscal controls,
tighter tax administration procedures, transparency and accountability. Two international auditing firms
auditfinancial statements. This has contributed significantly to the city's good fiscal performance, and
enhanced its reputation in the international market. Despite these efforts, the city still suffers from serious
information gaps, and inconsistencies vis-á-vis world standards.
A negative impact for the city is Brazil's long and controversial history of sub-sovereign debt crises
and central government bailouts. While it was clear that the issue was sold purely on the credit strength of
the municipality, previous local government bailouts indicated that there was an implicit government
guarantee that was incorporated into the investors overall view of the credit risk.

Conclusions and Lessons Learned
The transaction preparation process was crucial to the credit rating process, particularly the
preparation of financial statements prepared with international accounting standards and the general due
diligence process.
The underwriting and marketing process affected thefinal uptake of the issue as well as the offer
price. The underwriting syndicate ensured that the placement was given a geographically large coverage.
In addition, the small size of bond allocations to co-managers and small lot size illustrated the strong
market demand.
Even though two international auditingfirms audited the city'sfinancial statements, the city still
suffered from serious information gaps. In addition, interpretation of the financial statements was
rendered meaningless due to the inconsistencies and differences that exist between Brazilian accounting
standards and international accounting standards. Brazilian accounting standards do not necessarily
reflect the economic andfinancial reality of municipal operations.

Sesión 1 -

Appendix 5
Ibague, Colombia
I. Description of the Bond
Features
Offer Date
Issuer
Agent
Amount
Tranche 1
Tranche 2
Collateral
Interest Margin
Maturity
Placement

Concept
July 1996 and July 1997
Municipality of Ibague
Fiduciaria Global SA
Pesos 8 million
Pesos 4.8 billion
Pesos 3.2 billion
Future gasoline taxes
Variable over DTF +3.5%
máximum
Five years
Bogotá, Medellin and Occidente

The Títulos were placed in two sepárate tranches of 4.8 million and 3.2 million pesos. The first
tranche was sold in July 1996 and the second in July 1997. The second tranche sale was conditional on the
gasoline tax cash flow performance for the first tranche issue. The gasoline tax was the security and
collateral for the bonds.
Both tranches were placed with brokers on the exchanges of Bogotá, Medellin and Occidente. The
maturity period was five years. The interest margin for thefirst tranche should have been based on the
variable rate DTF plus 2.75%, in line with Central Bank rate guidelines for this type of maturity. However,
this particular type of transaction fell outside the standard regulation and the actual market spread was DTF
+ 3.5%. DTF is the average rate for a 90-day period for banks and otherfinancial institutions that particípate
in the certifícate of deposits market. The relatively high spread for thefirst tranche reflected tight local
market conditions at the time, theriskprofile of the municipality, and the perceived uncertainty regarding
gasoline tax income. The second tranche attracted a margin of DTF +2%. The interest and principal payment
schedules were arranged in monthly, quarterly or semi-annual installments, with semi-annual principal
payments restricted to about 17.5% of the total. The amortization choice could be delegated to issue holders.
However, two additional conditions formed part of the structure: 40% of the principal could be repaid by the
48th month and 60% by the 60a1 month.
Duff & Phelps of Colombia gave the issue an A+ rating. The gasoline tax revenue performance during
thefirst 12 months of the issue was fully satisfactory. This factor, combined with a hike in the tax rate from
14% to 20%, eased the marketing period and allowed for the successful placement of the second tranche.
A. Purpose of the Issue
The issue proceeds were used to helpfinance a series of civil works aimed at improving road and
other transport infrastructure within the city's jurisdiction. Although there were alternative funding
instruments in place for infrastructure investment, the city's financial advisers suggested a bond transaction
structured around gasoline tax revenues to complement these sources. The city could raise the gasoline tax if
necessary, a fact which helped it receive the A+ rating. Another crucial advantage of this type of revenuebacked transaction is that it is not included in the measure of the city's direct or official debt stock. All
"titularizaciones" of this nature are exempt from the debt indicators used by the govemment to manage
municipal debt capacity, although the law has tightened on this since the time of the Ibague issue.

Sesión 1 -

The projects selected for finance under the new bond facility were as follows:
• Autopista Barrio Eduardo Santos - Descanso del Papa - construction, extensión and maintenance
• Maintenance and extensión of road systems
• Construction of viaducts and access roads in local districts
• Construction and expansión of various rural roads
• Provisión of 20% of the funding required for the construction and development of Avenida
Ambala and Avenida 79.
The issue was thefirst of its kind for the municipality and only the second in the entire country using
gasoline sales tax revenues as collateral for a bond issue. The effort was also seen as a means to enhance the
city's image and reputation in the market place. Despite this, the issue was perceived as a one-off transaction
rather than a debt program. It is important to note that the municipality had a healthyfinancial standing at the
time, matched by a reasonably ambitious plan to invest in services and facilities over the médium term. In
contrast with many other local govemment issues in Latin America at the time, the Ibague transaction was
not intended to refinance short or long term debt obligations.
B.

The Structure
Figure 3, below illustrates the structure of the Ibague bond transaction.
Figure 3
ibague - Tax Collection & Bond Payments

Source: Municipality of Ibague

The advisors for the transaction were Fiduciaria Global, backed by Corporacion Financiera de
Occidente. Global was also appointed agent/manager for the transaction.
Corporacion Financiera de Occidente is a limited liability company registered under Colombian
corporate laws. With appropríate certificates from the Superintendencia Bancaria, it can act as a financial
advisor to municipal govemments. Occidente's role in the bond transaction was to assist the authorities with
its design and structure. Thefirm also undertook limitedfinancial and organizational due diligence of the
municipality and supported the work on gasoline market assessments and projections.

Sesión 1 - 100

Fiduciaria Global was established in 1992 as a limited liability company with a mandate to establish
fiduciary and financial advisory operations. The fiim is registered and approved by the appropriate
authorities, including the Superintendencia Bancaria. Global is based in Bogotá, and is a subsidiary of a
larger group that has changed its ñame to UCN. The firm had a track record with the type of fiduciary
operations envisaged under the Ibague bond issue. Their role in the transaction was to co-ordinate the work
of Occidente and Ibague, and manage the transaction process. Some of Global's other tasks were as follows:
•
•
•
•

Administration of revenues
Co-ordination of the rating process and rating company
Support with the preparation of the circular
Management of appropriate permits, approvals and consents.

C. Subscription and Sale
The issue was difficult to place. Revenue backed transactions of this nature were fairly new in
Colombia at the time of the sale. Although municipalities could raise gasoline taxes, originally the máximum
rate allowed by central government was 14%. This was later increased to 20%, providing a strong Ímpetus
for municipalities to raise funds. Other assets eligible for collateral purposes include debts, company shares,
fixed assets and cash-flow. In more recent years, municipalities have also tried to collateralise duties on beer
and liquor sales. The underlying objective in all cases is that the funds raised need to be assigned to
infrastructure development.
Despite these considerations, revenue-backed operations of this nature are difficult to establish and
market. The government was, and remains, generally unhappy with the concept, the state accounting bodies
are unclear on its accounting treatment and most investors were concerned with the legal aspects particularly in connection with the title to, and the transfer of, public revenues of this nature. In addition,
most investors were initially concerned about the fact that the collateral was structured around future rather
than existing revenues. The advisers and the municipality staff recognized these problems and agreed to
structure the operation in two tronches, the second contingent on the performance of the first. This measure
provided additional support and incentives to the investors.
The issue was placed entirely in the domestic market. There were some significant differences
between the tranches:
• Thefirst tranche was placed with a relatively small number of institutions through 5 brokers
operating in the three Stock Markets in Colombia. The brokers were awarded quotas against a firm
underwriting commitment and an average commission fee of 1%. The brokerage arrangement was not
subject to an open tender. The order period was long and the margin correspondingly high. The original
buyers of thefirst tranche were as follows:
• The second tranche was acquired by one single broker in the Medellin area, for subsequent
marketing and distribution to institutions in the various exchanges. The commission rate in the second
tranche was negligible. The success of thefirst transaction impacted on this, as well as on the actual uptake
by institutional investors.

Sesión 1 - 10

Figure 4
Ibague First Tranche Uptake
1
4
2
1
2
1
1
2
1

Investors (Millions)
Insurance Company
Pensión Funds
Insurance Companies
Investment Fund
Fiduciaries
Mutual Fund
Bank
Cajas Compensación
Company
Total

Amount
350
1,400
730
100
250
100
200
1,550
120
4,800

Source: Municipality of Ibague

The bonds remain with a limited number of institutional investors. Trading has been narrow. The
capital market in Colombia is small and the level of liquidity tight. Ownership has been mainly with the
following bond holders:
Figure 5
Ibague Second Tranche Uptake
1
5
3
1
3
1
7
2
3

Investors (Millions)
Insurance Company
Pensión Funds
Insurance Companies
Broker
Fiduciaries
Bank
Employees Funds
Cajas Compensación
Companies/Others
Total

Amount
1,350
2,850
1,730
600
300
200
850
1,550
120
8,000

Source: Municipality of Ibague

The broker shown in the table above is the same one that acquired the issue on behalf of Ibague.

II. Impact of the Issue
A. Benefits

Credit Rating
The credit rating process was very beneficial to the municipality. Duff & Phelps awarded the issue a
credit rating of A+. The rating is quite high by local municipalriskstandards and reflects mainly the strong
collateral base. The two tranche series helped considerably in reducing theriskprofile, especially the second
trance which was contingent on the revenue collection performance of the municipality during the first year
of operations. The second issue was placed a year later and Duff & Phelps has maintained its A+ rating for
both issues.

Sesión 1 - 10

In contrast with some of the other case study credit ratings, the strong collateral base for the Ibague
transaction simplified enormously the due diligence effort by the rating agency.
The gasoline tax revenues, and the structure put in place for their collection and transfer to bond
holders, effectively "ring-fenced" the issue from the municipality's financial standing and shifting conditions
in business, financial and political environment. This is a significant consideration and an important lesson
for other municipalities wishing to enter the bond market - although the rating approaches differ
considerably between different Latin American countries. It should also be noted that the process in Ibague
was eased by the fact that the city was classified under the "green light" category, meaning that its financial
performance and debt structure was manageable and that it could issue bonds without central govemment
approvals. It is highly unlikely that the issue would have been processed had the city been classified in any
other category - with or without the collateral base.
The selection of Duff & Phelps was relatively easy: it is the only international agency operating in
Colombia that provides domestic ratings.
B. Difficulties and Weaknesses
The analysis of the Ibague case reflects several difficulties and weaknesses. The offering circular does
not provide detailed information about the economic structure of the city. Although the bonds are
collateralized by the gasoline tax, thefiscal condition of the city is important to gauge political risk. The
trastee may be able to collect gasoline taxes as long as the city maintains a positive operating balance. If the
city'sfinancial condition deteriorates, it may consider negating thefiduciary contract and capture the
gasoline revenues. The probability of this political risk is much reduced if the city is in goodfinancial health.
From this perspective the offering circular should review the economic base andfinancial condition of the
city.
Another important omissionfrom the circular and rating report is information on demographic factors.
There is no data to provide details on density, growth rates and age distribution of the city's population. A
declining or rapidly aging population will always tend to put a significant constraint on publicfinances since it shrinks revenues or accelerates service demands, especially on health care. High population growth
is likely to be the real situation in Ibague. This can be regarded as a positive feature in the context of the
issue since it could be partially attributed to high in-migration of working-age population, and in the context
of the collateral, of a greater number of vehicles and skilled workers. Another important aspect of
demographics is the size of the dependent population (the portion under 15 and over 65). These groups place
demand on services but contribute less to income. The circular does not cover this basic requirement.
Although thefinancia1 information in the prospectos provided an interesting account of performance
and expectations, under normal circumstances, it would have been virtually impossible to decipher its
validity, depth and degree of timeliness. Much of the data is collected manually and reported late. Without
the gasoline tax collateral and a soundfiduciary set up to handle gasoline tax revenues, the transaction may
not have stood the due diligence tests of most discemingfinancial investors, and in this sense it might have
failed. The rating report also attached importance to this collateral aspect, largely ignoring the inadequacies
of thefinancial reporting system and thefinancial performance. The deal was effectively "ring-fenced" from
other existing and future activities by the municipality. This includes the actual institutional/administration
capabilities since the collection and distribution of taxes were responsibilities mandated to fiduciary
concerns.
C. Conclusions and Lessons Learned
The issue strategy and structure were of paramount importance to the success of the issue. The main
consideration here was a broad investor approach and the use of the collateral base.
Sesión 1 - 10

The issue was targeted at the institutional sector with considerable pre-sale marketing efforts. The first
series was more difficult to place, although the strategy was largely adhered to. The key element behind
success was the collateral and two series tranching approach. This provides an important lesson, particularly
for landmark operations of this nature. The collateral base was backed by a tightfiduciary arrangement.
Without the collateral and fiduciary arrangements, it is unlikely that the issue would have enjoyed the same
reception. Thisfiduciary arrangement was much less well defined in the case of Valle del Cauca, with
serious consequences, especially in terms of rating and market credibility for the issuer. Most future
collaterised issues in Colombia will most likely have to follow the arrangements made by Global Fiduciaria
in the case of Ibague.
The main lessons learned from the issue are as follows:
•
•
•

Its "one-off' nature
Deficient information system, budget controls andfinancial planning

One-off Issue

The one-off nature of the issue is a weakness and a lesson for future transactions. Although the issue
was divided into two series, the overall market perception is for a one-off transaction linked to a specific
investment program.
•

Information, controls and planning

A common denominator with similar issues (by municipalities or provinces) is the weakness of the
information system. This will need to be remedied over time if the city is ever to appear in the market
without collateral. The same requirement could apply even for collaterised issues if at any time in the future
the credit rating agencies decide to introduce tighter disclosure and efficiency criteria. The remaining
constraint is the lack of detailedfinancial planning. The latter should include a better awareness of different
financial instruments. Budgeting and control is also a major deficiency.
In conclusión, Ibague can be regarded as a successful issue, taking place in more difficult market
conditions than similar issues by small municipalities in Latin America. Ibague is a usefiil case to highlight
the potential for raising disclosure and due diligence standards. Its fundamentáis areright.Eventually it
should consider issuing unsecured obligations. However, significant improvement infinancial reporting,
controls and planning are necessary before smaller municipalities can enter the bond market with bonds
secured by the city's full faith and credit.

Sesión 1 - 10

Appendix 6
Valle del Cauca, Colombia
I. Description of the Bond
Department of Valle del Cauca, 50.000 Million Colombian Pesos Notes
Collateralised Against Liquor Sale Taxes
Feature
Issuer
Amount
Series
Successful series
Placement
Target investors
Law
Security
Security arrangements
Underwriter
Agent
Use of Proceeds

Detail
Departamento Valle del Cauca
P$ 50 billion
Planned in 3 series varying maturities
Series C, seven year maturity
Bolsas de Valores: 3 domestic markets
Institutional and retail
Colombian - Codigo de Comercio
Tax on Liquor sales
Fiduciary for bond holder repayment
Banco de Colombia
Fiducolombia SA
Debt restructuring/capex programme

The notes were issued in December 1996 by the Department of Valle del Cauca for a period of seven
years, maturing in December 2003. They are guaranteed by taxes collected on regional liquor sales. These
taxes were structured to cover up to 130% of the debt service obligation (including principal and interest). In
the event of tax revenue short falis, the Department would be forceó pledged other current operating
revenues. Under the prevailing capital market laws, all tax revenues that secure bonds are permanenüy tied
to the specific debt obligation, and can not be used by the issuer for any other purpose. The law itself
provides stiff penalties, and even prosecution, for non-compliance.
The bonds were placed in three domestic stock exchanges over a two-day period - during the 23rd and
the 26 a of December of 1996, dates considered quite unusual for this type of transaction and market. The
main reason for selecting these dates was the department'sfiscal shortfall.
The issue was directed at domestic institutional and personal investors and is listed in the stock
exchanges of Bogotá, Medellin and Occidente. Institutional investors purchased most of the bonds in the
Bogotá Stock Exchange. The proceeds were intended to cover thefinancing requirements of projects under
the capital budget (25 billion pesos) and to refinance existing debt (25 billion pesos). The total transaction
was structured around three sepárate tranches, with an amount of 50,000 million pesos each.
•
•
•

Tranche A:five years maturity - not placed
Tranche B: six years maturity - not placed
Tranche C: seven years maturity- successful placement

The inability to place the first two tranches was a function of poor issue price structure, relatively Iong
maturities and bear market conditions. For the first two tranches the market was tight and the department had
to compete with other transactions. The market improved considerably towards the period coinciding with
the third tranche. The high price margin paid by the city was the most important factor in the sale of the third
trance. In addition, longer maturities were not a concern by the time the third trance was sold, as the market
entered into a bull market phase.

Sesión 1 - 10

A. Purpose ofthe Issue
The bond concept and preparation process was led by the administration itself. The entity had
gathered considerable expertise from a previous issue with similar characteristics in 1991. The issue was
required to meet debt obligations and to help fund capital expenditure. A summary of this investment
program is shown below:
Concept
Public Works
DATT
Education
Health
Hacienda
Community Development
Agriculture
Indeval
Acuavalle
Government
Valorización
Visevalle
Social Programmes
Tota)

Works
(Number)
47
2
39
53
10
31
9
10
7
4
3
4
2
221

Amount
(Mil Peso)
5,556.4
355.2
2,539.2
3,226.8
2,687.8
2,743.0
546.8
915.0
2,399.1
296.0
3,300.0
352.0
190.0
25,107.3

In terms of the refinancing outstanding debt, the department targeted the following specific
obligations:
Institution
Banco Uconal
Caja Agraria
Caja Agraria
Corfiboyaca
Banco del Estado
Banco Industrial
Banco Industrial
Banco Ganadero
Las Fes
Bancafe

Amount
(Mil Peso)

Margin

1,800.0
1,500.0
2,080.0
3,600.0
800.0
3,000.0
1,833.3
6,600.0
2,500.0
1,286.6

DTF + 6.0
DTF + 4.5
DTF + 4.5
DTF + 6.0
DTF + 6.0
DTF + 6.0
DTF + 6.0
DTF + 5.0
DTF + 5.0
DTF + 5.0

Issue Date Maturity Date
Apr 95
Oct 94
Jun 94
Apr 95
Dec 94
Jan 95
Jul 95
Apr 96
Apr 94
Oct 93

4/00
10/99
06/99
04/00
12/99
06/00
07/00
04/01
04/99
10/97

B. Subscription and Sale
The bonds were placed in just two days between the 23"1-26th of December. The pre sale efforts by
Banco de Colombia contributed to the rapid placement. It is also important to note that market conditions
were highly favorable (bull market). The pre-sale work was performed among the bank's regular clients. The
bank's treasury department led the work. The issue did not have a road show. The number of purchase
tickets was limited, as shown below:

Sesión 1 - 10

Buyer (Mil Peso)
Personal Investors
Real Sector
Banks
Financial Corporations
Trustees
FCO
Pensión Funds
Total

Tickets
2
5
2
5
9
15
30
68

Figure 6
Valle De Cauca Uptake
Amount
% of total
260
0.52%
1,450
2.90%
3,570
7.14%
15,650
31.30%
4,010
8.02%
7,750
15.50%
17,310
34.62%
50,000
100.00%

The make-up of the bondholders has not changed significantly since the initial offering, as shown
in the table below for July 1998:

Buyer (Mil Peso)
Personal Investors
Real Sector
Banks
Financial Corporations
Trustees
FCO
Brokerage Firms
Insurance Companies
Pensión Funds
Total

Figure 7
Valle De Cauca Current Bondholders
Tickets
Amount
% o f total
1
100
0.20%
39
10,900
21.80%
5
5,100
10.20%
9
5,300
10.60%
1
160
0.32%
5
10,450
20.90%
19
120
0.24%
6
1,180
2.36%
15
16,690
33.38%
68
50,000
100.00%

II. Impact of the Issue
The most challenging factors for the bond preparation was the due diligence process and projections
related to liquor revenues. The liquor tax was the centerpiece of the collateral package. It also almost entirely
defined the rating process.
The liquor tax projection was inadequate, despite a direct involvement by the department in liquor
sales (through one of its companies) and the historical record of sales from the previous issue.
Due diligence was conducted by the department's own economics and financial staff, supported by a
financial advisor and Banco de Colombia. Thefinancial advisor was afinancial firm with experience in
similar bond issues - Comisionista de Colombia. In complete contrast with the other domestic bond case
studies, the involvement of the advisor was fairly limited. This type of situation tends to develop with
follow-up issues that have the same structure as the initial offering. Indeed, a keyfinding from case studies is
that landmark issues are much more difficult to prepare and sell than subsequent operations; the problem
being compounded by the lack of experience at the level of the issuer. In Valle del Cauca, the advisor went
as far as to relinquish responsibility for the actual quality of the information in the placement prospectus. In
contrast with other countries, however, in Colombia this exclusión is not possible i.e. the advisor is jointly
and severally responsible with the issuer for the circular contente.

Sesión 1 - 1 0

Conclusions and Lessons Learned
The key factor to take into consideration in this case is the regulatory framework which facilitated the
issue process and its placement. The critical elements are the tax raising powers of the department, its debt
structure and the formal govemment approval process.
Valle has considerable leverage on raising taxes. However, the administration has also inherited major
service delivery responsibilities under the decentralization process. This has had a negative impact on
Valle'sfiscal health, and as a result will likely hinder the sale of unsecured bonds. The department's debt
structure, however, is not reported as being excessively high, thus allowing it to qualify for bond sales under
the "green light" categoiy. However, thefinancial reporting systems that determine the department's fiscal
condition are not reliable. Consequently, the "green light" allowing the department to issue bonds may be
misleading.

Sesión 1 - 10

Appendix 7
Latin America Sub-National Bond Issues
DataJsaued*. Place

Argentina
8/10/93

Province

11/23/95
2/12/96
5/1/96
5/1/96
5/1/96
5/1/96

Province
Province
Province
Province
Province
Province

Province
6/23/94
10/4/95
Province

lasuer

.

-

Coraooa
Buenos Aires
Buenos Aires
Buenos Aires
Buenos Aires
Entre R O
Í S
Entre Ríos
Entre Rios
Entre RÍOS

Currency

USO
USD
USO
DM
DM
ARP
ARP
ARP
ARP

8/2/96
9/18/96

Province
Province

Mendoza
Buenos Aires

USO
SFR

3/1S/97
4/3/97
4/11/97
05/11/97
5/13/97
5/23/97
5/28/97
í/iaOT
8/14/97
8/22/97
9/1/97

Municipality
City
City
Province
City
Province
Province

Guaymallen
Buenos Aires
Buenos Aires
Buenos Aires
Buenos Aires
Buenos Aires
Buenos Aires

Province
Province

Mendoza
Tierra del Fueqo

USD
USD
USD
USD
ARP
LIT
USD equ.
USD
USO
USD
USD

Province
Buenos Aires
Province
Tucuman

BoBvta

Amount

: Type

60
100
100
107
169.22
13
22
22
20

GOV
GOV
GOV
GOV
GOV
GOV
GOV
GOV
GOV

125.00 GOV
161.19 GOV
3.00 GOV
250.00 GOV
250.00 GOV
250.00 GOV
150.00 GOV
0.1 (Irill) GOV
150.00 GOV
69,000.0 GOV
GOV
250.00 GOV
55.00 GOV

200.00

Maifcet

Euro-Pnv
Euro
Euro
Euro
Euro
Dom
Dom
Dom
Dom

Euro
Euro
Dom
Euro
Dom
Euro
Euro
Euro
Euro
Euro
Euro
Euro
Dom&Glob

Retino

CoDaterel '

BB-ÍS&P)
BB-(SiP)
B1 (Moodys); BB-(S&P)

BB (Duff & phelps Int.); AA (Duff 88% Oil royalty, Argentine collection
&Phelps Argentina);AA+ (Ratto & Account, US collection account Debt
Humphres)
So/vico Reserve
B1 (Moodys); BB- (S&P)
AAA (SíP), AAA(M/B calificadora
de riesgo)
81 (Moodys); BB-(S&P)
B1 (Moodys)
B1 (Moodys);
B1 (Moodys);
B1 (Moodys);
51 (Moodys);
BB- (S&P)
81 (Moodys);

<

BB-(S&P)
BB- (S&P)
BB- (S&P)
B8- (S&P)
BB- (S&P)
.

11/23/94
10/8/97

Municipality
Municipality

Gobierno Municipal de La Paz
Gobierno Muniapal de La Paz

USD
USO

3 GOV
3.5 GOV

Dom
Dom

n/r
n/r

CTA Receptora Sinking Found
CTA Receptora Sinking Found

11/6/97
11/7/97
12/5/97

Municipality
Muru'cipality
Muniapaiity

Gobierno Municipal de La Paz
Gobierno Municipal de La Paz
Gobierno Municipal de La Paz

USD
USO
USD

2.5 GOV
3 GOV
2 GOV

Dom
Dom
Dom

n/r
n/r
n/r

CTA Receptora Sinking Found
CTA Receptora Sinking Found
CTA Receptora Sinking Found

BazB
6/1/93
8/1/94

City
City

Sao Paulo
Sao Paulo

REALES
REALES

99.00 GOV
99 GOV

Dom
Dom

8/12/96
7/2/96
11/1/96
6/1/97
6/2897
9/1/97

Municipality
City
City
City
State
City

Rio de Janeiro
Rio de Janeiro
Rio de Janeiro
Rio de Janeiro
Sao Paulo
Río de Janeiro

USD
USD
REALES
REALES

Muniapaiity
Muniapaiity
Munkapalily
Municipality
Munidpaiily
Munídpafrty
Municipality
Municipality
Municipality
Municipality
Munia'pality
Municipality

Palmira
Pereira
Santiago Calí
Santiago Cali
Cartago (Valle)
Medeilin
3ucaramanga
Santiago Cali
Itagui
ñionegro
Santiago Calí
Medeilin

Colombia
1991
1991
1991
1991
1992
3/17/93
9/1/93
8/1/94
9/22/95
9/26/95
11/16/95
4/24/96

REALES

125.00
125
17
21

61 (Moodys); Bt (S&P); B8-(duff
Phelps)
BB+(S&P)

GOV
GOV
GOV
GOV
GOV
21 GOV

Euro
Euro
Dom
Dom
Dom

n/r

1,000.00 GOV
2,600.00 GOV
4,000.00 GOV
4,000.00 GOV
700.00 GOV
15,000.00 GOV
4,978.00 GOV
46,437.00 GOV
11,660.00 GOV
2,000.00 GOV
20,003.00 GOV
20,000.00 GOV

Dom
Dom
Dom
Dom
Dom
Oom
Dom
Dom
Dom
Dom
Dom
Dom

n/r
n/r
n/r
n/r
n/r
n/r
n/r
n/r
A
n/r
AA-, A+
AA

_

-

Sesión 1 - 10

Sesión 1-110

EL ENDEUDAMIENTO DE LOS GOBIERNOS SUBNACIONALES:
EL CASO DE LAS PROVINCIAS ARGENTINAS

Alejandro Zentner
Juan Sanguinetti
Versión Preliminar

Sesión 1 - 1

Indice

I.

Introducción

113

II. Los efectos del endeudamiento subnacional sobre la gestión macroeconómica. Un breve
marco analítico
115

A.

La visión tradicional

...—

—...

116

III. El enfoque de la disciplina del mercado crediticio

119

IV. El uso de controles internos al sector publico

120

V. El endeudamiento de las provincias Argentinas

A. El contexto institucional

126

126

VI. El marco regulatorio del endeudamiento subnacional en Argentina

138

A. Presupuestos y reglas de endeudamiento provincial............................................................. 138
VII. Síntesis y conclusiones

146

Bibliografía
Anexos

148
150

Sesión 1 - 1

Abstract
El análisis de los controles al endeudamiento subnacional constituye una importante línea
de investigación en los estudios sobre federalismo fiscal, y en particular, de aquellos
concentrados en la relación entre la descentralizaciónfiscal y la gestión macroeconómica.
Existe abundante evidencia empírica que indica que el marco institucional de las relaciones
fiscales intergubernamentales resulta clave para que un proceso de descentralización
resulte exitoso, en el sentido de fomentar la solvenciafiscal de los gobiernos subnacionales
y brindar los incentivos que conduzcan a una gestión pública más eficiente en términos de la
cantidad y calidad de los servicios públicos provistos. En este sentido, la existencia de un
régimen de transferencias intergubernamentales con cierta cuota de discrecionalidad, la
posibilidad de utilizar los bancos públicos para obtenerfinanciamiento adicional o contraer
deuda flotante con empleados públicos y proveedores, la ausencia de mecanismos de control
al endeudamiento subnacional o el uso indiscriminado de garantías para la contratación de
préstamos con instituciones financieras, son factores que debilitan la restricción
presupuestaria que enfrentan los gobernantes. El presente trabajo enfatiza los aspectos
institucionales del control del endeudamiento subnacional para el caso Argentino. Se
analizan los diferentes mecanismos de control del endeudamiento de las provincias
argentinas instrumentados por el gobierno federal. En particular, se analiza la distorsión
que introduce la utilización de los recursos del régimen de coparticipación federal de
impuestos como colateral y de que manera esta distorsión impide que el mercado financiero
funcione como disciplinador.

I.

Introducción

En los últimos años, el análisis de la relación entre descentralización y gestión macroeconómica ha
tomado preponderancia en la literatura sobre federalismo fiscal. Esta literatura ha puesto en evidencia que
en los países que han adoptado una organización federal para su gobierno, la gestión macroeconómica y la
estabilización de la economía se ven seriamente influidas por las normas y prácticas del federalismo fiscal.
Algunos autores han adoptado una posición "pesimista", subrayando que los procesos de
descentralización podrían acarrear costos o "peligros" que en alguna medida pueden atenuar las ventajas
de la descentralización, en particular, se refieren a los procesos de descentralización llevados a cabo
últimamente en los países en desarrollo y en las economías en transición (Bird et al, 1995; Prud'homme,
1995; Tanzi, 1995).
Al respecto, Tanzi (1995) señala que existe suficiente evidencia empírica en los países en desarrollo
que bajo ciertas circunstancias las jurisdicciones subnacionales han contribuido -en muchos casos
significativamente- al emperoramiento de los problemas macroeconómicos. En dicho trabajo argumenta
que "en diferentes países en desarrollo los gobiernos locales impulsaron aumentos en los gastos por
encima de los incrementos en sus recursos, incrementando así la deuda y ocasionalmente forzando a los
gobiernos centrales a salir en su rescate".
Para Ter-Minassian (1997) esta aparente contradicción entre descentralización y gestión
macroeconómica es mucho mayor cuando los gobiernos subnacionales no operan en condiciones de
limitaciones presupuestarias estrictas, es decir, cuando hay plena autonomía en los gobiernos
subnacionales.

Sesión 1 - 1

Claramente, la percepción de los políticos sobre la dureza de la restricción presupuestaria resulta
clave para que un proceso de descentralización resulte exitoso, en el sentido de fomentar la solvencia
fiscal de los gobiernos subnacionales y brindar los incentivos que conduzcan a una gestión pública más
eficiente en términos de la cantidad y calidad de los servicios públicos provistos.
A su vez, existe cierta evidencia que determinados arreglos o marcos institucionales en las
relaciones fiscales intergubernamentales o ciertas características de lasfinanzas subnacionales pueden
debilitar la dureza de la restricción presupuestaria, y por lo tanto, la disciplinafiscal (Jones, Sanguinetti y
Tommasi, 1997; Stein et al, 1999; Dillinger and Webb, 1998; Nicolini et al 1999). En este sentido, la
existencia de un régimen de transferencias intergubernamentales con cierta cuota de discrecionalidad, la
posibilidad de utilizar los bancos públicos para obtener financiamiento adicional o contraer deuda flotante
con empleados públicos y proveedores, la ausencia de mecanismos de control al endeudamiento o el uso
indiscriminado de garantías para la contratación de préstamos con institucionesfinancieras debilita la
restricción presupuestaria que enfrentan los hacedores de políticas.
En particular, la literatura pone de manifiesto que las normas y prácticas asociadas al
endeudamiento subnacional constituyen un aspecto clave en todo proceso de descentralización. Según cual
fuese su diseño institucional-legal esta aparente contradicción entre desecentralización y estabilización
puede exacerbarse.
Los controles al endeudamiento y la posibilidad latente de un salvatage por parte del gobierno
central afectan directamente la percepción del mercadofinanciero acerca de la restricción presupuestaria
que enfrentan los gobiernos subnacionales. En este sentido interesa conocer de que manera dicho nivel de
gobierno accede a los mercadosfinancieros y cual es el papel del gobierno central.
En Argentina las provincias tienen una gran autonomía para contraer endeudamiento. La
constitución nacional no hace referencia alguna a las facultades de los gobiernos provinciales para
contraer deuda pública. No obstante, en casi todas las constituciones provinciales se establecen límites
específicos a los servicios de la deuda, y en algunos casos, al destino de estos créditos.
En la Argentina sólo existe información de la deuda provincial para la historia reciente. En los
últimos años, la deuda consolidada de las provincias y la Ciudad de Buenos Aires ha crecido en forma
importante (más de un 100% desde 1994). Para fin del año 1999 se espera que el endeudamiento
consolidado de las provincias alcance una cifra cercana a los U$S 18.000 millones (U$S 19.500 millones
si se considera la deuda aun no consolidada de la Ciudad Buenos Aires con la Nación), una deuda que
representa un 50% de los ingresos anuales de dicho nivel de gobierno.
Sin embargo, la magnitud agregada del endeudamiento subnacional no es la variable más
informativa para advertir el potencial problema sobre la gestión macroeconómica nacional -nótese que la
deuda provincial agregada es reducida en relación con la deuda del gobierno nacional que supera en 1999
los U$S 130 mil millones.
En cambio, el potencial riesgo puede percibirse con mayor claridad analizando la gran
heterogeneidad entre las performances fiscales de las distintas provincias. Existen jurisdicciones que han
tenido una política fiscal ordenada y responsable, en las cuales la deuda no representa problema alguno
(San Luis y La Pampa, por ejemplo). Por el contrario, en otras provincias, los altos y recurrentes déficit
públicos han generado deudas importantes que obligaron a comprometer una alta proporción de sus
fondos coparticipados, afectando seriamente suflujo de ingresos corrientes, comprometiendo la prestación
de los servicios básicos (salud, educación, justicia y seguridad) y obligando al gobierno nacional ha
realizar rescates a algunas provincias.
Al respecto, Nicolini et al (1999) argumentan que: "Entre los episodios que han tenido mayor
repercusión pública y que por lo tanto resultarían interesantes analizar se destacan: la crisis recurrente que
enfrenta la provincia de Jujuy y que desencadenó la renuncia del gobernador a fin del año pasado, la crisis
de Tucumán que enfrentó durante 1993, la crisis que se desencadenó en la provincia de Santiago del

Sesión 1 - 1

Estero y que culminó con el incendio de la casa de gobierno y posterior intervención del gobierno federal
y las sendas crisisfiscales en Río Negro y Córdoba que se desencadenaron luego de la crisis del tequila.
Varios han sido los mecanismos que el Poder Ejecutivo Nacional utilizó para aportar recursos adicionales
a las provincias que atravesaban dificultades fiscales yfinancieras, entre los más importantes, deberíamos
mencionar los anticipos de coparticipación, los Aportes del Tesoro Nacional, préstamos del Banco de la
Nación Argentina u otras instituciones bancarias oficiales (por ejemplo, BICE), préstamos en títulos
públicos (BOTESO 10) y redescuentos del Banco Central a los bancos provinciales, entre otros." Además,
no puede dejar de mencionarse las recurrentes intervenciones a la provincia de Corrientes en los años
1992,1995 y 1999.
Esta situación ha llevado a que se plantearan, desde distintos ámbitos, estrategias tendientes a
solucionar el problema de la deuda de las provincias argentinas. FADE (1999), por ejemplo, ha presentado
una propuesta para modificar el perfil de vencimientos de deuda pública en las provincias. Sin embargo,
este trabajo sólo propone una tecnologíafinanciera para atenuar la carga de la deuda en el corto plazo sin
analizar como rediseñar la relación Nación-Provincias de modo de plantear una solución de fondo. Aún
más, tal como está planteada la propuesta -sin una redefinición del marco institucional de la relación
Nación-provincias- puede considerársela como un nuevo episodio de rescate.
Otro propuesta fue la incluida en el proyecto de ley para reformular el esquema de coparticipación
federal de impuestos del Senador R. Ortega, en el cual se plantea la necesidad de acordar entre las
provincias metas similares a las planteadas en el tratado de Maastricht para limitar el endeudamiento
subnacional.
El propósito de este trabajo es estudiar la problemática del endeudamiento subnacional en el caso de
las provincias argentinas. Es extremadamente importante este análisis, dado que si se parte de un
diagnóstico equivocado las recomendaciones de política pueden no ser las más apropiadas. Un análisis
integral debe incluir un estudio de las condicionesfinancieras de las provincias para plantear una
propuesta de corto plazo, pero además debe incorporar un análisis del marco institucional del
endeudamiento provincial de modo de instaurar un nuevo esquema de comportamientofiscal provincial en
el cual los comportamientos fiscales de una provincia no perjudiquen la estabilidad macroeconómica
nacional y de las demás provincias.
El trabajo se organiza de la siguiente manera. El segundo capítulo, comienza con un análisis de la
literatura sobre los efectos del endeudamiento subnacional sobre la gestión macroeconómica en un país
federal. En el tercero se presenta el enfoque de la disciplina del mercado para los gobiernos subnacionales.
El cuarto capítulo contiene un análisis de los controles al endeudamiento subnacional, adoptando un
criterio normativo y empírico. El capítulo quinto profundiza sobre la problemática de la deuda provincial,
presentando información sobre su evolución y composición. En el sexto capítulo se presenta de una
manera estilizada la naturaleza de los controles al endeudamiento subnacional para el caso Argentino. Este
análisis incluye el papel de los controles internos de cada jurisdicción subnacional, el rol del gobierno
central y del Banco Central, el papel de los bancos y entidadesfinancieras y de las calificadoras de riesgo.
En el séptimo capítulo se analiza acerca de si el mercadofinanciero ha tenido un rol disciplinador para los
gobiernos subnacionales. En el último capítulo se presenta una síntesis y las principales recomendaciones
de política.

II.

Los efectos del endeudamiento subnacional sobre la gestión
macroeconómica. Un breve marco analítico

Una controversia de la teoría de lasfinanzas públicas en países con varios niveles de gobierno es
acerca de qué nivel de gobierno debe ser el responsable de la política de estabilización. En la división
tradicional de funciones entre el gobierno nacional y los gobiernos subnacionales que se remonta a

Sesión 1 - 1

Musgrave (1959) se asigna al nivel central la responsabilidad de la política de estabilización. Sin embargo,
contemporáneamente esta visión ha sido rebatida, asignándosele al nivel subnacional de gobierno mayores
prerrogativas en cuanto a la capacidad de desarrollar una políticafiscal autónoma.

A. La visión tradicional
La conveniencia de que sea el gobierno nacional quien absorba la mayor parte de las fluctuaciones
en los ingresos ha sido ampliamente analizada en la literatura, por lo cual aquí sólo es menester realizar
una mención sintética.
El primer argumento a favor de la centralización de la políticafiscal es que el gobierno nacional
posee más instrumentos de política macroeconómica que los gobiernos subnacionales. Es sabido que el
uso de una moneda única a nivel nacional implica que la política monetaria deba permanecer centralizada.
A su vez, dados los estrechos vínculos entre la política fiscal y la política monetaria se arriba a una
conclusión que, en principio, alienta la centralización de la política fiscal.
En segundo lugar, a partir del análisis del multiplicador Keynesiano se puede advertir la ineficacia
de otorgar la política anticíclica al nivel subnacional de gobierno. Esto se debe a dos motivos. El primero
es la reducida magnitud relativa que, desde el punto de vista fáctico, generalmente poseen los
presupuestos subnacionales considerados aisladamente. El segundo proviene del menor tamaño relativo
del multiplicador fiscal que resulta del carácter abierto de las economías provinciales. Dicho carácter
provoca que los estímulos fiscales promovidos por los gobiernos subnacionales sean en gran medida
exportados. En ambos casos se revela la ineficacia, en el sentido de impacto relativo despreciable, de
asignar a los gobiernos subnacionales el rol de la estabilización del ciclo.
En tercer término, también como consecuencia del carácter abierto de las economías de los
gobiernos subnacionales, los responsables de ejecutar las políticas locales no poseerán los incentivos
adecuados para llevar a cabo una política suficiente. Por un lado, las autoridades de los gobiernos
subnacionales enfrentan el costo pleno de las medidas de política que emprenden, y por el otro, los
beneficios de estas políticas se derraman sobre las regiones lindantes. A pesar de la similitud de este
argumento con el anterior -de hecho, ambos se encuentran íntimamente relacionados- existe una fina
diferencia. Mientras, en este último caso lo que resulta reducido son los incentivos a llevar a cabo la
política macroeconómica, en el caso anterior son los efectos de la política macroeconómica los que
resultan despreciables.
En cuarto lugar, se debe señalar que las prerrogativas que se asignen a los gobiernos
subnacionales en cuanto al manejo de la deuda pueden desembocar en un resultado subóptimo. Es decir, si
se permite el libre endeudamiento de los gobiernos locales en un escenario de costos marginales crecientes
se generará un endeudamiento excesivo, dado que cada gobierno solo evaluará su propio costo de
endeudamiento no internalizando los efectos externos producidos sobre los demás tomadores de deuda.1
Este costo se magnifica si, en un contexto en el cual no existe información perfecta que permita distinguir
claramente entre deudores potencialmente solventes o insolventes, algún participante no es capaz de
honrar su deuda. Si este fuera el caso puede ser posible que el mercado penalice injustamente a deudores
solventes.
Por último, y también relacionado con el manejo de la deuda, existe la posibilidad de que el
gobierno nacional actúe explícita o implícitamente como prestamista de última instancia, desarticulando
1

Este hecho no sólo es aplicable al endeudamiento de los gobiernos subnacionales sino también al gobierno nacional. Sin
embargo, una mayor cantidad de deudores de m n r magnitud internalizarán una menor parte de los costos extemos de sus
eo
decisiones que u único deudor que los nuclée a todos. Es decir, mientras los pequeños deudores t m n la tasa de interés c m
n
o a
o o
u dato, u gran deudor observa que la oferta que enfrenta tiene pendiente positiva. Obviamente, aunque el gobierno nacional
n
n
concentre el endeudamiento público, a n no se internalizan los costos extemos sobre los deudores del sector privado.
ú
Sesión 1 - 1

los incentivos a evaluar y monitorear el comportamiento de los deudores subnacionales. Ante la
percepción de rescate, el mercadofinanciero, actuando racionalmente, no disciplinará en una magnitud
suficiente a los gobiernos locales dado que anticipan que en el peor de los escenarios el gobierno nacional
brindará el apoyofinanciero. Esto es, la estrecha relación entre el gobierno nacional y los gobiernos
subnacionales puede ablandar la restricción presupuestaria de estos últimos.
Los argumentos mencionados permiten advertir el conflicto entre las funciones estabilizadora y
asignativa. Es decir, la descentralización de la políticafiscal, que surge como la receta óptima cuando el
principio bajo análisis es la asignación de recursos, presenta algunos contrapesos al centrar la atención
sobre los aspectos macroeconómicos.

B. La revisión de los argumentos históricos
Recientemente la tesis que asigna la responsabilidad de la política fiscal al gobierno nacional ha
sido cuestionada (Spahn, 1997; Tanzi, 1995). En el caso de una región o país grande, como la Unión
Europea o los Estados Unidos, los ciclos pueden no estar correlacionados. La ausencia de correlación
entre los ciclos regionales y agregados puede provenir del hecho de que cada subregión este especializada
en la producción de commodities cuyos precios son independientes de la situación macroeconómica del
país. En este caso, aunque el país no perciba la necesidad de realizar política alguna, puede ser óptimo
desde el punto de vista de los gobiernos subnacionales estimular la economía local a pesar del reducido
tamaño del multiplicador.
Además, los argumentos históricos suponen una economía cerrada. El esquema IS-LM
incorporando los aportes de Mundell indica que en una economía con tipo de cambio flexible todo
estímulo fiscal sería compensado con una variación del tipo de cambio. Sin embargo, esta ineficacia
agregada de la políticafiscal no es válida a nivel regional dado que si se proporciona una respuesta fiscal
con variaciones regionales pero de suma cero, de modo de no alterar el tipo de cambio, se pueden
conseguir efectos sobre el empleo relativo de las distintas regiones.
En el caso de los salvatages existe una controversia de sí una mayor centralización o
descentralización de la políticafiscal aumenta o disminuye la probabilidad de su ocurrencia.
En un sentido, el incentivo a rescatar opera a través de las externalidades, análogamente al
argumento "too big to fail". Un gobierno subnacional de mayor tamaño puede causar un mayor costo
externo sobre las demás jurisdicciones -incrementando la prima de riesgo de la nación y de los demás
gobiernos subnacionales, por ejemplo-, y por lo tanto incrementar los daños potenciales para el gobierno
nacional de no rescatar (Wildasin, 1997). Debe advertirse que con este criterio, la solución consiste en
fomentar aun una mayor descentralización, de modo de reducir el tamaño de las jurisdicciones y las
externalidades que estas pueden generar.
En el otro sentido, el incentivo a rescatar opera a través del costo relativo del salvataje en relación
al tamaño del presupuesto del gobierno nacional. Si existe un costo político fijo de dejar caer una
jurisdicción entonces el gobierno nacional tendrá más incentivos a rescatar a aquellas jurisdicciones en
que el costo del salvataje sea reducido en relación al tamaño del presupuesto del gobierno nacional. De
este criterio se desprende que un alto tamaño relativo de los gobiernos subnacionales hace más creíble la
condición de no salvataje (Eichengreen et al, 1996).
Otra falencia de los argumentos históricos es que sólo abordan las consideraciones de la demanda.
Si se incorporan al análisis las cuestiones de oferta puede arribarse al resultado de que los gobiernos
subnacionales poseen mejor información y generalmente responden con mayor rapidez a las necesidades
locales.

Sesión 1 - 1

Por último, existe evidencia que indica que la política monetaria se encuentra más desvinculada de
la política fiscal en los casos en que esta última se encuentra más descentralizada, lo que implica la
posibilidad de una mayor disciplina fiscal (Euther et al, 1996). Debe notarse que esta proposición no
resulta trivial por cuanto es posible que el gobierno nacional emita dinero para realizar un rescate de los
gobiernos subnacionales. Es decir, es posible que, en un contexto de altos costos de no rescatar a un
gobierno subnacional, el gobierno nacional se vea obligado a emitir dinero parafinanciar un rescate. En
cambio, si la política fiscal está centralizada, entonces el gobierno nacional controla plenamente la
decisión de gastar y emitir parafinanciar dicho gasto.
En este último caso se advierte una disputa entre los argumentos tradicionales y los más recientes
que está basada sobre un mismo hecho: el vínculo entre la política monetaria y lafiscal. Si se considera
que la políticafiscal debe ser compensatoria, el gobierno nacional debería concentrar la potestad sobre la
política monetaria y fiscal, de modo definanciar el déficit con emisión. Si, en cambio, el objetivo de la
política fiscal es mantener la disciplina fiscal, entonces existen argumentos en pos y en contra de
centralizar la política fiscal.
C.

El endeudamiento de los gobiernos subnacionales. La teoría de los Bailouts

La deuda de los gobiernos subnacionales posee principalmente dos características idiosincráticas
que lo diferencian de los demás tipos de endeudamiento. En primer lugar, el mercado de crédito se
encuentra distorsionado como consecuencia de la relación vertical entre los gobiernos locales y el
gobierno nacional. Esta relación vertical hace que, bajo determinados contextos institucionales, la
probabilidad de rescate ante un default de un gobierno subnacional pueda resultar elevada. En segundo
lugar, las conductas fiscales de cada gobierno local generan efectos horizontales, sobre sus pares, y
verticales, sobre el gobierno nacional. En particular, los comportamientos de los gobiernos poseen efectos
externos no computados por cada uno de ellos.
Uno de los elementos que ocasiona una distorsión en el mercado de deuda provincial es la
anticipación de un rescate a una provincia por parte del gobierno nacional. En esa sección se pretende
interpretar a estos potenciales bailouts como el resultado esperado de un esquema institucional inadecuado
de la relación vertical entre el gobierno central y los gobiernos subnacionales.
De manera abstracta, los bailout pueden ser analizados como el equilibrio de un juego dinámico
con información imperfecta, en el cual, primeramente el gobierno nacional se compromete a no rescatar.
Luego el gobierno subnacional decide una política fiscal dentro de un continuo de posibilidades que
incluyen opciones de política que van desde políticas que tienen bajo riesgo de culminar en un default
provincial (prudentes) a políticas en que con alta probabilidad no se pueda honrar los compromisos
asumidos (imprudentes). Posteriormente, se produce una realización de la naturaleza que, en función de la
política más prudente o imprudente que halla desarrollado la provincia, determina si dicho gobierno
subnacional cae o no en default. Por último, el gobierno central, ante la demanda provincial de un bailout,
decide si rescata o no.
En este juego la decisión del gobierno subnacional es secuencialmente racional, en el sentido de
que la elección de una política fiscal más prudente o imprudente, se toma considerando que respuesta
tendrá el gobierno central si la provincia cae en default. La restricción presupuestaria del gobierno
subnacional será blanda si, a pesar de que el gobierno subnacional adopte un comportamiento fiscal
imprudente, el gobierno central decide rescatar en última instancia. Es decir, si la percepción de las
provincias es que en última instancia el gobierno nacional dará una ayudafinanciera, las provincias tienen
incentivos a aumentar su exposición alriesgocon políticas imprudentes. Cabe destacar que la restricción
presupuestaria se vuelve menos restrictiva en la medida que el sectorfinanciero advierta que el gobierno
nacional se hará cargo de dichos compromisos.

Sesión 1 - 1

Debe notarse que a partir de esta estilización un bailout no es la ayuda financiera del gobierno
central a un gobierno subnacional que enfrenta un fuerte shock negativo, sino que además esta ayuda debe
ser consecuencia de un comportamientofiscal perverso que, previniendo el rescate posterior, incrementa
la probabilidad de sufrir una crisisfinanciera. Es decir, para que una ayudafinanciera del gobierno central
sea considerada un bailout, el gobierno subnacional previamente debe haberse comportado
imprudentemente en términosfiscales, de modo de incrementar su exposición a los shocks.
En cambio, si la ayudafinanciera no es antecedida por un comportamiento fiscal imprudente,
entonces dicha ayuda puede considerarse como una respuesta óptima dada la imposibilidad de escribir
contratos completos. Por ejemplo, el episodio de una ayudafinanciera no contemplada del gobierno
central ante una catástrofe no puede considerase un bailout, dado que el gobierno subnacional no posee
control sobre estos eventos. En cambio, esta ayuda puede ser el resultado de un contrato implícito entre
ambos niveles de gobierno.
El problema surge debido a que en muchas oportunidades no resulta fácil determinar si los
problemas financieros son consecuencia de comportamientos oportunísticos de los gobiernos
subnacionales que se exponenriesgosamente,sabiendo que en última instancia serán rescatados por el
gobierno central, o de fuertes shocks reales que a pesar de un sano comportamiento fiscal por parte del
gobierno local le generen una crisis financiera.

III.

El enfoque de la disciplina del mercado crediticio

La existencia de un mercadofinanciero provee importantes beneficios a los agentes económicos
que en él participan. Por ejemplo, posibilita a los consumidores elegir un apropiado sendero de consumo a
lo largo del ciclo de vida, permite a lasfirmas superar los obstáculosfinancieros de tener que solventar sus
proyectos de inversión con ingresos propios, y libera al gobierno de realizar costosos cambios en las
alícuotas impositivas para financiar los mayores gastos en períodos en que lo considere oportuno (permite
el tax smoothing).
Para lograr esta tarea, en un mundo de información perfecta, el mercadofinanciero debe permitir
que los agentes mantengan desequilibrios temporarios entre sus ingresos y sus egresos. Es decir, el
mercado debe proporcionar la liquidez necesaria a aquellos agentes que ex-post cumplan con sus
obligaciones, y por lo tanto satisfagan su restricción presupuestaria intertemporal. En un mundo de
información imperfecta el problema surge debido a la imposibilidad de determinar con certeza que los
desequilibrios sean sólo temporarios. En este sentido, antes de que se produzca la realización, sólo se tiene
una expectativa de que el prestatario cumpla con su obligación y confirme que el desequilibrio fue sólo
temporario. Sin embargo, puede ocurrir que la expectativa no sea convalidada posteriormente y el
prestatario se encuentre transitando en un sendero no sostenible.
Una de las soluciones postuladas por la literatura para resolver el problema antes señalado
proviene del enfoque del mercado disciplinador (Lañe, 1993; Bayoumi et al, 1995). Aplicado al caso de
los gobiernos subnacionales este enfoque postula que los agentes económicos deben pagar tasas de interés
que reflejen su situación de solvencia, de manera que aquellos agentes que incurran en deudas que sólo
con dificultad puedan ser servidas, enfrentarán tasas superiores para compensar el mayor riesgo de default,
lo que a su vez reduce el incentivo a endeudarse.
A su vez, uno de los elementos centrales de la hipótesis de disciplina de mercado es la existencia
de una relación no lineal entre el costo del endeudamiento y las variablesfiscales (Bayoumi et al, 1995).
En particular, se asume que el costo del crédito crece a tasa creciente con el nivel de endeudamiento, de
modo que si la solvencia del prestatario se torna muy riesgosa, el mercado le restringirá el acceso al
crédito. En otras palabras, el mercado controlará que sólo se le otorguen créditos a deudores solventes,

Sesión 1 - 1

induciendo a que los agentes desarrollen un comportamiento prudente e imponiendo el cumplimiento de
sus restricciones presupuestarias intertemporales2.
En la literatura referida al logro de la disciplina fiscal apelando al mercado se postulan un
conjunto de condiciones necesarias para que el mercado sea un mecanismo eficaz para inducir un manejo
responsable de lasfinanzas públicas provinciales. Entre las más importantes se destacan:
a) Los mercadosfinancieros deben ser razonablemente abiertos, de modo que cuando la solvencia
de un determinado prestatario sea dudosa, los prestamistas no se encuentren cautivos y tengan la
posibilidad de colocar su dinero en usos alternativos.
b) No deben existir prestatarios con privilegios en cuanto al acceso al crédito. En particular no
deben existir regulaciones que privilegien al sector público en el acceso al crédito.
c) Debe existir información sobre la solvencia de los demandantes de crédito. Si esta no está
disponible se presentan dos peligros: que se le otorguen fondos a prestatarios insolventes o que se
le racione el crédito a prestatarios solventes por emitir señales similares a las de deudores que no
hayan cumplido con sus obligaciones.
d) El sistemafinanciero debe poseer una fortaleza suficiente para poder asimilar sin problemas la
caída de un gran deudor.
e) No debe existir la certeza ni la percepción de que el prestatario vaya a ser rescatado si no puede
servir sus deudas.
Si no se satisfacen estas condiciones, el mercado no funcionará adecuadamente para prevenir un
excesivo nivel de endeudamiento. Es decir, a pesar de una mala performance económica y presupuestaria
provincial, acompañada por una suba insuficiente de las primas por default, la deuda puede continuar
incrementándose hasta alcanzar niveles insostenibles, implicando una severa y costosa restricción fiscal
futura.
Cabe destacar que estas condiciones son muy restrictivas y en general no se cumplen en la
mayoría de los países, principalmente en los países en vías de desarrollo, que no poseen un maduro
mercado de capitales. En particular, estos países no cuentan con sistemas de información que permitan
monitorear adecuadamente los comportamientos fiscales de los distintos gobiernos subnacionales y
muchos países instrumentan incentivos para colocar a los gobiernos locales en una posición ventajosa en
cuanto al acceso al crédito. Además, en la experiencia internacional no resulta difícil encontrar casos en
que los gobiernos centrales intervienen de modo de evitar un default de un gobierno subnacional3.

IV.

El uso de controles internos al sector público

La alternativa a la disciplina de mercado es establecer algún mecanismo de control interno al
sector público. En este caso se puede apelar a cinco variantes: a las reglas fiscales explícitas, a los
procedimientos presupuestarios, a la cooperación y coordinación de las decisiones de endeudamiento entre

Utilizando datos de panel sobre las tasas pagadas por los gobiernos subnacionales en los Estados Unidos y Canadá, Goldstein y
Woglom (1992) y Bayoumi, Goldstein y Woglon, (1995) h n encontrado que el mercado constituye u eficaz disciplinador para
a
n
los gobiernos subnacionales. Dichos trabajos analizan la relación entre los spreads de los bonos de varios gobiernos
subnacionales americanos y ciertas variablesfiscales, relacionadas con la solvencia y disciplina fiscal
3 En el capitulo VI se analizará, para el caso argentino, si el mercado crediticio disciplina a los gobiernos subnacionales.
2

Sesión 1 - 1 0

los distintos niveles de gobiernos, a los controles directos y a los controles administrativos o burocráticos
por parte del gobierno central.4
Cabe destacar que ningún país adopta una forma pura dentro de esta tipología, sino que en general
el control se efectúa mediante una combinación de mecanismos. Sólo en algunos casos es posible
identificar con certeza cual es el mecanismo utilizado.
Sin embargo, a pesar de esta dificultad, esta estilización resulta útil para racionalizar cuales son las
características preponderantes de la experiencia de los distintos países en el control de la deuda. Además,
el análisis de las virtudes de cada instrumento de control resulta indispensable si se pretende avanzar en el
diseño de una nueva arquitectura institucional para la coordinación de los comportamientos fiscales
subnacionales en la Argentina. A continuación analizaremos los principales aspectos de esta clasificación5.

A. Reglas predeterminadas
Aunque más recientemente este mecanismo ha adquirido una gran difusión debido a su
incorporación en el tratado de Maastricht, el control mediante reglas no constituye un mecanismo nuevo.
En efecto, la constitución y las leyes de muchos países establecen reglas que limitan directamente el
endeudamiento de los gobiernos locales, o bien distintas variables relacionadas al endeudamiento.
El estudio de la efectividad de las reglas constituye en la actualidad un activo campo de análisis
tanto teórico como empírico.6 En general, estos estudios, basados en análisis de corte transversal y en
datos de panel, sugieren que existe una asociación significativa y diferente entre los distintos tipos de
metas aplicadas sobre los resultadosfiscales y los resultadosfiscales finalmente observados.
Es menester destacar que existen distintas formas de limitar el endeudamiento por medio de
reglas. Se pueden imponer restricciones sobre el endeudamiento en sí mismo, limitando el monto o bien el
destino del endeudamiento, o se puede limitar el endeudamiento indirectamente, fijando metas sobre
variables que inciden directamente sobre el endeudamiento (déficit, por ejemplo) o que están íntimamente
relacionadas al endeudamiento (servicios de la deuda, por ejemplo).
Un tipo especial de regla predeterminada son los presupuestos equilibrados. De manera estilizada,
la fortaleza de esta restricción puede ser de cuatro niveles, dependiendo del estadio del proceso
presupuestario que regule. En primer lugar, se puede entender el equilibrio presupuestario como la
obligación de que el Ejecutivo presente al Congreso un proyecto de presupuesto equilibrado. En segundo
término, el equilibrio presupuestario puede significar que el Congreso deba aprobar un presupuesto
equilibrado. En tercer lugar, es posible que la regulación no sólo establezca la sanción de un presupuesto
equilibrado, sino que además se obligue al cumplimiento de esta meta en la ejecución del presupuesto por
parte del Poder Ejecutivo, aunque se permita que los desvíos entre la ejecución presupuestaria y el
presupuesto sean compensados en ejercicios posteriores. Por último, la versión más fuerte y restrictiva del
concepto impide que los desvíos sean compensados en ejercicios futuros, imponiendo que los ajustes
deban efectuarse en cada período. Además, esta regla no sólo puede ser mas laxa o másrígidaen función
del estadio presupuestario que limite, sino que muchas veces lo que difiere es la amplitud del sector
público que comprometen. Sólo una definición amplia del sector público hace que el espíritu de la
Latipologíaadoptada en Ter-Minassian (1997) incorpora cuatro clasificaciones: Ccooperación entre distintos niveles de
gobierno, reglas, controles directos y disciplina de mercado. Aquí se agrupan las tres primeras variantes, incorporando dos
nuevas categorías (procedimientos presupuestarios y controles burocráticos) dentro de la categoría que incluye a los controles
directos del sector público. Sin embargo, al analizar los controles internos al sector público, cada categoría se analizará de
manera discriminada.
5 Para u análisis completo de la problemática del endeudamiento subnacional véase Ter-Minassian (1997) y Monasterio et al
n
(1999).
6 Para u análisis integral, véase Poterba y Von Hagen (1999).
n
4

Sesión 1 - 1

restricción sea cumplido. De otro modo, si la restricción se aplica sólo a una parte del sector público,
existe elriesgode que el endeudamiento se acumule sobre los organismos no comprendidos en la norma7.
Debe notarse que la regla presupuestaria no resulta óptima, dado que una regla que impone el
equilibrio presupuestario período a período obliga a recortar el gasto y/o aumentar los impuestos en los
momentos de depresión económica. Dicho patrón de política económica no respeta ni las prescripciones
de la teoría "keynesiana standard', ni las prescripciones que surgen de la teoría de "tax smoothing".
Sin embargo, en algunos trabajos se ha encontrado que los países que han evidenciado déficits
persistentes han sido menos capaces de equilibrar intertemporalmente el consumo que los países con
déficit reducidos (Arreaza et al, 1999). El hallazgo sugiere, que las instituciones que limitan el déficit
mejoran la capacidad de realizar una política macroeconómica adecuada. De hecho muchos argumentos
indican que la política fiscal se rige más por principios de naturaleza política que por prescripciones de
optimalidad. Bajo esta concepción, el beneficio de la regla proviene en mayor medida de la limitación de
la negociación política.
Pero, lo que estos trabajos sugieren es que el presupuesto equilibrado puede mejorar la
performance de la política macroeconómica en relación a la ausencia de alguna otra regla. Lo que no se
demuestra es si mediante otras reglas puede alcanzarse un resultado simétrico en términos de la limitación
de la negociación política pero superior en términos de optimalidad. Desde ya que estas reglas alternativas
deben ser simples, lo cual constituye uno de los pilares del debate "reglas versus discreción". De otro
modo, las mismas pueden ser difíciles de monitorear o esquivadas mediante la "contabilidad creativa".
Otras reglas utilizadasfrecuentemente en la experiencia internacional son el establecimiento de
límites al gasto público, al cociente entre deuda y producto, al cociente entre la deuda y los recursos y al
cociente entre los servicios de la deuda y los recursos. En los dos últimos casos las reglas tratan de imitar
el comportamiento del mercado, en el sentido de que se permite el endeudamiento de acuerdo a la
capacidad de repago de la misma.
Una regla que es utilizada en muchos casos (por ejemplo, Alemania, Suiza y en muchos estados
de Estados Unidos) es la llamada regla de oro, la cual limita el destino del endeudamiento, al
financiamiento de los gastos de capital. El objetivo de esta restricción consiste en impedir traspasar la
carga de la deuda a generaciones futuras.8 La argumentación es que en el caso de la deuda originada por
gasto corriente la generación presente disfruta de los beneficios del gasto pero el costo de este gasto es
afrontado por las generaciones futuras. En cambio, en el caso de los bienes de capital, si bien parte de la
deuda es amortizada por las generaciones futuras, también la generación futura goza de un mayor stock de
capital.
El control mediante reglas posee ventajas y desventajas. Entre las ventajas se cuentan la
transparencia y equidad en el tratamiento de los distintos gobiernos subnacionales. Este mecanismo de
control no demanda negociaciones periódicas, que pudieran culminar en resultados inequitativos en el
tratamiento de las distintas regiones. A su vez, ante determinados contextos institucionales, estas
negociaciones podrían ser orientadas más por factores políticos de corto plazo que por el mantenimiento
de niveles sostenibles de endeudamiento.

Kenyon y Benker (1984) encontraron para los Estados Unidos que estas limitaciones sólo comprendían a una media del 6 % del
0
gasto total
8 Debe advertirse que en el caso de que la deuda pública sea solamente interna tanto los deudores c m los acreedores pertenecen
o o
al ms o país. En este caso, si bien la generación futura deberá afrontar la deuda, también la generación futura recibe los pagos
im
por amortización de este endeudamiento. Esto revela la falacia demostrada por los economistas clásicos Ricardo y Pigou. Es
decir, e el caso de que la deuda sea interna no es posible trasladar la carga de la ms a a generaciones futuras, sea esta
n
im
originada en gastos comentes o de capital. Este endeudamiento sólo genera efectos distributivos intrageneracionales. Sin
embargo, en el caso de que los acreedores sean externos sí es posible trasladar la carga de la deuda a generaciones futuras. En
este caso el endeudamiento originado en gastos comentes y e gastos de capital tiene efectos distributivos intergeneracionales.
n
7

Sesión 1 - 1

El principal costo es que las reglas imponen una pérdida deflexibilidad para ajustarse a los shocks
sobre la recaudación. Este costo se ve agravado en el caso de las reglas que impiden el endeudamiento,
como es el caso de la regla que obliga a mantener el presupuesto equilibrado.

B. Procedimientos presupuestarios
Una alternativa al control mediante reglas numéricas sobre variablesfiscales es diseñar el proceso
presupuestario de forma tal que de la interacción en este proceso se arribe a un resultado de
endeudamiento deseado. En este caso se pretende establecer el contexto, o las instrucciones del juego de
negociación por el cuál se decide el comportamientofiscal de los distintos estados, confiando en que el
resultadofinal no arroje un nivel de endeudamiento excesivo.
En relación al contexto de la negociación se presentan dos opciones: verticalista y colegiada.9 En
el primer caso, el mayor poder en la negociación del presupuesto se le asigna al Poder Ejecutivo con
relación al Legislativo, y al ministro definanzas en relación a los demás ministros. El poder legislativo
tiene limitada capacidad de influir sobre la asignación y el tamaño del presupuesto. En la segunda opción,
el presupuesto surge de una negociación más democrática, en la cual el Poder legislativo tiene la
capacidad de influir en las decisiones y el ministro definanzas se encuentra en una posición similar a los
demás ministros.
Las características de estos dos modelos institucionales estilizados generan resultados
presupuestarios distintos. De hecho, existe un intercambio entre estos modelos. En el caso verticalista
resulta más probable que se arribe a un resultado fiscal equilibrado, que no comprometa la restricción
fiscal y que, por lo tanto, haga del presupuesto un eficaz método de control. Este beneficio se logra al
costo de un menor respeto a la heterogeneidad y a la oposición minoritaria. En el modelo colegiado el
procedimiento presupuestario es más democrático, garantizando una mayor incumbencia de las minorías.
El mayor poder relativo de las minorías resulta en un menor énfasis en la disciplina fiscal.
El carácter colegiado o verticalista de la negociación puede provenir de que el sistema electoral
sea proporcional o mayoritario. Esto se debe a que los sistemas electorales proporcionales tienden a
generar gobiernos de coaliciones y los sistemas mayoritarios incentivan el bipartidismo. En los gobiernos
de coaliciones se deben consensuar las decisiones entre los distintos participes de la coalición. Dicho
consenso otorga un carácter colegiado a la negociación, dejando el poder de veto en cada miembro de la
coalición.
Una cuestión que resulta crucial en el caso de los procedimientos presupuestarios es el diseño
institucional de la votación.10 Por ejemplo, en 1974, los Estados Unidos reformaron la secuencia de la
votación de la ley presupuesto por parte del congreso. A partir de ese año el tamaño global del presupuesto
dejó de constituirse meramente por el resultado al que se arribefinalmente de la suma de cada uno de los
proyectos aprobados y pasó a predeterminarse con anterioridad al tratamiento en particular de cada
proyecto. Esto es, se dividió la votación presupuestaria en dos etapas, definiéndose en la primera etapa
sólo el tamaño del presupuesto y en la segunda la asignación del mismo a los distintos proyectos
competitivos.
El objetivo de este cambio institucional fue el de garantizar el equilibrio presupuestario. Sin
embargo, los grandes déficits en las Estados Unidos en los años ochenta generaron una reconsideración
acerca de los efectos que genera este mecanismo.
9

Alesina et al (1999) llaman a estas estilizaciones hierarchica y collegial.
El teorema de la imposibilidad de Arrow indica que el orden de la votación altera el resultado del escrutiño. Si no se i p n n
moe
restricciones e los procedimientos, si se carece de una estructura y reglas de votación, entonces se producirá un caos
n
legislativo.

10

Sesión 1 - 1

A partir de estos estudios se arribó a la conclusión de que no resulta obvio que este mecanismo
deba garantizar el equilibrio presupuestario. La intuición de este resultado es que los legisladores son
forward looking. Si cada legislador conoce perfectamente las preferencias de los demás legisladores,
entonces basa su decisión del tamaño del presupuesto realizado en la primera etapa computando el
resultado sobre la composición del presupuesto a que se arribará en la segunda etapa. Es decir, las
decisiones son secuencialmente racionales.
Otra característica del sistema de votación que puede afectar el resultado presupuestario es la
exigencia de mayorías simples o reforzadas para la aprobación del presupuesto. La exigencia de mayorías
especiales puede desembocar en un vacío institucional si el congreso rechaza el presupuesto, pero, a la
vez, estas mayorías especiales otorgan un carácter de consenso a las decisiones.
Es importante notar que las mayorías necesarias para la aprobación del presupuesto también
otorga al contexto de negociación un carácter más colegiado o uno más verticalista.
Una cuestión relevante acerca del procedimiento presupuestario es el grado en que el Poder
Legislativo tiene la capacidad de realizar enmiendas al presupuesto. Bajo un sistema de reglas cerradas el
Poder Legislativo tiene la capacidad de veto, es decir puede votar favorablemente o en contra del
presupuesto planteado por el Poder Ejecutivo que es quien plantea la agenda. La alternativa a las reglas
cerradas es un sistema de reglas abiertas. Bajo este sistema el Poder Legislativo puede discutir y enmendar
cada punto particular del proyecto de presupuesto planteado por el Poder Ejecutivo.
En cuanto al control presupuestario por parte del Poder Ejecutivo una opción aplicada en algunos
estados de los Estados Unidos es la llamada Une item veto. Dicho mecanismo otorga al Poder Ejecutivo el
poder de veto sobre algunas partidas presupuestarias, limitando los efectos de las presiones de los grupos
de interés sobre el Poder Legislativo.

C. Cooperación y coordinación entre los distintos niveles de gobierno
En algunos países la forma de control de endeudamiento utilizada consiste en coordinar las metas
de endeudamiento mediante acuerdos de cooperación entre los distintos niveles de gobierno. En este
mecanismo los gobiernos subnacionales son copartícipes con el gobierno central en la definición de las
metas macroeconómicas. Este hecho hace que los gobiernos subnacionales no sólo se sientan partícipes de
los objetivos macroeconómicos, sino también que sean responsables de las consecuencias que un
endeudamiento excesivo podría acarrear. La cooperación y coordinación del endeudamiento constituye la
forma de control más respetuosa de la autonomía que deberían gozar los gobiernos locales en un sistema
federal de gobierno (Gandenberger, 1997).
Sin embargo, este tipo de control parece más adecuado en países que posean una alta disciplina
fiscal. Si esto no sucede, el proceso de negociación puede desembocar en conductasfiscales insostenibles
(Ter-Minassian, 1997). En particular, si no existe una fuerte imputabilidad sobre todos los actores
involucrados de los costos que puede generar un alto nivel de endeudamiento, se corre elriesgode que el
endeudamiento sea percibido como un recurso común, al cual todos los actores involucrados en la
negociación intenten depredar.
Los mecanismos de negociación de endeudamiento pueden ser bilaterales o multilaterales. En el
primer caso, del cual puede citarse como ejemplo Dinamarca, la negociación se realiza entre el gobierno
central y cada gobierno local aisladamente. En el segundo, la discusión se desarrolla en un ámbito en el
cuál se negocian, simultáneamente, los límites del gobierno central y de todos los gobiernos
subnacionales.
Si los acuerdos son bilaterales se corre elriesgode que los gobiernos subnacionales no reciban un
trato equitativo. Este hecho, a su vez, puede desembocar en que los gobiernos subnacionales que no
Sesión 1 - 1

satisfagan las metas prefijadas aleguen que su incumplimiento se explica por un trato no equitativo en la
negociación.
Australia es un país en el cual la negociación es multilateral y cooperativa. En dicho país, las
negociaciones de endeudamiento y la consistencia de este con los objetivos macroeconómicos son
desarrolladas en el ámbito de una institución federal llamada Loan Council, en la cual están representados
todos los estados y el gobierno nacional.
En el caso de Alemania, las negociaciones también poseen un carácter multilateral. La
negociación se realiza en dos etapas. En la primera sefija el endeudamiento agregado del gobierno central
y de los gobiernos locales, y se distribuye verticalmente que porcentaje de este endeudamiento puede ser
tomado por los gobiernos locales y quefracción por el gobierno central. En la segunda etapa, la fracción
de la deuda pública asignada a los gobiernos subnacionales se distribuye horizontalmente entre estos.
En España, a partir de 1992, existe también un procedimiento en dos etapas similar al caso
alemán. Sin embargo, en este caso la negociación es bilateral entre el gobierno central y cada comunidad
autónoma. Los límites al endeudamiento para cada comunidad se fijan en los llamados "Escenarios de
Consolidación Presupuestaria".

D. Controles directos
En algunos países, principalmente en aquellos que poseen una estructura unitaria de gobierno, el
nivel Nacional controla directamente el endeudamiento de los gobiernos locales. Dado que los gobiernos
locales deben pedir autorización al gobierno central para endeudarse, este tipo de control resulta el modo
más efectivo para limitar el excesivo endeudamiento. Sin embargo, se contrapone a esta virtud el hecho de
que bajo este sistema de control se restringe excesivamente la necesaria autonomía de los gobiernos
subnacionales en un esquema descentralizado. Esta desventaja provoca que difícilmente se adopte un
sistema de controles directos en un país federal.
Los controles directos pueden adoptar distintas formas. Estas pueden ser: límites al
endeudamientofijados anualmente para cada gobierno subnacional, centralización del endeudamiento por
parte del gobierno central y posterior repaso de los créditos a los gobiernos subnacionales, examen y
autorización de las operaciones de crédito que deseen desarrollar los estados subnacionales por parte del
gobierno central, prohibiciones al endeudamiento en determinados tipos de deudas, como por ejemplo
deudas con el exterior. En este último caso, el argumento es que este tipo de endeudamiento puede
comprometer los objetivos de la política monetaria y cambiaría desairolladas por el gobierno central.
Un país unitario que aplica el control directo de endeudamiento de sus gobiernos subnacionales es
Japón. Las pautas anuales de endeudamiento se negocian en un planfinanciero anual, denominado "Plan
Fiscal Anual de los Gobiernos Locales" que se aprueba por el parlamento nacional simultáneamente con la
aprobación del presupuesto nacional. Chile es otro país unitario que restringe directamente el
endeudamiento de los gobiernos locales. En Chile las municipalidades tienen vedado el acceso a los
mercados de crédito interno y extemo, salvo autorización expresa del gobierno central (Yañez et al, 1993).
Un país federal que instrumenta el control directo de endeudamiento es India. En este país los
gobiernostienenvedados los mercados internacionales de crédito. A su vez, en el caso del endeudamiento
interior, los estados que mantienen deudas con el gobierno nacional deben solicitar autorización en el caso
de que deseen endeudarse. Cabe destacar que todos los estados se encuentran endeudados con el gobierno
nacional. Además, en la India, el gobierno central ha regulado el sistemafinanciero de forma tal que
mantiene un mercado cautivo de colocación de deuda, el cual le brinda un sistema adicional de control,
dado que el acceso a este mercado esta reservado para emisiones aprobadas por el gobierno central.

Sesión 1 - 1

E. Control burocrático
Es usual que los países diseñen mecanismos de control burocrático. En estos casos los gobiernos
centrales articulan procedimientos que los gobiernos subnacionales deben satisfacer para endeudarse. Los
procedimientos consisten en el cumplimiento de ciertos requisitos exigidos por un determinado organismo
del Gobierno Nacional. Estos requisitos pueden incluir la presentación de información acerca del estado
de lasfinanzas públicas del gobierno subnacional, acerca de las características del endeudamiento en
términos de tasas, amortizaciones, etc. y acerca del destino de los fondos.
Además, en algunos casos, el cumplimiento de estos trámites puede resultar necesario para que los
prestatarios accedan a ciertos privilegios otorgados a los créditos con los gobiernos subnacionales, lo cual
constituye un incentivo para que estos se cumplimenten.
Este método de control resulta más utilizado en el caso de algunas democracias federales, en las
cuales el gobierno nacional no tiene la prerrogativa de restringir directamente el endeudamiento de los
gobiernos subnacionales y no existe una instancia institucional en la que los representantes de cada nivel
de gobierno acuerden ciertos estándares o procedimientos en lo referido al crédito público. Cabe destacar
que como este tipo de control generalmente es paralelo a algún otro mecanismo más restrictivo, en muchas
oportunidades no se lo considera como una alternativa en sí misma. Sin embargo, en el caso argentino los
procedimientos administrativos para el endeudamiento constituyen un elemento esencial, razón por la cual
esta alternativa de control se incorpora separadamente.
Debe notarse que puede confundirse este tipo de control, con los controles directos, especialmente
cuando los gobiernos subnacionales deben pedir autorización para endeudarse, como en el caso Argentino.
La diferencia en este caso es de grado y no de tipo. Es decir, si bien bajo los controles burocráticos y
administrativos el objetivo es fundamentalmente informar acerca de las operaciones de crédito y
cumplimentar una serie de procedimientos administrativos, en el caso de los controles directos los
gobiernos subnacionales, o bien tienen prohibo el acceso a determinados tipos de endeudamiento, o bien
deben solicitar autorización del gobierno central si pretenden endeudarse.

V.

El endeudamiento de las provincias Argentinas
A. El contexto institucional

La organización del sector público argentino contempla cuatro niveles de gobierno: el federal, las
provincias (con sus 23 jurisdicciones), el Gobierno de la Ciudad de Buenos Aires y más de 1.300
gobiernos locales incluyendo municipios, comunas, comisiones municipales, juntas de gobierno,
comisiones de fomento y comunas rurales. En la constitución sancionada en 1853 se adoptó la forma
republicana y federal de gobierno, previendo un sistema con tres poderes: un poder ejecutivo a cargo de
un presidente, un poder legislativo compuesto por un congreso bicameral y un poder judicial, presidido
por la Corte Suprema de Justicia. En dicha constitución las provincias delegaron al gobierno federal una
serie de poderes y atribuciones, incluyendo entre otras cosas, la potestad sobre la recaudación de ciertos
impuestos, aunque con ciertas limitaciones. Las provincias tienen sus propias constituciones, algunas
sancionadas con anterioridad a la Nacional, y eligen su propio gobernador, sus legisladores y jueces sin la
intervención del gobierno federal.
La constitución argentina establece en su articulado las competencias y funciones asignadas a cada
nivel de gobierno. En particular describe los recursos e impuestos que cada jurisdicción puede establecer

Sesión 1 - 1

para poderfinanciar las funciones que presta. Básicamente, determina que la Nación utilizará para
financiar sus erogaciones los recursos del comercio exterior y los impuestos indirectos, mientras que las
provincias sefinanciarán mediante tributos sobre la producción y el consumo de bienes específicos.
Establece, además, la facultad al gobierno federal de imponer impuestos directos por tiempo determinado.
No obstante, la incorporación de nuevas funciones y responsabilidades en el gobierno federal
obligó a recurrir a nuevas fuentes de recursos. Esto derivó, primero, en la creación de nuevos impuestos en
concurrencia con los provinciales, y luego, el diseño de regímenes de coparticipación federal de
impuestos, mediante los cuales la Nación imponía gravámenes sobre el consumo, la renta y el patrimonio,
constitucionalmente reservados a las provincias, para luego compartirlos en una proporción fija con las
jurisdicciones que adherían.
Desde principio del año 1988 el marco general de las relacionesfinancieras entre la Nación y las
provincias se estructura sobre la base de la ley 23.548 de coparticipación federal de impuestos. Dicha ley
establece como impuestos coparticipables a todos aquellos tributos nacionales existentes y a crearse con la
excepción de los derechos de exportación e importación, los impuestos incluidos en regímenes especiales
de coparticipación y aquellos con asignación específica. La masa coparticipable se distribuye del siguiente
modo: a) 42,34% en forma automática a la Nación, b) 56,66% en forma automática al conjunto de las
provincias y c) el 1% restante se asigna al fondo de Aportes del Tesoro Nacional, destinado a atender
situaciones de emergencia y desequilibriosfinancieros en las provincias. La distribución secundaria de las
transferencias se determina por coeficientes fijos establecidos en la ley. Es importante señalar que aún
cuando la distribución primaria y secundaria no tiene en la norma una fimdamentación explícita, ésta
permitió darle a las relaciones financieras intergubernamentales un encuadre legal definido y transparente,
limitándose los aportes discrecionales que caracterizaron los años 80 (Jones, et al 1997).
Con la reforma constitucional del año 1994, el sistema de coparticipación federal de impuestos
adquiere rango constitucional.11
En términos del gasto público, Argentina constituye el país más descentralizado de América
Latina, con aproximadamente el 48% de los gastos ejecutados desde el nivel subnacional de gobierno12.
En contraste, la distribución de recursos tributarios se encuentra sumamente concentrada en el nivel
federal de gobierno, ya que cerca del 70% de los recursos tributarios son generados por dicho nivel. Esta
distribución de funciones y potestades lleva a un alto grado de desequilibrio fiscal vertical que debe ser
cubierto mediante mecanismos de transferencias de recursos desde el gobierno federal. El principal
mecanismo de transferencias lo constituye el anteriormente mencionado régimen de coparticipación
federal de impuestos. Por este sistema, durante 1999 el gobierno federal transfirió a las provincias y la
Ciudad de Buenos Aires más de $ 900 millones mensuales.

B. La evolución de lasfinanzas y el endeudamiento provincial durante el plan de convertibilidad
Luego de una década de fuerte inestabilidad macroeconómica, el la década de los noventa se
implementa un ambicioso programa de reformas económicas estructurales. Estas incluyen la reforma
tributaria, la privatización de todas las empresas públicas, la concesión de servicios públicos, una fuerte
reducción del déficit público, la descentralización de los servicios de salud y educación, el levantamiento
de restricciones a las inversiones extranjeras y la apertura y liberalización de la economía, entre otras
medidas. Asimismo, en abril del año 1991 se lanza un exitoso programa de estabilización -plan de
La nueva constitución establece en sus cláusulas transitorias la obligación de sancionar antes quefinalizara el año 1 9 u
96 n
nuevo esquema de coparticipación federal de impuestos. Vencido el plazo estipulado por la Constitución Nacional, en el año
1999, el gobierno federal intentó negociar con las provincias u proyecto de reforma, pero fracasó.
n
12 Si se excluyen del cálculo los gastos por seguros sociales (seguridad social y obras sociales) la participación de los gobiernos
subnacionales e el gasto público argentino trepa al 64%.
n
11

Sesión 1 - 1

convertiblidad-, por el cual sefija la paridad del dólar a un peso (equivalente entonces a 10.000 australes)
y se le prohibe al Banco Central la emisión de moneda sin respaldo. Por su parte, en septiembre de 1992
se dicta una nueva carta orgánica para el Banco Central que refuerza su autonomía del poder político. Con
relación a las provincias la nueva carta orgánica restringe la posibilidad de otorgar líneas de redescuento
(préstamos) a ios bancos provinciales13.
Como resultado de estas reformas económicas, la economía Argentina se estabiliza y logra un
crecimiento sólo interrumpido por la crisis del tequila y, más recientemente, por la crisis internacional. No
obstante esos episodios, el crecimiento de la economía alcanza un 6% anual promedio para el período
1991 y 1998.
Para los gobiernos provinciales el programa económico generó un sustancial crecimiento en los
recursos. Si consideramos solamente el primer año del plan de convertibilidad los recursos girados en el
marco del régimen de coparticipación federal de impuestos aumentaron en términos nominales un 22% -si
se consideran la totalidad de los recursos girados desde el gobierno federal el aumento en los ingresos
provinciales alcanza al 50%-. Aunque en menor magnitud, también se observó un incremento en los
ingresos tributarios propios (ver Cuadro 1).
Cuadro 1
Finanzas Públicas Provinciales y de la Ciudad de Buenos Aires. Período 1989-1999
(en millones de pesos corrientes^
1992
1993
1994
1996
1998
1999
1991
1995
1997

1989*

1990*

Ingresos corrientes
Tributarios propios
Otros ingresos propios
Transferencias nacionales

12,330
3,242
2,496
6,592

11,620
3,791
1,452
6,377

14,643
4,678
1,422
8,543

21,591
7,364
2,246
11,981

24,777
8,794
2,730
13,253

26,778
9,681
3,272
13,825

26,002
9,167
3,411
13,424

27,930
9,772
3,745
14,413

30,831
10,892
3,610
16,329

32,012
11,614
3,648
16,749

31,057
11,008
3,409
16,640

Gastos corrientes
Gastos de consumo
Intereses de la deuda
Transferencias corrientes

11,149
7,959
516
2,674

11,876
8,730
219
2,927

13,997
10,153
266
3,578

19,816
14,347
353
5,115

23,477
16,975
471
6,031

25,033
18,181
571
6,281

25,197
18,341
721
6,135

25,481
17,821
1,004
6,656

27,628
19,182
1,052
7,394

29,665
20,524
1,181
7,960

30,894
21,747
1,279
7,868

78

189

216

212

739

588

689

1,147

1,800

1,080

998

2,115

2,225

2,250

2,584

3,788

4,529

4,745

4,672

5,115

5,081

4,034

Ingresos totales

12,409

11,809

14,859

21,804

25,515

27,365

26,691

29,077

32,631

33,091

32,055

Gastos totales

13,264

14,101

16,247

22,400

27,265

29,561

29,942

30,153

32,743

34,746

34,928

-596
•596

-1,749
-2.064

-2,196
-2.364

-3,252
-3.480

-1,076
-1.724

-112
-1,377

-1,655
-2.233

-2,873
-3.340

Ingresos de capital
Gastos de capital

Resultado financiero:
con privatizaciones
-855
-2,292
-1,388
sin privatizaciones
-855
-2,292
-1.388
Nota: (*) Actualizados a pesos de abril de 1991
Fuente: Susbsecretaría de Programación Regional

Los gobiernos provinciales reaccionaron a este flujo de recursos incrementando en forma
significativa sus erogaciones, en especial, las de carácter corriente. En los primeros años del plan de
convertibilidad, este fuerte crecimiento en los recursos significó una disminución de la necesidad de
financiamiento de las provincias, aún cuando el crecimiento de las erogaciones había sido, también,
importante. A partir del año 1993 esta tendencia se modifica. Lasfinanzas provinciales comienzan a
transitar por un sendero de desequilibriofiscal que alcanza su máxima expresión en el año 1995 con un
déficit de $ 3.500 millones, como respuesta al fuerte impacto en las finanzas subnacionales ocasionado por
el efecto tequila.
En la década de los 8 era una práctica c m n el endeudamiento de los tesoros provinciales con sus propios bancos. Por el otro
0
o ú
lado, el Banco Central otorgaba líneas de redescuento a los bancos provinciales para otorgarle liquidez.

13

Sesión 1 - 1

Como es ampliamente conocido, la crisis desatada por la devaluación mexicana, en diciembre de
1994, impacto con fuerza en la economía Argentina. Durante el año 1995, la economía experimentó una
severa crisis bancaria y crediticia, que socavó la confianza de los inversores locales y extranjeros. Esto
provocó una importante fuga de capitales, la pérdida de reservas internacionales del Banco Central, un
aumento en la tasa de interés y una caída en el nivel de actividad de la economía.
La crisis definanciamiento puso de manifiesto la fragilidad de lasfinanzas públicas provinciales.
La brusca caída de los niveles de actividad -durante el año 1995 el PIB disminuyó un 4,4%-, la pérdida de
recursos tributarios tanto nacionales como provinciales, y la rigidez de los presupuestos provinciales
generaron un fuerte desequilibrio presupuestario. Esta situación se vio agravada en algunas provincias
como consecuencia de la crisis que azotó a la banca provincial producto de la fuerte desconfianza de los
ahorristas. Esta desconfianza desencadenó corridas bancarias en los bancos provinciales ocasionando
importantes pérdidas de depósitos que debieron ser cubiertas con recursos corrientes de la tesorería
provincial. Este factor agravó la crisis de liquidez de los gobiernos provinciales14. Como consecuencia de
estos hechos las provincias tuvieron serios problemas para hacerfrente a los pagos de salarios y haberes
jubilatorios. En algunas jurisdicciones se llegaron a acumular retrasos de entre cuatro y seis meses de
salarios a los empleados públicos.
Producto de esta crisis se produce, en el año 1995, un significativo aumento del endeudamiento
provincial. El stock de la deuda aumenta un 31%, pasando de $ 11.120 millones, en 1994, a prácticamente
$ 14.600 millones en 199515. Dicho stock de la deuda representa el 5,6% del Producto Interno Bruto de
1995 y el 55% de la totalidad de los ingresos provinciales de dicho año (ver Cuadro 2).

Cuadro 2
Deuda provincial. Evolución 1995-99
Período

en mili, de U$S

var. %

en % del PIB

en % de IT (1)

1994
1995
1996
1997(2)
1998
1999

11,120,400
14,572,868
15,644,445
14,822,582
15,786,618
16.745.720

31.0%
7.4%
-5.3%
6.5%
6.1%

4.3%
5.6%
5.7%
5.1%
5.3%
5.9%

40.6%
54.6%
53.8%
45.4%
47.7%
52.2%

(1) Ingresos Totales
(2) El stock de deuda en el año 1997 disminuye como consecuencia de la cancelación de
deuda de la Drovincia de Buenos Aires
Fuente: Elaboración propia en base a información de la Subsecretaría Programación
Rfininnal. rlfi las nrnnias nrnvinnias v nalifinadoras rift riesnn.

Una fuente habitual definanciamiento de los déficit provinciales ha sido el endeudamiento con los
propios bancos provinciales. Al respecto, información disponible sobre el stock de la deuda provincial
para 1996 indica que las provincias de Buenos Aires, Córdoba y la Municipalidad de Buenos Aires habían
Las provincias que se vieronfinancieramente m s afectadas por la crisis de sus bancos fueron Río Negro, Córdoba y Mendoza.
á
E los tres casos el gobierno federal debió asistirfinancieramente con recursos extraordinarios. Al respecto véase World B n
n
ak
(1996), Nicolini et al (1999) y Damill (1997).
15 En Argentina existen grandes dificultades con la información fiscal de las provincias. Con respecto a la deuda subnacional, la
información fue elaborada con datos de la Secretaría de Programación Económica y Regional, datos de las mismas provincias e
información de las calificadoras deriesgos.La información presentada no incluye deuda sin consolidar del Gobierno de la
Ciudad de Buenos Aires con el gobierno nacional por $ 5 0 millones.
10
14

Sesión 1 - 1

tomado préstamos de sus respectivos bancos por aproximadamente $ 3.000 millones (ver Cuadro 3). En el
año 1995 el stock de deuda con los bancos públicos y privados supera los $ 5.000 millones, representando
un 36,6% del stock de la deuda provincial (ver Cuadro 4).
Cuadro 3
Endeudamiento Provincial. Principales Entidades Bancarias acreedoras
1996
en miles de $

Entidad

en %

1997
en miles de $

1998
en % en miles de $

en %

5,982,698

100%

4,186,834

100%

5,041,041

100%

784,479
Galicia y Buenos Aires
441,381
Río de la Plata
BBV Banco Francés
551,478
247,351
Bansud
Banco de la Nación Argentina
26,065
Quilmes
181,775
Ciudad de Buenos Aires
497,557
Argemofin
140,336
Banco de la Provincia de Córdoba
181,216
Cargil Trading
60,083
Credit Lyonnais
101,206
Banco Hipotecario Nacional
74,659
BankBoston
66,113
Banco Provincia de Buenos Aires
2,044,019
Comafi
33,905
BERSA
25,006
Otros
526069
Fuente: Subsecretaría de Programación Regional

13.1%
7.4%
9.2%
4.1%
0.4%
3.0%
8.3%
2.3%
3.0%
1.0%
1.7%
1.2%
1.1%
34.2%
0.6%
0.4%
8.8%

1,459,167
527,673
538,419
294,522
18,733
73,566
244,700
143,844
99,507
123,654
89,904
69,795
69,809
20,570
65,623
21,365
325983

34.9%
12.6%
12.9%
7.0%
0.4%
1.8%
5.8%
3.4%
2.4%
3.0%
2.1%
1.7%
1.7%
0.5%
1.6%
0.5%
7.8%

1,627,986
664,379
615,009
332,493
227,072
213,572
202,854
163,966
162,354
130,685
75,867
69,478
66,595
59,200
55,532
36,018
337981

32.3%
13.2%
12.2%
6.6%
4.5%
4.2%
4.0%
3.3%
3.2%
2.6%
1.5%
1.4%
1.3%
1.2%
1.1%
0.7%
6.7%

Total préstamos

Cuadro 4
Deuda Provincial. Composición
(en millones de pesos y en % del total)
Período

Préstamos con
ent. financieras

Títulos públicos

Organismos
Internacionales

Deuda
Flotante

Resto (1)

Total

1994
1995
1996
1997
1998
1999(2)

s/d
5,333
6,242
4,638
5,240
5,590

s/d
215
346
1,587
1,802
1,890

s/d
1,485
1,771
2,326
3,032
3,225

s/d
4,301
3,499
2,901
2,542
2,861

s/d
3,239
3,787
3,370
3,170
3,180

11,120
14,573
15,644
14,823
15,787
16,746

1995
36.6%
29.5%
22.2%
100%
1.5%
10.2%
100%
1996
39.9%
22%
11.3%
22.4%
24.2%
22.7%
1997
31.3%
10.7%
15.7%
19.6%
100%
33.2%
100%
1998
11.4%
19.2%
16.1%
20.1%
33.4%
17.1%
19.0%
100%
1999 (2)
11.3%
19.3%
(1) Incluye deuda con el Gobierno Nacional y deuda consolidada.
(2) A junio de 1999.
Fuente: Elaboración propia en base a información de la Subsecretaría Programación Regional y calificadoras de riesgo.

La otra fuente históricamente importante definanciamiento de los gobiernos provinciales ha sido
la postergación de los pagos de salarios, haberes jubilatorios y endeudamiento con proveedores y
contratistas. En 1995, como consecuencia de la crisis de liquidez que afectó a la mayoría de los gobiernos
provinciales se fueron acumulando compromisos de pagos que no pudieron ser efectivizados a tiempo.

Sesión 1 - 1 0

Esto generó una deudaflotante que superaba los $ 4.000 millones (30% del stock de la deuda). Asimismo,
en Argentina el gobierno nacional constituye otro acreedor importante de las provincias. A través de los
créditos en títulos públicos (BOTESO 10) y redescuentos del Banco Central a los Bancos Provinciales el
gobierno nacional había asistidofinancieramente a las provincias que atravesaban por dificultades
financieras16. En 1995, este componente de la deuda provincial alcanzaba los U$S 1.500 millones.
En 1996 las nuevas autoridades provinciales impulsan una importante cantidad de reformas
estructurales. Estas reformas incluyen, el recorte de los presupuestos, el dictado de leyes de emergencia
económica, la privatización de la banca provincial, de las empresas de energía eléctrica y el traspaso de los
sistemas jubilatorios a la Nación17. Estas reformas, aún en marcha, generaron una sustancial mejora en las
finanzas provinciales. El déficit provincial disminuyó sustancialmente, al pasar de un nivel de $ 3.500
millones a $ 1.700 millones en 1996 (ver Cuadro 1).
A su vez, durante 1996, la mayoría de las provincias emprendieron acciones tendientes a mejorar
el perfil de su endeudamiento, aprovechando que las condiciones de los mercados resultaban más
favorables que las existentes en 1995. Esta reprogramación de la deuda con entidadesfinancieras y
bancarias permitió acceder a tasas de interés más bajas, a plazos de vencimientos más extensos, e incluso,
a períodos de gracia. La tasa de interés promedio del endeudamiento con bancos disminuyó, para las
provincias, 4 puntos al pasar del 16,4% en el año 1995 al 12,4% en 1996 (ver Informe Económico
Regional, Abril de 1997).
No obstante, a pesar del alivio financiero que significó para los tesoros provinciales este proceso
de reprogramación de la deuda, el endeudamiento de las provincias y la Ciudad de Buenos Aires aumentó,
producto de la necesidad definanciamiento y el reconocimiento de pasivos contingentes de los bancos
provinciales. En efecto, la deuda provincial aumenta, en 1996, un 7,4%, con lo cual el stock a fin del
ejercicio alcanza una cifra de U$S 15.600 millones, un 5,7% del PIB (ver Cuadro 2).
En 1996 el stock de la deuda con bancos y entidadesfinancieras aumenta significativamente,
alcanzando los U$S 6.000 millones, representando el 40% de la deuda total (ver Cuadro 3). Paralelamente
a este crecimiento de la deuda con entidadesfinancieras se produce una disminución de la deuda flotante.
El otro componente de la deuda que registra una suba importante es la deuda con organismos
multilaterales de crédito.
En 1997 lasfinanzas provinciales vuelven a mejorar, registrándose una nueva disminución en el
déficit sin privatizaciones (ver Cuadro 1), considerando el resultado global con privatizaciones el déficit se
reduce sustancialmente18. El stock de la deuda provincial diminuye por efecto de la cancelación de la
provincia de Buenos Aires de su deuda con su banco oficial19. Con esta cancelación la deuda provincial
alcanza a los $ 14.822 millones, disminuyendo 5,3% con relación al stock de 1996.
En 1997 las provincias continúan aprovechando la gran liquidez en los mercadosfinancieros para
reprogramar el perfil de vencimiento de la deuda. A su vez, en varias provincias se instrumentaron
programas definanciamiento con el mercado de capitales a través de la emisión de títulos provinciales con
calificación20. Las jurisdicciones que han emitido este tipo de instrumentofinanciero son: Buenos Aires,

Un análisis de los préstamos en BOTESO 1 a las provincias puede verse en Sanguinetti J. (1999).
0
Para u informe m s detallado de estas reformas, véase World Bank (1996), Nicolini et al (1999) y Sanguinetti J. (1999).
n
á
18 Los ingresos por privatizaciones, en el año 1 9 , superaron los U$S 15 0 millones.
97
.0
19 En los primeros m s s del año 1 9 el gobierno de la provincia de Buenos Aires canceló pasivos con el Banco de la Provincia
ee
97
de Buenos Aires por $ 1.906 millones. La deuda se había originado por elfinanciamiento de la construcción de la central de
generación eléctrica de Bahía Blanca y deudas contraídas por la Dirección de Vialidad. Para cancelar el capital adeudado, la
provincia entregó a s banco, bonos cupón cero del Tesoro de los Estados Unidos con vencimiento en el año 2025. Por otra, se
u
estableció el pago mensual de los intereses (Informe Económico Regional, Diciembre de 1997).
2 0 En realidad este proceso comenzó a mediado del año 1 9 cuando la provincia de Buenos Aires coloca el primer tramo del
94
Programa de Eurobonos a Mediano Plazo. La primera emisión se realizó el 1 de julio de 1 9 por U$S 1 0 millones a u
4
94
0
n
plazo de 3 años con amortización del capital afin del período. Los títulos devengaron u a tasa de interés del 9,5%.
n
16
17

Sesión 1 - 1

Mendoza, Tucumán, Chubut, Formosa, Tierra del Fuego y la Ciudad de Buenos Aires. En el Cuadro 5 se
indican las características principales de estos instrumentos financieros.
Cuadro 5
Títulos públicos provinciales. Principales características
Jurisdicción

Bono

Ciudad de Bs. Aires Tango

Calificación

Plazo

Tasa

emitido
fmllns USS)

de
colocación

S&P

Moodv's

D&P

7 / 8 años

10%

500

Internacional

BB-

B1

B1

BBBBBB-

BA3
BA3
BA3

Otras

Buenos Aires

Series 1, 2 , 3 y 4
Serie 5
Serie 6

3 años
5 años
7 años

9,5%/11,5%
10%
7.75%

500

internacional
Internacional
Internacional

Chubut

BOGAR (1)

8 años

Libor + 3,5%

15.2

Local

AA+ (1)

Entre Ríos

Bonos Ciase C

6,7 años TE BCRA +5,22%

70.0

Local

AA+ (1)

Mendoza

Eurobono
Aconcagua 2007

6 años
10 años

10%
10%

93.8
250.0

Internacional
Internacional

Neuquén

TIPRE

6 años

Libor 180+4%

70.0

Local

10 años

Bono global +2%

120.0

Internacional

Santiago del Estero
Tierra del Fuego

Albatros

6 años

9%

55.0

Internacional

Tucumán

EUROBONO
(Serie 2 v 3)

10 años

9.5%

364.3

Internacional

B1
B1

BB

BBAAA (1)

AAA(1)
BBB1

BBBB-

Fuente: Economía & Regiones
(1) Calificación para el mercado local

En 1998 el ciclo de lasfinanzas públicas provinciales se revierte mostrando un empeoramiento en
el resultado financiero sin privatizaciones -el déficit alcanza una cifra de $ 2.200 millones-. Este
desempeño en lasfinanzas subnacionales obedece principalmente a un menorritmode crecimiento en los
ingresos, en particular, los de origen nacional y un aumento sostenido en los gastos provinciales. Como
resultado, la deuda de las provincias aumentó un 6,5% ubicando el stock de la deuda provincial en U$S
15.800 millones.
Del análisis de la composición de la deuda se advierte un significativo incremento de la deuda con
organismos multilaterales de crédito, en gran parte, como consecuencia de los créditos otorgados a las
provincias en el marco de la privatización de los bancos provinciales21. Asimismo, los créditos de los
bancos y entidadesfinancieras con garantía de los recursos coparticipados aumentaron cerca de $ 600
millones en 1998. Dentro de estos créditos merece ser destacado los créditos otorgados por el Banco de la
Nación Argentina por más de $ 230 millones (ver Cuadro 3). Es importante mencionar esto dado que una
de las medidas adoptadas en la crisis del tequila había sido restringir la cartera de préstamos del Banco
Nación a los gobiernos provinciales con el objeto de endurecer la restricción presupuestaria provincial.
Otro hecho que merece se destacado es el uso del Fondo Fiduciario para el Desarrollo Provincial (FFDP)
para asistirfinancieramente a la provincia de Jujuy. Si bien el FFDP había sido creado exclusivamente
para asistir a las provincias que se comprometían a privatizar sus bancos públicos, en 1998, el gobierno
Estos créditos fueron otorgados para sanear las entidadesfinancieras previo a s privatización. Hasta el m m n o el Banco
u
o et
Interamericano de Desarrollo y el Banco Mundial h n aportado m s $ 1.316 millones a este programa (Informe Económico
a
á
Regional julio de 1999).

21

Sesión 1 - 1

nacional otorgó un préstamo a la provincia de Jujuy por U$S 40 millones mediante un convenio de
asistencia financiera que establecía una serie de condicionalidades fiscales22.
En 1999 las finanzas de los gobiernos provinciales continuaron empeorando. En algunas
provincias la situación fiscal se agravó significativamente, generándose conflictos sociales, como en la
provincia de Tucumán, Río Negro, Tierra del Fuego y Corrientes23. La proyección del resultado financiero
consolidado para las 24 jurisdicciones subnacionales indica un desequilibrio de más de $ 3.300 millones.
Por su parte, la información disponible de la deuda provincial indica, a junio de 1999, un stock de deuda
de U$S 16.700 millones, que a fin del ejercicio podría alcanzar los U$S 18.000 millones.
En la composición de la deuda del año 1999 se destaca la alta participación de los créditos con
entidades bancarias yfinancieras (33,4%), el stock de la deuda con organismos multilaterales (19,3%) y la
deuda consolidada (19%).
Si se analiza el stock de la deuda provincial (cerca de U$S 18.000 millones a fin del ejercicio
1999) en relación a la deuda federal (U$S 130.000 millones) se arriba a la conclusión de que la magnitud
agregada del endeudamiento no es el principal problema. En cambio, la performance fiscal de las
provincias presenta una gran heterogeneidad. Es decir, en ciertas jurisdicciones lasfinanzas poseen un
compromiso tal que está latente la posibilidad de que el gobierno federal tenga que salir en su rescate.24
En el Cuadro 6 se indica la evolución de esta deuda provincial para el período 1994-1999,
advirtiéndose aumentos significativos en algunas de ellas. Los casos más elocuentes en cuanto a
crecimiento del stock nominal de la deuda corresponden a Corrientes, Chaco, Chubut, Jujuy, Mendoza,
Neuquén, San Luis y Tierra del Fuego.
En el cuadro 7 se analiza la sostenibilidad de ese endeudamiento medida a partir de la relación con
los ingresos provinciales. Del cuadro se advierte la delicada situación en la que se encuentra un gran
número de jurisdicciones, en particular, se destaca la comprometida situación de las provincias de
Corrientes, Chaco, Chubut, Fomosa, Jujuy, Misiones, Río Negro y Tucumán, con un stock de la deuda que
supera los recursos provinciales.

Según la información brindada por funcionarios de la provincia estas condicionalidades no se habrían cumplido
satisfactoriamente.
2 3 Con relación a los conflictos sociales cabe subrayar que ninguna provincia entró u default financiero ya que los préstamos con
n
entidadesfinancieras, bancos, e inclusive con el gobierno nacional presentan como garantía losrecursoscoparticipados, con lo
cual n n a dejaron de ser cancelados. En cambio, esta crisis si pusieron en peligro la prestación de los servicios públicos
uc
básicos. En corrientes, por ejemplo, los au n s prácticamente no tuvieron clases durante 1999.
lmo
2 4 Probablemente el caso m s dramático sea la provincia de Corrientes, que luego de varios meses de graves conflictos sociales las
á
nuevas autoridades debieron intervenir la provincia y hacer frente a una pesada deuda salarial con fondos nacionales.
22

Sesión 1 - 1

Cuadro 6
Deuda provincial. Evolución 1994-1999 por jurisdicción
(en miles de pesos corrientes)
Jurisdicción
TOTAL

1994

1995

1996

1997

1998

1999*

11,120,400 14,572,868 15,644,445 14,822,582 15,786,618 16,745,720

G.C.B.A.
Buenos Aires
Catamarca
Córdoba
Corrientes
Chaco
Chubut
Entre Ríos
Formosa
Jujuy
La Pampa
La Rioja
Mendoza
Misiones
Neuquén
Río Negro
Salta
San Juan
San Luis
Santa Cruz
Santa Fe
Santiago del Estero
Tierra del Fuego
Tucumán

1,390,400
2,929,700
306,700
1,000,000
532,900
539,900
101,300
414,300
365,400
224,200
0
597,500
311,000
254,700
58,200
364,700
281,600
281,100
0
77,100
397,000
169,000
24,000
499.700

1,246,579
2,827,475
241,467
1,318,000
656,570
568,445
250,929
551,588
731,563
438,498
33,040
485,838
1,160,143
519,971
280,030
583,904
543,452
426,014
16,973
135,591
510,919
310,776
111,415
623,688

1,099,800
3,275,411
323,323
1,121,678
717,836
526,021
344,444
608,083
723,769
594,606
83,810
506,033
1,086,000
538,973
307,670
738,835
514,925
439,877
59,981
82,750
496,633
389,613
133,119
931.255

1,165,900
1,487,370
385,831
1,179,228
761,803
832,345
408,891
523,592
732,499
634,831
82,726
448,268
1,357,000
636,312
304,790
791,308
494,982
373,742
86,701
40,917
579,685
349,400
154,013
1.010.448

1,095,900
1,545,100
397,151
1,104,459
948,491
1,020,112
474,329
600,921
768,221
676,223
81,768
462,890
1,059,386
720,514
397,792
869,589
517,883
405,093
88,449
137,046
938,950
333,014
161,296
982.041

864,900
1,596,000
434,981
1,141,203
1,041,050
1,187,536
689,582
643,459
812,516
717,751
78,972
471,864
1,058,424
782,160
508,446
946,614
546,136
399,602
88,355
176,548
1,021,739
278,678
225,934
1.033,273

Fuente: Elaboración propia en base a información de la Subsecretaría de Programación Regional y

calificadoras de riesgo
* A junio de 1999

Sesión 1 - 1

Cuadro 7
Deuda provincial. Evolución 1994- 1999 por jurisdicción
(En porcentaje de los recursos totales)

Jurisdicción
TOTAL
G.C.B.A.
Buenos Aires
Catamarca
Córdoba
Corrientes
Chaco
Chubut
Entre Ríos
Formosa
Jujuy
La Pampa
La Rioja
Mendoza
Misiones
Neuquén
Río Negro
Salta
San Juan
San Luis
Santa Cruz
Santa Fe
Santiago del Estén
Tierra del Fuego
Tucumán

1994

1995

1996

1997

1998

1999*

40.2%
50.4%
37.8%
78.8%
46.7%
89.3%
76.6%
14.4%
41.7%
68.3%
42.1%
0.0%
120.3%
21.5%
42.6%
7.5%
61.1%
37.9%
51.5%
0.0%
16.3%
18.1%
26.4%
7.3%
61.0%

54.1%
46.5%
36.9%
63.3%
65.4%
111.8%
80.6%
35.6%
57.4%
132.9%
86.0%
7.9%
90.7%
103.5%
90.6%
31.3%
107.6%
73.1%
79.8%
3.9%
25.0%
24.8%
49.9%
36.2%
74.3%

53.3%
40.2%
40.3%
79.9%
50.0%
114.8%
68.0%
44.6%
50.2%
129.8%
99.7%
18.7%
85.2%
83.5%
74.3%
33.9%
108.2%
63.1%
67.2%
12.4%
11.9%
23.3%
59.6%
41.3%
105.3%

44.9%
40.3%
14.3%
76.3%
48.8%
104.9%
94.5%
46.4%
44.8%
116.9%
113.9%
16.6%
72.3%
111.1%
84.4%
31.7%
121.7%
59.9%
53.6%
17.2%
6.1%
23.8%
45.1%
43.4%
105.5%

47.1%
34.5%
16.3%
77.6%
44.0%
128.5%
108.4%
50.4%
50.7%
120.3%
107.2%
15.5%
74.2%
65.0%
88.0%
40.7%
126.3%
61.9%
57.1%
16.8%
21.4%
36.1%
42.9%
45.6%
96.6%

52.1%
26.9%
16.9%
84.1%
48.1%
145.3%
129.3%
135.4%
54.7%
127.8%
120.1%
15.1%
74.1%
82.4%
105.3%
54.3%
146.9%
62.4%
57.9%
16.6%
27.6%
42.4%
34.9%
62.8%
105.0%

Fuente: Elaboración propia en base a información de la Subsecretaría de Programación
Regional y calificadoras de riesgo
* A junio de 1999

El otro aspecto que merece ser destacado de la deuda provincial es su corta madurez. Al respecto,
en el Cuadro 8 se presenta el stock de la deuda consolidada (sin deudaflotante) para el año 1998 y los
vencimientos para dicho año. Del cuadro se desprende la importante concentración de los vencimientos de
la deuda provincial. Otro indicador importante del perfil de la deuda provincial lo constituye el porcentaje
de los recursos del régimen de coparticipación federal de impuestos afectados al pago de amortizaciones e
intereses de préstamos con entidadesfinancieras, organismos internacionales y el gobierno nacional. El
cuadro 9 presenta dicha información, del cual se desprende que algunas provincias tienen afectada una alta
proporción de estos recursos (Tucumán, Río Negro, Formosa, Corrientes y Chubut tienen afectado al
pago de préstamos más del 50% de los recursos coparticipados).

Sesión 1 - 1

Cuadro 8
Perfil de la deuda financiera provincial. Año 1998
(en miles de pesos corrientes)
Stock de la deuda
Consolidada
(31/12/1998)
(1)
TOTAL
G.C.BA.
Buenos Aires
Catamarca
Córdoba
Corrientes
Chaco
Chubut
Entre Ríos
Form osa
Jujuy
La Pampa
La Rioja
Mendoza
Misiones
Neuquén
Río Negro
Salta
San Juan
San Luis
Santa Cruz
Santa Fe
Santiago del Estero
Tierra del Fuego
Tucumán

Amortizaciones e
intereses

13,244,844
840,900
1,270,100
357,127
828,440
861,098
874,319
381,329
533,897
710,721
533,675
81,415
308,901
978,240
705,979
317,582
799,097
506,883
379,113
58,541
117,965
515,190
304,327
115,382
864.623

2,644,931
206,040
343,891
65,642
135,084
189,181
104,147
22,980
94,185
93,152
121,899
2,877
91,908
143,000
76,229
86,684
170,087
91,946
25,500
4,223
2,456
53,508
199,996
22,500
297.818

(2)

(2)/m

20.0%
24.5%
27.1%
18.4%
16.3%
22.0%
11.9%
6.0%
17.6%
13.1%
22.8%
3.5%
29.8%
14.6%
10.8%
27.3%
21.3%
18.1%
6.7%
7.2%
2.1%
10.4%
65.7%
19.5%
34.4%

Fuente: Elaboración propia en base a información de la Secretaría de Programación Económica y Regional

Sesión 1 - 1

Cuadro 9
Retenciones a l o s r e c u r s o s d e c o p a r t i c i p a c i ó n federal d e i m p u e s t o s
(en % de los recursos coparticipados ley 23.548)
1998
total
retenciones por
retenciones (1)
p r é s t a m o s (2)
TOTAL
G.C.B.A.

27.9%

17.3%

Ene/Oct 1999
total retenciones
retenciones por
(1)
p r é s t a m o s (2)
29.8%

19.7%

0.8%

0.3%

0.8%

0.3%

Catamarca

52.4%

22.4%

46.3%

17.3%

Córdoba

14.8%

14.3%

3.1%

2.5%
57.4%

Buenos Aires

Corrientes

48.3%

47.7%

57.9%

Chaco

20.1%

19.6%

30.3%

29.7%

Chubut

14.2%

13.6%

52.1%

51.6%

Entre Ríos

18.6%

18.1%

21.5%

20.9%

Formosa

24.6%

24.0%

59.7%

59.2%

Jujuy

71.2%

40.3%

56.8%

42.8%

2.0%

1.4%

1.8%

1.3%

La Rioja

77.1%

41.6%

62.2%

24.2%

Mendoza

44.9%

6.8%

52.4%

10.2%

Misiones

22.2%

21.7%

37.6%

37.0%

Neuquén

2.0%

1.4%

2.5%

1.9%

Río Negro

96.5%

63.2%

98.7%

63.4%

Salta

53.9%

22.5%

65.6%

34.2%
4.6%

La Pampa

San Juan

33.1%

3.4%

34.1%

San Luis

22.4%

1.7%

21.3%

1.1%

Santa Cruz

2.0%

1.5%

6.1%

5.5%

Santa Fe

2.4%

1.8%

7.1%

6.6%

Santiago del Estero

67.3%

45.4%

50.1%

28.4%

Tierra del Fuego

32.8%

32.2%

43.4%

42.9%

Tucumán

95.8%

58.7%

91.6%

57.7%

Fuente: Elaboración propia en base a información de la Secretaría de Programación Económica y Regional
Nota: (1) Incluye las retenciones efectuadas por el Banco de la Nación Argentina en virtud del aporte al Consejo Federal
de Inversiones, adelantos de coparticipación de la Secretaría de Hacienda, aportes y contribuciones al SIJP,
amortizaciones e intereses de los préstamos en BOTESO 10, préstamos de entidades financieras públicas y privadas
(FFDP, BHN, Banco Nación, BICE, etc.) y redescuentos del BCRA.
(2) Excluye las retenciones del Consejo Federal de Inversiones y aportes y contribuciones al SIJP.

Sesión 1 - 1

VI.

El marco regulatorio del endeudamiento subnacional en Argentina
A. Presupuestos y reglas de endeudamiento provincial

Las constituciones provinciales establecen, con alguna excepción, algún tipo de limitación o
restricción a la capacidad otorgada al ejecutivo para contraer deuda pública . Esta limitación incluye
procedimientos específicos sobre los mecanismos para autorizar al poder ejecutivo a endeudarse (la
exigencia de cierta mayoría especial para su autorización en la legislatura), restricciones en cuanto al
destino y el uso de los fondos (en muchas jurisdicciones subnacionales esta vedado por la constitución
provincial su uso parafinanciar gasto corriente) o limitaciones expresas sobre los montos de los servicios
de la deuda26.
Con respecto a los límites sobre la magnitud del endeudamiento provincial, las constituciones
provinciales establecen, en general, que los servicios de la deuda no pueden superar el 25% de las rentas
provinciales27. Los textos constitucionales no definen claramente el concepto de rentas generales, aún
cuando se debería interpretar como servicios de la deuda el pago de amortizaciones e intereses. Algunas
jurisdicciones tienen cláusulas más estrictas al considerar que para el cálculo de los recursos se tome en
cuenta el menor de los ingresos de los últimos tres años (ver Anexo I: límites al endeudamiento
subnacional).
No obstante la existencia de dichos límites al endeudamiento, en diferentes momentos, las
provincias han superado estos límites constitucionales sin tener consecuencia alguna.
En el Cuadro 10 se exhiben, para 1995 y 1998, los servicios de la deuda provincial en relación con
los ingresos provinciales y los límites al endeudamientofijados por las constituciones provinciales. En el
cuadro se puede observar de que manera las provincias de Córdoba, Corrientes, Formosa, Jujuy y Río
Negro han superado los límites de endeudamiento establecidos por sus propias constituciones
provinciales.
Adicionalmente a las cláusulas constitucionales que limitan el endeudamiento subnacional,
algunas jurisdicciones han impulsado, en este último tiempo, leyes tendientes a restringir el gasto público
y el nivel de déficit fiscal28.

San J a constituye la única jurisdicción que e su constitución no establece límites expresos a la capacidad de endeudamiento.
un
n
Para u análisis m s exhaustivo, véase Jones, et al (1997).
n
á
En el caso de las provincias de Catamarca, Córdoba y Jujuy el límite se establece en el 20% de ios recursos provinciales. E la
n
provincia de Buenos Aires no sefija un límite específico.
2 8 Al respecto, la provincia de Salta sancionó en 1 9 u a ley de responsabilidadfiscal de similares características a la sancionada
99 n
por el gobierno federal.
25

26

Sesión 1 - 1

Cuadro 10
Servicios de la deuda y límites al endeudamiento subnacional
(en % de los recursos provinciales)

Concepto

Limites provinciales
Constitucionales

Total
G.C.B.A.
Buenos Aires
Catamarca
Córdoba
Corrientes
Chaco
Chubut
Entre Ríos
Formosa
Jujuy
La Pampa
La Rioja
Mendoza
Misiones
Neuquén
Río Negro
Salta
San Juan
San Luis
Santa Cruz
Santa Fe
Santiago del Estero
Tucumán
Tierra del Fuego

30%
s/l
20%
20%
25%
25%
S/l

25%
25%
20%
25%
25%
S/l

25%
S/l

25%
25%
s/l
25%
s/l

25%
25%
S/l

25%

Servicios de la deuda (1)
1995

1998

9.0%
9.4%
4.4%
8.1%
19.3%
14.8%
10.9%
16.0%
15.7%
41.9%
9.6%
0.3%
7.7%
13.7%
7.4%
5.2%
31.4%
6.6%
5.4%
0.0%
3.3%
0.5%
14.2%
9.3%
1.8%

12.5%
9.6%
5.9%
13.5%
27.6%
24.9%
11.3%
8.7%
14.5%
25.7%
30.0%
1.3%
18.9%
15.9%
9.5%
10.2%
39.5%
19.4%
8.2%
1.2%
0.8%
3.1%
20.1%
31.0%
5.7%

Fuente: Elaboración propia en base a información de la Secretaría de Programación Económica y Regional
(1) Incluye amortizaciones e intereses

B. El control federal sobre el endeudamiento subnacional en la Argentina
Como anticipamos anteriormente, el control de los comportamientosfiscales provinciales en la
Argentina no se encuadra en ninguna de las clasificaciones habitualmente abordadas en la literatura.
Debido a ello se agregó una nueva clasificación denominada "control burocrático". Es menester destacar
que el control burocrático no es utilizado monopólicamente por la Argentina, sino que muchos países
emplean este sistema de control. Sin embargo, esta opción de control no ha sido considerada en la
literatura como una alternativa en sí misma debido a que generalmente su uso es concomitante con algún
otro método más restrictivo.

Sesión 1 - 1

El control burocrático consiste en la necesidad de formalizar un conjunto de procedimientos
administrativos para poder acceder al endeudamiento. Estos procedimientos involucran el cumplimiento
de una serie de requerimientos formales que incluyen, entre otras cosas, la presentación de información
acerca del estado de lasfinanzas subnacionales, el destino de dicho financiamiento -sea para financiar
obra pública, para refinanciar deuda o parafinanciar el desequilibrio comente-, el porcentaje o monto de
afectación en garantía de los recursos provenientes del régimen de coparticipación federal de impuestos y
la presentación de toda información que avala la legalidad de la operación sobre la base de las normas
(constitucionales) de cada jurisdicción. En el Anexo II se presenta la información requerida por la
Secretaría de Programación Económica y Regional para autorizar una operación de crédito de un gobierno
subnacional.
En la Argentina el control burocrático encuentra su origen legal en la resolución del Ministerio de
Economía 1.075 y en la comunicación A282 del Banco Central. La resolución 1075 asigna a la Secretaría
de Programación Económica y Regional (SPEyR) la función de control del endeudamiento subnacional
que utilice como colateral los recursos del régimen de coparticipación federal de impuestos y todo aquel
endeudamiento en moneda extranjera. Por su parte, la comunicación A282 prohibe a la banca privada el
otorgamiento de créditos al sector público con la única excepción -excepción que en los hechos resulta la
regla- que la operación esté autorizada por el Ministerio de Economía.
Para el endeudamiento en moneda extranjera, la resolución 1.075 establece que "Las provincias,
sus organismos o empresas, están sujetas a la supervisión, coordinación y aprobación del Ministerio de
Economía y Obras y Servicios Públicos, respecto de todo acto, contrato, gestión o negociación de créditos
a celebrarse por los que se originen, o puedan eventualmente originarse, obligaciones de pago en moneda
extranjera."
Por su parte, para los casos de afectación de los recursos del régimen de coparticipación en
garantía de endeudamiento provincial, la resolución 1.075 establece que "los importes que correspondan a
cesiones de derechos sobre impuestos coparticipables o sobre la garantía otorgada por la Nación a las
Provincias que suscribieron el Pacto Fiscal aprobado por la ley 24.130, serán retenidos por este Ministerio
en una cuenta especial, hasta tanto los acreedores notifiquen la revocación de dichas cesiones".
En la práctica, en cumplimiento de estas normativas, las jurisdicciones que deseen afectar los
recursos del régimen de coparticipación, o que pretendan tomar créditos en moneda extranjera, deben
acudir a la SPEyR para que ésta autorice la operación. Es importante mencionar que la normativa
señalada, tanto la 1075 como la comunicación A282 del BCRA, no establecen restricciones específicas en
cuanto al grado de afectación de los recursos coparticipados que las provincias pueden ofrecer como
colateral, o respecto a topes específicos en relación a los créditos tomados, sino que la autorización de la
SPEyR constituye, en los hechos, un requisito más que las provincias deben cumplimentar para obtener un
crédito en moneda extranjera o un crédito con afectación de recursos coparticipados, puesto que la
autoridad de aplicación no tiene suficientes argumentos legales para no autorizar una operación de crédito
de los gobiernos subnacionales.29
En función de los requisitos exigidos por la SPEyR para autorizar el endeudamiento subnacional
(ver Anexo II), el gobierno federal cumple con dos roles principales. En primer lugar se asegura que la
operación haya cumplimentado con todos los requisitos legales (constitucionales) establecidos en la
provincia (ley autorizando el endeudamiento, ley autorizando la afectación de los recursos coparticipados,
dictamen de la Fiscalía de Estado y el Contador General de la provincia autorizando la operación), afin de
resguardarse de eventuales denuncias por ilegalidad de la operación. En segundo lugar se asegura que los
recursos afectados en garantía sean suficientes para cancelar los compromisos asumidos (intereses y
amortizaciones), es decir, resguarda los intereses del prestatario.

Por la Constitución nacional las provincias conservan todo el poder no delegado al gobierno nacional. En este aspecto las
provincias mantienen plena autonomía para endeudarse.

29

Sesión 1 - 1 0

Es necesario destacar que se han concretado algunas operaciones de crédito de la Banca privada
con las provincias sin cumplir con la normativa vigente. Además, como se analizará con mayor detalle en
la próxima sección, las entidades bancarias yfinancieras poseen un fuerte incentivo a exigir a la provincia
la intervención del Ministerio de Economía debido a que de este modo se obtienen privilegios en cuanto a
la determinación de los capitales mínimos y a las previsiones y clasificación de deudores.
Para concluir vale la pena argumentar porque este sistema de control no se encuadra en ninguno
de las clasificaciones de control interno al sector público consideradas en la literatura. En primer lugar, es
claro que no existe ninguna regla impersonal, ni de diseño presupuestario en el proceso de control. Es
decir, no existe un marco de referencia explícito al cual el organismo de control -la SPEyR- deba
controlar el cumplimiento de ciertos estándares o metas fiscales en la jurisdicción que requiere la
autorización para endeudarse. Tampoco, el sistema se ajusta a un mecanismo de coordinación bilateral, ni
multilateral, en el cual los distintos niveles de gobierno se ponen de acuerdo en cumplir cierto cronograma
definanciamiento o un techo para la contratación de crédito público. Por último, el control no se ajusta a
un sistema de control de endeudamiento directo. Aunque esta última alternativa parece la más cercana a la
caracterización del caso argentino, el poder del gobierno nacional en el control de endeudamiento
provincial en nuestro país no se acerca ni remotamente al poder de control que gozan los países
generalmente considerados dentro de esta clasificación.

C. Los privilegios de las provincias en el acceso al crédito
En nuestro país existe un fuerte incentivo para que el sistemafinanciero conceda préstamos a las
provincias. Esto se debe, no sólo a los costos asociados a evaluar a una gran cantidad de pequeños
prestatarios que compiten con las provincias por los fondos, es decir a la existencia de economías de
escala en la evaluación de los préstamos a las provincias, sino también a los menores requerimientos de
capitales mínimos y previsiones para este tipo de operaciones de crédito.
En particular, la comunicación "A" 2253 de Banco Central establece que: "a los fines de la
determinación de la exigencia de capitales mínimos, se ponderen al 0% los préstamos que se otorguen a
las Provincias, a la Municipalidad de Buenos Aires y a los bancos oficiales de esas jurisdicciones,
garantizados con la afectación especial de fondos de coparticipación federal de impuestos, en operaciones
que cuenten con la pertinente intervención del Ministerio de Economía y Obras y Servicios Públicos de la
Nación".
A su vez, la comunicación "A" 2216 del Banco Central establece que, a los efectos de las
previsiones mínimas y clasificación de deudores, los préstamos con -"afectación en prenda de fondos de la
coparticipación federal de impuestos, en operaciones que cuenten con la pertinente intervención del
Ministerio de Economía y Obras y Servicios Públicos"- se incluyan en la categoría normal. Es decir, esta
norma exige el mínimo del 1% de previsionamiento para este tipo de operaciones.
Estas normas legales generan grandes incentivos a que los bancos concedan préstamos a las
provincias. Por otra parte, como se mencionó anteriormente, para obtener estos beneficios las entidades
financieras demandan a las provincias que deseen endeudarse la intervención del Ministerio de Economía.
Recientemente la comunicación "A" 2793 del Banco Central redujo los beneficios a los préstamos
provinciales. Esta comunicación establece la ponderación entre el 40% y el 60% para los créditos
otorgados al sector público provincial que cuenten "con garantía de coparticipación de impuestos
mediante la cesión directa o indirecta de los correspondientes derechos y que cuenten con la pertinente
intervención del Ministerio de Obras y Servicios Públicos de la Nación." Esta medida obliga a los bancos

Sesión 1 - 1

que otorgan préstamos a las provincias con esta garantía a incrementar sustantivamente sus
inmovilizaciones de fondos.
Sin embargo, debido a que esta medida ha desencadenado algunas protestas por parte de las
entidades bancarias yfinancieras nacionales, que concentran el crédito provincial, se decidió prorrogar en
dos oportunidades la entrada en vigencia de la norma. En una primera instancia la prórroga operaba hasta
el 1/1/2000, pero recientemente la norma se prorrogó nuevamente30.

D. Disciplina de mercado en el caso argentino
Como se indicó en apartados anteriores a partir de la estabilización macroeconómica, el
endeudamiento con entidades bancarias y financieras ha constituido una importante fuente de
financiamiento de los déficits presupuestarios provinciales. En este sentido, en el año 1999, la
participación de la deuda bancaria en el stock total de la deuda para el consolidado de las 24 jurisdicciones
subnacionales ha sido del 33%, alcanzando en junio de 1999 un nivel de aproximadamente $ 5.600
millones
En general este endeudamiento presenta como garantía (colateral) los recursos de libre
disponibilidad de las provincias provenientes del régimen de coparticipación federal de impuestos.
El sistema de los préstamos con garantía de coparticipación funciona del siguiente modo: cuando
una provincia toma un préstamo garantizado con recursos del régimen de coparticipación, afecta parte de
los mismos al cumplimiento de las obligaciones contraídas. El Banco Nación, que cumple el rol de agente
financiero del régimen, es el encargado de transferir diariamente los recursos de dicho régimen a las
provincias. Al concretarse la operación de préstamo, el Banco Nación recibe la información referida a los
plazos y pagos de amortizaciones e intereses, junto con las directivas de asignar parte de los recursos de la
provincia prestataria al pago de sus obligaciones. De esta forma, del monto a transferir a la provincia, el
Banco Nación retiene los importes correspondientes a los pagos de intereses y amortizaciones, que son
girados automáticamente a la entidad que otorgó el crédito. Como queda explícito, las provincias no
interfieren en este mecanismo
Con respecto a estetipode operaciones de préstamos la calificadora Moody's considera que el uso
de garantías de coparticipación permite neutralizar el riesgo provincial sobre el instrumento financiero.
Para Moody's: "Los recursos provenientes de la coparticipación constituyen una importante herramienta
definanciación, especialmente en combinación con técnicas definanciamientos estructuradas. Bajo la Ley
de Coparticipación y la legislación sobre Fideicomiso" (ley 24.441) "el derecho a recibir el pago de
recursos coparticipables futuros puede transferirse a unfideicomiso y de esta manera aislarlo del riesgo
crediticio de la entidad que los transfiere" (provincia). "Además, se puede utilizar protección crediticia
adicional para lograr una calificación de la entidad que realiza la transferencia (provincia), y en algunos
casos, que exceda también "el techo" de la calificación soberana."
En este apartado se analiza que implicancias posee el hecho de que las provincias utilicen como
garantía de sus préstamos un recurso cuya administración está fuera de su control. La hipótesis es que la
disciplina fiscal no sería aplicable en el caso argentino debido a las distorsiones generadas por el marco
institucional vigente en el mercado de préstamos a provincias. Es decir que no se espera encontrar una
relación estadísticamente significativa entre el costo del crédito de las distintas jurisdicciones y las
variables fiscales.

La importancia de esta circular para las entidadesfinancieras argentinas se pone de manifiesto e la inclusión en recientes
n
operaciones de crédito de una corrección de la tasa de interés e función de los cambios de los requisitos mnm s de liquidez
n
íi o
exigidos por el BCRA.

30

Sesión 1 -

En el marco institucional vigente no se satisface una de las condiciones esenciales para que el
mercado de crédito cumpla con el rol de disciplinados que consiste en que los prestamistas internalicen el
costo de cesación de pagos por parte de los gobiernos provinciales en caso de que estos no puedan afrontar
la devolución de los créditos por encontrarse en dificultades financieras. Esta condición no es satisfecha
en el caso de los préstamos con garantías de coparticipación ya que si una provincia atraviesa por
dificultades financieras no tiene la posibilidad de declararse en cesación de pagos.
La utilización de recursos de coparticipación como garantía de los préstamos constituye una
tecnologíafinanciera31 que, al asegurar la consistencia temporal de los compromisos asumidos por las
provincias, levanta la restricción de liquidez que estas enfrentarían en caso de tener un comportamiento
fiscal irresponsable.32 Es decir, el Banco Nación retiene de los fondos que le corresponden a cada
provincia en el marco del régimen de coparticipación federal de impuestos los importes correspondientes a
los pagos de intereses y amortizaciones, los cuales son girados automáticamente a la entidad que otorgó el
crédito sin que las provincias puedan interferir en este mecanismo.33 Por lo tanto, puede argumentarse que
este mecanismo constituye un sistema automático de pagos, y no sólo un colateral.
Una pregunta relevante es si esta tecnologíafinanciera disciplina a los gobiernos subnacionales
por sí misma. Esta pregunta es relevante por cuanto en el caso de los agentes privados una tecnología de
este tipo, que hace creíbles las intenciones de honrar la deuda, evita el racionamiento del crédito y elimina
el costo de monitoréo.
Sin embargo, en el caso de los gobiernos subnacionales, la alta tasa de descuento de los políticos y
la posibilidad de salvatajes alteran este resultado. Es decir, la estrecha relación entre el gobierno nacional
y los gobiernos subnacionales desalienta el control por parte de los prestamistas. Esto ablanda la
restricción presupuestaria de los gobiernos subnacionales permitiendo comportamientos imprudentes y, de
este modo, aumenta la probabilidad de que en el futuro se produzca un rescate por parte del gobierno
nacional cuando el gobierno subnacional no tiene posibilidad de ajuste. Debe recordarse que la
proposición original era si el mercado era capaz de sustituir las reglas explícitas.34
Esto es, la estructura de las relacionesfinancieras intergubernamentales vigente en la Argentina,
junto al estrecho horizonte temporal internalizado por los políticos, induce a un gasto excesivo. Sin
embargo, el controlfiscal y la restricción de liquidez que el sistemafinanciero impondría a las provincias,
en ausencia de una tecnología de este tipo, contrarrestaría la tendencia al comportamiento fiscal
imprudente. Por lo tanto, esta innovación tecnológica, al eliminar el control y la restricción de liquidez,
posibilita que los funcionarios provinciales realicen sus ambiciones de corto plazo y aumenta la
probabilidad de un salvataje por parte del estado nacional.
Otro elemento que hay que destacar es que, como se analizó con anterioridad, en Argentina existe
un fuerte incentivo para que el sistemafinanciero conceda préstamos a las provincias, lo que impide el
cumplimiento de la segunda condición postulada para que el mercado funcione con eficacia.

Se debe mencionar que a n n existiendo u a tecnología de este tipo los prestamistas pueden percibir una garantía implícita por
u o
n
parte del Gobierno Nacional, por lo cual la política propuesta debe ser m s amplia que la eliminación de este mecanismo.
á
32 Cabe mencionar que mientras el sector privado puede utilizar suriquezatotal c m colateral, no es c m n observar que los
o o
o ú
gobiernos c m r m t n m s allá de sus recursos.
o po ea
á
33 Sin embargo, dada la estructura legal de este sistema, Standard & Poors argumenta que las provincias podrían alterar los
términos y condiciones de su endeudamiento, de modo que no se eliminaría completamente la probabilidad de default.
(Creditweek, 1997)
34 Dillinger et al. (1998) a g m n a que la garantía de coparticipación fuerza a las provincias a honrar su deuda y por lo t n o
ru e tn
at
refuerza la restricción presupuestaria subnacional. Sin embargo, esta garantía hace que el mercado permita que las provincias
realicen comportamientos imprudentes yriesgosos,que sólo sean sostenibles si la economía no sufre u shock negativo. Pero,
n
e el caso que la economía sufra u shock negativo, las provincias se verán obligadas a solicitar u bailout por parte del Estado
n
n
n
Nacional. Es decir, esta garantía hace que el mercado no controle a las provincias, y que por lo tanto vaya penalizando los
comportamientos imprudentes sin saltos discretos.
31

Sesión 1 -

En nuestro país no se cuenta con una fuente de información que permita testear acabadamente la
hipótesis planteada. Sin embargo se dispone de la información de 64 operaciones de préstamos con
garantía de coparticipación otorgados a 15 provincias, a partir de los cuales se puede intentar una
aproximación. Estos préstamos están nominados en dólares y corresponden al período Julio de 1996-Julio
de 1997. Las tasas de interés y las fechas de cada operación se exhiben en el gráfico Al del anexo IH.
A su vez, también se cuenta con un cálculo de la vida promedio para cada préstamo.35 Dada la
heterogeneidad de las distintas operaciones en cuanto a la estructura temporal de amortización del capital
se procedió a estimar una regresión que vinculara estas variables36
Luego se calculó el spread de cada operación en relación al valor de la curva de regresión
estimada para la misma vida promedio. Por último se promediaron los spreads de las operaciones
correspondientes a cada provincia. Los resultados expresados en puntos básicos se presentan en el Cuadro
11. En el cuadro también se exhiben indicadores fiscales que usualmente se consideran signos de una
buena performance fiscal.
Es posible pensar que existe un problema de selección endógena a la muestra, lo cual probaría que
las provincias que presentan peor situaciónfiscal no posean acceso al mercadofinanciero. En ese caso, no
se podría testear con la muestra la hipótesis de racionamiento de crédito. Sin embargo, en el Cuadro 11 se
puede observar que provincias que presentan una muy mala performance fiscal han tenido acceso al
mercadofinanciero. Por ejemplo, Formosa es la provincia con mayor ratio entre la deuda y los recursos
totales, presentó un déficit alto en 1996 y posee un gran desequilibrio vertical. Sin embargo, esta provincia
tuvo acceso al crédito a una tasa reducida.
En el Cuadro 12 se presentan los resultados de las regresiones entre los spread y cada una de las
variables fiscales tomadas individualmente.37 En el Anexo DI se exponen gráficos de dispersión que
vinculan los spread con cada una de las variablesfiscales consideradas.
De las regresiones se puede observar que en el caso de considerar todas las provincias de la
muestra los signos de las regresiones son los esperados por la teoría de la disciplina de mercado. Sin
embargo, sólo es significativa la variable "% de recursos afectados".
Cabe destacar que esta no es una de las variablesfiscales que forma parte de la teoría de disciplina
de mercado sino que se hace relevante debido al contexto institucional Argentino. Sin embargo, si el
mercado penalizara la mayor proporción de recursos afectados, entonces esta innovación institucional
podría ser de utilidad. Debido a ello resulta oportuno advertir que los gobiernos subnacionales no tienen la
capacidad de declarar una cesación de pagos de estos préstamos aún cuando posean el total de recursos
afectados.
Además, al extraer las provincias de Santa Fe y Córdoba de la muestra la afectación de
coparticipación deja de ser una variable estadísticamente significativa y los resultados de todas las
variables empeoran. En este caso, la deuda y el desequilibrio vertical poseen el signo contrario al esperado
en la teoría del mercado disciplinador y ninguna de las variables es estadísticamente significativa.
La vida promedio de u préstamo constituye una medida que informa acerca de la estructura temporal de la amortización del
n
capital. Por ejemplo, una operación ha cinco años con amortización mensual posee una vida promedio de 2,5 años; y una
operación a cinco años con una única amortización alfinal posee una vida promedio de 5 años. Sí las operaciones poseen
períodos de gracia entonces se alarga la vida promedio, dado que se retarda la amortización del capital.
3 6 La relación entre el costo y el plazo de las operaciones generalmente se estiman con una relación en logaritmos. Kiguel y
Lopetegui 1997. En este caso la elasticidad del spread a la vida promedio resultó de positiva y significativa. |Log(Tasa)=-|
|2,2995+0,04652*Log(vida promedio); t=5,38, R'=31,8%j
35

Estas estimaciones están sujetas a serios problemas econométricos y sólo deben considerarse como una aproximación. Se
regresan las variables tomadas individualmente debido a que sólo se cuenta con 1 observaciones. Parece adecuado t m r
5
o a
variablesfiscales del año 1 9 debido a que la mayor cantidad de préstamos corresponden al año 1 9 y en ese año sólo se
96
97
posee la información agregada del año anterior.

37

Sesión 1 - 144

Cuadro 11
Préstamos de entidades financieras a provincias
Spread en puntos básicos e indicadores fiscales del año 1997
Deuda/Rec. totales

Deficit/Rec.
totales

Rec.
Propios/Rec.
tntaloc

79.9%
10.0%
CATAMARCA
17.8%
CORDOBA
51.5%
-8.2%
44.9%
CORRIENTES
114.8%
5.3%
13.1%
CHACO
86.3%
8.7%
15.5%
CHUBUT
70.1%
23.5%
35.6%
ENTRE RIOS
50.1%
-12.5%
27.6%
FORMOSA
129.8%
7.0%
8.9%
82.2%
22.4%
38.4%
MENDOZA
74.3%
18.5%
MISIONES
9.0%
108.2%
12.2%
27.3%
RIO NEGRO
SALTA
63.0%
-3.4%
26.0%
SAN JUAN
67.2%
-5.4%
15.0%
SANTA FE
23.3%
2.0%
41.4%
TUCUMAN
105.3%
15.5%
20.4%
TIERRA DEL FUEGO
41.3%
3.8%
29.4%
Fuente: Elaboración propia en base a información de la Subsecretaría de Programación Regional

%de

Spread

afectación
59.4%
37.5%
44.0%
18.7%
8.0%
24.6%
66.4%
60.2%
36.6%
90.6%
49.9%
44.1%
3.1%
45.0%
47.1%

36
-87
11
50
15
4
-25
58
22
86
48
42
-164
134
87

Cuadro 12
Regresión lineal de los spreads c o n variables fiscales
Variables Explicativas
Deuda/Recursos totales
Déficit/Recursos totales
Recursos Propios/Recuros totales
% de afectación de recursos

Coeficiente
109.9
252.6
-251.9
157.1

Estadístico t
1.78
1.47
-1.58
2.04

R2
0.197
0.142
0.161
0.243

N° de observaciones
15
15
15
15

Rearesión sin Santa Fé v Córdoba
Deuda/Recursos totales
Déficit/Recursos totales
Recursos Propios/Recuros totales
% de afectación de recursos
Fuente: elaboración propia

Coeficiente
-14.1
93.5
126.4
47

Estadistico t
-0.29
0.82
0.99
0.82

R2
0.007
0.057
0.083
0.058

N° de observaciones
13
13
13
13

En suma, de la escasa información disponible no se encuentra que el mercado este disciplinando
eficazmente a los gobiernos subnacionales mediante una penalización sobre el costo del endeudamiento.
Esto es particularmente relevante cuando se extrae a las provincias de Córdoba y Santa Fe. 38
Es decir, el mercado presenta una distorsión de tal magnitud que, ante el empeoramiento de la
capacidad de repago de los prestatarios, ni siquiera se observan incrementos en las primas por default.
Como mencionan Sanguinetti y Porto (1996) "puede ocurrir que la capacidad de endeudamiento de los
gobiernos subnacionales no esté limitada por la performance actual o futura y que entonces el mercado no
cumpla con su rol de disciplinador. Es el caso de por ejemplo, los préstamos bancarios con garantía de
coparticipación."
38

Debe notarse que en la Argentina los rescates han sido mas frecuentes en el caso de las provincias pequeñas. Sanguinetti J.
(1999).

Sesión 1 - 145

Cabe destacar que de este análisis no debe concluirse que una eliminación de la garantía de
coparticipación permitiría sustituir las reglas explícitas de control fiscal de los gobiernos subnacionales
por la disciplina de mercado. En numerosas oportunidades el gobierno nacional ha rescatado a provincias
por deudas que no poseían esta garantía. La disciplina de mercado debe ser considerada mas un
complemento que un sustituto de las reglasfiscales explícitas.
Sin embargo, aunque no se puede confiar plenamente en la disciplina de mercado para prevenir
comportamientos no sostenibles, el mercado puede ser de gran ayuda si las instituciones fiscales son
diseñadas de modo que el mercado participe en el controlfiscal. Para ello, como línea de investigación, es
relevante analizar como el mercado reacciona a las reglas fiscales.

VII. Síntesis y conclusiones
En este trabajo se ha analizado la cuestión del endeudamiento provincial en la Argentina y la
necesidad de avanzar sobre un rediseño del marco institucional afin de evitar los posibles costos externos
que el comportamiento fiscal de una provincia puede ocasionar sobre las demás provincias y sobre la
Nación. A su vez, se analizaron los distintos tipos de control de endeudamiento utilizados en la
experiencia internacional a fin de evaluar que alternativas existen en la elección de un mecanismo de
control que se ajuste a las necesidades y características de la Argentina.
Cabe puntualizar algunos resultados del trabajo:
Las características de la economía Argentina, en el sentido de país chico y el hecho que los
casos de bailouts han sido másfrecuentes para las provincias de menor tamaño, hacen que la
visión tradicional que asigna la política fiscal compensatoria al gobierno nacional sea más
apropiada.
En la experiencia internacional, aún los países grandes -para los cuales puede resultar
oportuno que los gobiernos subnacionales posean una política fiscal autónoma- cuentan con
algún mecanismo de coordinación de modo de impedir resultados ineficientes como
consecuencia de comportamientos oportunistas (free raider).
La disciplina de mercado no constituye para el caso de la Argentina un sustituto a los
controles internos al sector público.
El control del endeudamiento provincial en la Argentina resulta inefectivo, limitándose, en el
caso del control federal un trámite burocrático. Además, existe una distorsión en la regulación
financiera que alienta el acceso al crédito por parte de las provincias.
El gobierno federal actuó reiteradamente como prestamista de última instancia, otorgado
créditos a provincias con serias dificultades financieras mediante el uso del Banco de la
Nación Argentina, el BICE, el Fondo Fiduciario, títulos de tesorería (BOTESO 10), anticipos
de la coparticipación y Aportes del Tesoro Nacional.
Del menú de opciones de mecanismos para el control del endeudamiento subnacional utilizados en
la experiencia internacional, los métodos más adecuados son las reglas y la coordinación en el ámbito de
una instituciónfiscal federal.
En el caso de la instauración de reglas -tipo Maastricht- debe considerarse la gran heterogeneidad
de las distintas provincias en cuanto a sus indicadores fiscales. Dicha heterogeneidad imposibilita la
instauración de reglas generales para todas las provincias del tipo deuda sobre recursos, deuda sobre

Sesión 1 - 146

producto bruto geográfico, etc., ya que estas reglas serían inoperantes para algunas provincias e
imposibles de cumplir para otras.
Por lo tanto, en el caso de las reglas, las opciones son instaurar reglas sobre algúnflujo -déficit,
por ejemplo-, lo cual implicaría en cierto sentido perdonar las malas performances fiscales pasadas, o
diseñar un tratamiento asimétrico, instaurando una regla generalizada sobre el stock de deuda medida en
relación, o bien a los recursos, o bien al producto, y dando un período de convergencia a las reglas
generalizadas durante el cual se exige la amortización gradual de endeudamiento para las provincias que
en el momento inicial no las satisfagan.
Paralelamente a la instauración de una regla sobre el endeudamiento subnacional será importante
diseñar una instancia institucional de cooperación y negociación interj urisdiccional. El nuevo esquema de
coparticipación federal de impuestos sería el instrumento más adecuado para introducir en el proceso de
federalismo argentino esta nueva institución.
Por último, en el corto plazo será importante trabajar sobre la oferta de crédito, aprobando una
circular que desaliente los préstamos a provincias con garantía de los recursos coparticipados
incrementando las exigencias de capitales mínimos para estas operaciones. Por otra parte, sería
recomendable penalizar con un nivel superior de requerimiento de capitales mínimos a las operaciones de
préstamos con garantía de los recursos coparticipados que afecten una cierta proporción de estos recursos,
por ejemplo, cuando un préstamo afecte un porcentaje superior al 50%.

Sesión 1 - 147

Bibliografía
Arreaza, Sorensen, Yosha (1999) "Consumption smoothing through fiscal policy in OECD and EU
countries". En "fiscal Institutions andfiscal performance" (1999).
Artana y López Murphy (1997) "Descentralización Fiscal y Aspectos Macroeconómicos: una pespectiva
latinoamericana". Descentralización Fiscal en América Latina. Cepal.
Banco Mundial (1996). "Argentine provincialfinances study".
Bayumi, Goldstein and Woglom. (1995) "Do credit Markets Discipline Sovereing Borrowers? Evidence
from U.S. States. Journal of Money Credit and Banking" Noviembre de 1995.
Bishop, Damrau and Miller (1989) "1992 and beyond: Market Discipline Can Work in the EC monetary
unión".
Damill (1997): Análisis de la situaciónfiscal de la Provincia de Mendoza". Mimeo
Dillinger and Webb (1998) "Fiscal Management in Federal Democracies: Argentina and Brazil" mimeo,
World Bank.
Eichengreen and von Hagen (1996) "Fiscal policy and monetary unión: is there a tradeoff between
federalism and budgetary restrictions?" Working paper. National Bureau of Economic Research.
Garegnani Lorena (1998). "Finanzas provinciales y ciclo económico". UNLP. Mimeo.
Gamboa (1997) "El rescate Financiero de los Gobiernos Estatales por el Gobierno Federal: Comparación
de los casos de Estados Unidos, Brasil y México." Banco de México.
Huther y Shah (1996) " A simple measure of good governance and its application to the debate on the
appropriate level of fiscal decentralization". Mimeo.
Informe Económico Regional. Subsecretaría de Programación Regional. Ministerio de Economía y Obras
y Servicios Públicos, varios números.
Kiguel y Lopetegui (1997) "Entendiendo elriesgopaís". CEMA 125.
Kenyon y Benker (1984). "Fiscal Discipline: Lesson from the state experience". National Tax Journal.
Oates, Wallace E. (1972) "Fiscal Federalism". New York.
Jones, Sanguinetti, Tommasi (1997): "Politics, Institutions and Public Sector in the Argentine Provinces"
in James Poterba and Jurgen von Hagen (eds) Budget Institutions and Fiscal Outcomes Chicago Press.
Poterba y Rueben (1999) "State fiscal institutions and the U.S. municipal bond market". En "fiscal
Institutions andfiscal performance" (1999).
Piffano, Sanguinetti J. y Zentner (1998a) "Lasfinanzas provinciales y el ciclo económico"- CEDI, Mimeo.
Piffano, Sanguinetti J. y Zentner (1998b) "Instituciones Fiscales" CEDI, Mimeo.
PruTiomme, R (1995): 'The Dangers of Decentralization" The World Bank Research Observen
Monasterio, Sánchez Alvarez y Angel (1999) "Controles internos al endeudamiento versus racionamiento
del crédito. Estudio especial del caso de las comunidades autónomas españolas".
Musgrave (1959) 'Teoría de la Hacienda Pública", Ed Aguilar, Madrid.
Nicolini, Posadas, Sanguinetti J., Sanguinetti P., Tommasi (1999). "Decentralization, Fiscal Discipline in
Sub-National Government, and the Bailout Problem: The Case of Argentina". BID.

Sesión 1 - 148

Lañe (1993)" Market Discipline", IMF Staff Papers.
Spahn (1997) "Descentralización Fiscal en América Latina." Cepal.
Sanguinetti P. y Porto A. (1996) "Reflexiones sobre el ciclo económico, coparticipación, endeudamiento y
finanzas provinciales". Universidad de La Plata.
Sanguinetti J. (1999) "Restricción de presupuesto blanda en los niveles subnacionales de gobierno: el
caso de los salvatages en el caso Argentino". DI Seminario Internacional de Federalismo Fiscal La Plata,
junio de 1999.
Sanguinetti J., Zentner y Rotsztein (1997) "Disciplina fiscal y endeudamiento provincial: ¿Se puede
confiar en el mercado como disciplinador?" Mimeo, Subsecretaría de Programación Regional.
Shome Parthasarathi (1995): Federalismo Fiscal: Ingresos, Gastos y Gestión Macroeconómica. Proyecto
Regional de Política Fiscal. Serie de Política Fiscal 64. CEPAL.
Series de Estudios Fiscales (1996) "Una propuesta para la distribución primaria de la coparticipación
federal de impuestos".
Talvi y Vegh (1997) "Can optimal policy be procycücal?" Mimeo.
Tanzi (1995) Fiscal Federalism and Decentralization: A Review of Some Efficiency and Macroeconomics
Aspects" Annual World Bank Conference on Development Economics 1995.
Ter Minassian (1997): "Descentralización y gestión macroeconómica. Documento para el seminario
internacional de Federalismo Fiscal y Gobiernos Locales. Mimeo.
Wildasin (1997) "Externalities and Bailouts. Hard and Soft Budget Constraints in Intergovernmental
Fiscal Relations". Policy Research Working Paper 1843.
Word Bank (1996): Córdoba - Evaluación del Sertor Público: Propuestas para una ReformaVolumen I y
H. Report N° 15132 AR.
Zentner A. (1999): "Cuestiones macroeconómicas del federalismo fiscal: el caso Argentino" El Seminario
Internacional de Federalismo Fiscal La Plata, junio de 1999.

Sesión 1 - 149

Anexo I
Límites a la capacidad de endeudamiento subnacional
Provincia
GCBA
Buenos Aires

Autorización

Límite de los
servicios

2/3 del consejo

Prohibición de
utilizarlo para

30% de las rentas

2/3 de los
presentes de
c/cámara

Otras especificaciones

no dice nada sobre límites.

Catamarca

2/3 de los
presentes de el
cámara

20% de las rentas gastos ordinarios deben especificarse los recursos
para los servicios y amortización

Cordoba

2/3 de los
presentes de c/
cámara

20% de las rentas

Corrientes

2/3 del total de
c/cámara

25% de la renta

Chaco

2/3 legislatura

25% de la renta

Chubut

Vz +1 de los

Entre Rios

V +1 de los
k
miembros

25% de la renta

2/3 de los votos
presentes

como renta se considera la
menor de los últimos 3 años

los servicios no excederán

anualmente

25% de las
renta**

Formosa

el 25%

gastos ordinarios

miembros

Jujuy

2/3 legislatura

La Pampa

2/3 legislatura
2/3 de los
miembros de la
cámara

25% de la renta

Mendoza

2/3 de los
presentes de
c/cámara

Misiones

2/3 de los
presentes

Neuquén

los servicios son sobre el stock y
la nueva deuda.

25% de la renta

La Rioja

gastos ordinarios

2/3 de la totalidad
de miembros

20% de las rentas gastos ordinarios
déficit

sólo para obras públicas y
colonizacion
sólo para obras públicas.Para el
computo de renta se utilizará el
menor de los últimos tres años,
se debe epecificar los recursos
que se utilizaran para los
servicios de la deuda y su
amortización.
la ley debe especificar los
recursos para amortizaciones y
servicios

25% de la renta

gastos ordinarios
no dice nada sobre límites al
endeudamiento.

Sesión 1 - 150

Autorización

Provincia
Rio Negro

Límite de los
servicios

2/3
de
los 25% de la renta*
miembros de la
legis.

2/3 de los
presentes de
c/cámara

Salta

Prohibición de
utilizarlo para
déficit

Otras especificaciones
la ley debe especificar los
recursos para hacer frente a
servicios y amort. y solo puede
utilizarse paraobraspúbiicas,
reforma agraria y calamidades.

25% de la renta*

no dice nada sobre límites ni
mayoría especial.

San Juan
San Luis

2/3 legislatura

Santa Cruz

la renta se calcula como la
menor de los 3 últimos años

25% de la renta

2/3 votos de los
presentes

Santa Fe

deben especificarse los recursos
para su amortización y el objeto
de su destino.
no dice nada sobre mayoría
especiales.

25% de la renta

Santiago del
Estero

2/3 legislatura

Tucumán

2/3 legislatura

Tierra
Fuego

del

2/3 de los
miembros

25% de la renta

25% de la renta

déficit

especificar los recursos para los
servicios

gastos de
prohibición de emisión de títulos
funcionamiento y públicos y privados que tengan
servicios de adm. como objeto el reemplazo de la
moneda de curso legal.

* incluye las amortizaciones y servicios** se toma por renta el promedio de los recursos ordinarios Fuente: Jones et
al (1997)

Sesión 1 - 1 5 1

Anexo II
Los requerimientos pedidos por la Secretaría de
Programación Económica y Regional para autorizar
el endeudamiento subnacional
a) Objeto y destino delfinanciamiento, indicando en su caso si se trata de refinanciamiento de una
deuda ya existente (Nota de la Provincia/Municipio solicitando el trámite a la Subsecretaría de
Programación Regional).
b) Propuesta de la entidadfinanciera que especifique claramente montos, modalidades de pago y
calendario de vencimientos de amortizaciones, intereses, gastos y garantías; debiendo acompañar
asimismo el proyecto de contrato de mutuo -donde conste el porcentaje de coparticipación afectado en
garantía- o prospectos de emisión en caso de bonos de deuda voluntaria.
c) Stock de deuda y perfil de vencimientos de capital e intereses actualizado.
d) Garantías. Informar el detalle (en porcentaje) de: a) Garantías sobre los recursos de
Coparticipación de Impuestos Nacionales o Provinciales, b) Otras garantías.
e) Estado de ejecución presupuestaria de recursos y gastos.
f) En cumplimiento de lo requerido en el punto c) se solicita Informe descriptivo del estado de las
operaciones acordadas con anterioridad en virtud de la Resolución MEyOSP N°1075/93, indicando el
grado de desembolso de las mismas, así también el grado de cancelación de las obligaciones
contraídas. Debiendo acompañar un detalle mes por mes, del total de las retenciones de los Recursos
de Coparticipación Federal (con la aclaración de los montos reintegrados por las Instituciones
Bancarias en los casos en que los porcentajes cedidos superen los servicios de amortización en
intereses).
g) Planta de personal abierto por escalafón.
h) Leyes que autoricen a) el endeudamiento a contraer y/o reestructuración de la deuda tomada y b)
la afectación de la coparticipación federal de los fondos que correspondan a la Provincia, en el caso de
que la misma se otorgue en garantía de operación. Dichas normas legales deberán ser dictadas en el
marco de la Constitución Provincial.
i) Autorización del Poder Ejecutivo Provincial para suscribir en representación de la Provincia la
operación a contratar, en el momento de su dictado (usualmente posterior a la sanción de las leyes
mencionadas).
j) Dictamen de la Fiscalía de Estado u organismo municipal competente.
k) Certificación del Contador General - u organismo municipal que corresponda- acerca de la
procedencia del porcentaje afectado de los niveles de coparticipación federal y de la factibilidad de la
operación conforme al régimen de administración financiera provincial (Ley de Contabilidad).
Asimismo, dicho organismo deberá certificar la efectiva afectación de los fondos recibidos por las
operaciones acordadas con anterioridad en virtud de la Resolución MEyOSP N° 1075/93, a los
destinos para los cuales fueron autorizadas.
1) La información deberá ser certificada por el organismo de contralor que corresponda (ej:
Contaduría y/o Tribunal de Cuentas y Escribanía General de Gobierno en su caso)
m) Asimismo deberá tenerse presente el inciso c) del Anexo a la Resolución MEyOSP N° 731/95, que
textualmente expresa: "En el caso que las negociaciones definanciamiento no lleguen a ser incluidas
de manera satisfactorias para las partes y por lo tanto no seafirmado el correspondiente contrato; las
partes no podrán reclamarse recíprocamente suma alguna en concepto de gastos y comisiones".
Sesión 1 - 152

SESION 2

INCENTIVOS TRIBUTARIOS PARA ATRAER
INVERSION EXTRANJERA DIRECTA:
EXPERIENCIAS RECIENTES Y
EVALUACION DE DISEÑOS ALTERNATIVOS

Sesión 2 - 153

GLOBALIZACION, IED Y NUEVAS ESTRATEGIAS
EMPRESARIALES EN AMERICA LATINA

Michael Mortimore
CEPAL

Sesión 2 - 155

Indice

I.

Panorama regional.

158

II.

México: fuente de eficiencia para las empresas transnacionales que operan en el
mercado de América del Norte
162

III. España: unos pocos inversionistas audaces dominan algunos de los principales sectores
de servicios de América Latina
164
IV.

Las consecuencias para América Latina de la reestructuración de la industria del
vestido de América del Norte
166

Sesión 2 -156

El informe del presente año comienza mencionando los importantes obstáculos que entorpecen la
comprensión cabal del fenómeno de la inversión extranjera, no sólo debido a las dificultades que se
plantean por el uso de las diferentes metodologías, prácticas contables, definiciones, cobertura nacional y
sectorial, etc. que aplican las distintas instituciones internacionales y nacionales que producen estadísticas
de inversión extranjera directa (IED), lo que se indicó en la edición de 1998, sino también porque existe
considerable confusión de carácter normativo o jurídico. Sin duda, la proliferación de tratados bilaterales de
inversión, sumada a las obligaciones multilaterales concertadas en el marco del Acuerdo General sobre el
Comercio de Servicios (GATS), el Acuerdo sobre las medidas de inversión relacionadas con el comercio
(MIC), el Acuerdo sobre los aspectos de los derechos de propiedad intelectual relacionados con el
comercio (ADPIC) y otros, y los capítulos sobre inversión de acuerdos de libre comercio, ya sea vigentes
-como el Tratado de Libre Comercio de América del Norte (TLC), el Acuerdo de Libre Comercio entre
Canadá y Chile, el Acuerdo de Complementación Económica entre Chile y México, el Tratado Bilateral
de Garantía a las Inversiones suscrito entre Bolivia y los Estados Unidos, etc.— o en etapa de negociación
-como el Área de Libre Comercio de las Américas (ALCA)--, ha contribuido a aumentar
significativamente los ingresos de IED a los países participantes. No obstante, estos instrumentos no han
ayudado mucho a lograr una visión más clara del fenómeno de la IED en sí. De hecho, es muy posible
que la plétora de nuevas obligaciones de carácter normativo o jurídico no coincida con ninguna iniciativa
multilateral unificada sobre inversión extranjera que pudiera acordarse en el futuro en el seno de la
Organización Mundial del Comercio (OMC), sobre todo si se trata de una iniciativa en que se reflejen
mejor las necesidades específicas de los países en desarrollo. En consecuencia, parecería aconsejable que
los países latinoamericanos ejercieran mayor cautela al contraer nuevas obligaciones hasta tanto se aclare
la situación multilateral y se llegue a una mejor comprensión del fenómeno de la IED como tal. Quizá sea
buen momento para que los encargados nacionales de adoptar decisiones se concentren en definir
aquellas actividades prioritarias en que estiman que la IED puede contribuir sustancialmente a los
objetivos nacionales de desarrollo, para luego trazar y aplicar políticas que favorezcan esos tipos de IED
en particular.
Los desafíos estadístico y normativo vinculados a la IED no afectaron en absoluto la afluencia de
inversión extranjera a América Latina y el Caribe, ya que los flujos netos correspondientes a 1998 (y las
estimaciones disponibles para 1999) superaron a los de 1997 y ascendieron a más de 76 000 millones de
dólares. En el capítulo titulado Panorama regional del presente informe se proporciona información más
detallada sobre los ingresos de IED y demás aspectos estadísticos conexos de la IED en la región.
Asimismo, se dedican capítulos separados al análisis exhaustivo del país receptor (México), el país
inversionista (España) y la rama de actividad (el vestido) destacados en la edición de este año.

Sesión 2 -157

I.

Panorama regional

En 1998 los flujos de inversión extranjera directa mundial superaron los 650 000 millones de
dólares, lo que representó un incremento de casi 39% respecto del año anterior y más del doble del
promedio correspondiente a 1991-1996. Estas cifras indican que continúa la aceleración de la expansión
internacional de las empresas transnacionales en el contexto del proceso de globalización. Los ingresos de
IED correspondientes a 1998 estuvieron mucho más concentrados que antes en los mismos países
industrializados (más del 70% en comparación con alrededor del 60% durante todos los años noventa) que
constituyeron los países de origen de cerca del 90% de las salidas de capital. De esta manera se invirtió la
tendencia ascendente registrada hasta entonces, según la cual los ingresos a los países en desarrollo
aumentaban, al recibir estos menos del 29% del total de losflujos, en comparación con el 40% que habían
captado a mediados de los años noventa.
En términos absolutos, el nivel de losflujos de IED que ingresaron a los países en desarrollo se
redujo ligeramente durante 1998, aunque en la distribución entre las regiones se aprecia que América Latina
y el Caribe se puso casi a la par de Asia en desarrollo (40.6% frente a 44.3%) y se alejó mucho de otras
regiones en desarrollo (15.1%). Indudablemente la crisis asiática que comenzó en 1997 afectó los ingresos
de IED destinados a Asia oriental, meridional y sudoriental (que disminuyeron alrededor del 14%, al pasar
de 88 366 a 76 292 millones de dólares), y especialmente losflujos dentro del continente asiático. Al mismo
tiempo, la IED que ingresó a América Latina y el Caribe se mostró muy dinámica (con un aumento de casi
11%, al pasar de 69 404 a 76 727 millones de dólares). Así pues, si bien los países en desarrollo en general
perdieron terreno frente a los países industrializados como destino de la IED durante 1998, la región de
América Latina y el Caribe resultó ser la excepción y registró un sostenido dinamismo.
En 1998 los países de mayor tamaño de la región que pertenecen a la Asociación Latinoamericana
de Integración (ALADI) recibieron 84% del total de ingresos de IED a la región, en tanto los centros
financieros (8.5%) y el resto de los países (7.5%) de la Cuenca del Caribe captaron crecientes volúmenes en
comparación con la tendencia de los años noventa. Entre los países más grandes, Brasil registró la mayor
afluencia de IED (41.6%), seguido a gran distancia por México (13.3%), Argentina (8%), Chile (6%) y
Venezuela (5.8%). Losflujos de IED que ingresaron a los centrosfinancieros se concentraron en las Islas
Caimán (4.6%), Bermuda (3.1%) y Panamá (1.6%). Los que se dirigieron a otros países de la Cuenca del
Caribe tuvieron por principal destino El Salvador (1.1%), Trinidad y Tabago (1%), la República
Dominicana (0.9%), Guatemala (0.9%) y Costa Rica (0.7%).
Según las cifras preliminares correspondientes a 1999, Brasil ha seguido siendo el foco principal de
la IED en la región. Principalmente gracias a la venta de Yacimientos Petrolíferos Fiscales (YPF) a la
empresa española de petróleo Repsol, Argentina superará a México y conquistará el segundo lugar entre los
países de la ALADI. Entretanto, Chile duplicará aproximadamente sus ingresos de IED en 1999, en gran
medida como consecuencia de la compra de Enersis y Endesa Chile por parte de la empresa española de
electricidad, Endesa España. Así pues, las fusiones y adquisiciones, en particular a cargo de empresas
españolas, han seguido siendo un factor muy dinámico en los ingresos de IED a los países más grandes de
América Latina en 1999.
Habida cuenta de que los ingresos netos de IED a la región se concentraron en tal medida en Brasil,
cabe puntualizar que cerca de la mitad de la IED que recibió ese país estaba dirigida a unos pocos sectores
de servicios, en que se estaban vendiendo activos estatales como parte de los procesos de liberalización y
desregulación de dichos sectores (telecomunicaciones y electricidad) o en que la venta de empresas privadas
nacionales era un factor fundamental (serviciosfinancieros y comercio al menudeo). En total, 37% de los
ingresos generados por el proceso de privatización brasileño provino del sector de las telecomunicaciones y

Sesión 2 -158

33% del sector de la generación y distribución de electricidad. Las empresas europeas, sobre todo las
españolas, tuvieron una participación muy activa en Brasil en estos ámbitos.
INGRESOS NETOS DE INVERSIÓN EXTRANJERA DIRECTA A AMÉRICA LATINA,
POR SUBREGION, 1990-1999

(En

Subregión/País

199019
94

1. Centroamérica y el Caribe 1397
2. Centros financieros del
2 506
Caribe
3. Asociación
1 28
4 3
Latinoamericana de
Integración (ALADI)

millones

de dólares

19
95

19
96

19
97

1998

Participación
1998 (%)

1999a

205
0

2 18
0

421
5

5 776

7.5

550
0

2 427

3 19
1

453
1

6 486

8.5

500
0

2 750
7

4 46
1 1

6 60
0 4

6 45
4 6

84.0

7 40
5 2

653
1
474
11200
467
3
3 13
2
491
9 16
8
26
4
326
2
17
3
2 13
8

804
9
71
3
1 60
9 5
5 219
573
0
65
9
1 81
2 3
270
1 75
8
16
2
556
3

6150
82
7
3 93
1 1
468
3
308
3
81
3
1 28
0 3
26
5
1 90
3
14
6
445
3

9.5
1.4
49.5
7.2

3.0
0.3
6.9

2 00
1 0
80
0
31000
890
0
350
470
1 00
0 0
10
0
1500
10
0
1 20
0

4 63
6 4

69404

7 77
6 2

1 00
0.

y porcentajes

8 90
5 2

Argentina
B olivia
Brasil
Chile
Colombia
Ecuador
México
Paraguay
Perú
Uruguay
Venezuela

291
7
8
5
1 73
0
1219
88
1
23
9
5 430
9
8
75
8
86
3

5 279
393
489
5
297
5
969
470
9 56
2
15
5
2 000
17
5
985

Total (1+2+3)

1 20
8 2

32182

4.7

1.3 •
15.9

0.4

Fuente: CEPAL, Centro de Información de la Unidad de Inversiones y Estrategias Empresariales de la División de
Desarrollo Productivo y Empresarial, sobre la base de información proporcionada por el Fondo Monetario Internacional, la
Conferencia de las Naciones Unidas sobre Comercio y Desarrollo (UNCTAD) y los bancos centrales de cada país.
a Cálculos de la Unidad de Inversiones y Estrategias Empresariales, sobre la base de proyecciones de los bancos centrales de
los países correspondientes u otras fuentes, a diciembre de 1 9 .
99

En la región en general, no cabe duda de que, como consecuencia del proceso de globalización y de
las políticas nacionales tendientes a facilitar nuevas operaciones de las empresas transnacionales
(liberalización comercial yfinanciera y desregulación de ciertos sectores), el papel que estas desempeñan ha
cobrado enorme importancia. Este hecho queda claramente demostrado al comparar los activos y las
actividades sectoriales de las 500 empresas más importantes en términos de ventas de América Latina
durante la década de 1990. Al desglosar las ventas totales correspondientes a este grupo de grandes
empresas, clasificadas en tres categorías -empresas transnacionales, empresas privadas locales y empresas
estatales—, se advierte que la participación de las empresas transnacionales ha registrado un brusco
aumento de 26.6% a 38.7%, la de las empresas privadas locales se mantuvo constante, en alrededor del
40%, y las ventas de las empresas estatales cayeron estrepitosamente, de 35% a 19.1%. En términos de sus
actividades sectoriales, la manufactura se mantuvo estable, en cerca del 42% de las ventas totales, en tanto
los servicios saltaron de 30% a 41.2% y las actividades primarias bajaron de 27.8% a 17.4%. Estas cifras
confirman decididamente la idea de que la búsqueda de acceso a los mercados en materia de servicios ha

Sesión 2 -159

sido una de las estrategias dominantes de las empresas transnacionales de la región durante los años
noventa. El estudio de caso sobre Brasil quefigura en la edición de 1998 también corrobora esta conclusión.
De acuerdo con información similar correspondiente a las 100 empresas manufacutreras más
importantes en términos de ventas, se observa que la participación de las empresas transnacionales aumentó
de 52.8% a 60.7%, que la de las empresas privadas locales se mantuvo estable en cerca del 40% y que la de
las empresas estatales bajó de 5.2% a 1.2% de las ventas totales. Cabe señalar que la proporción de las
ventas de las empresas transnacionales del sector automovilístico aumentó de 25.1% a 30.4%, lo que indica
que alrededor de la mitad del total de las ventas de manufacturas de dichas empresas corresponde a esa
única industria.
Los datos disponibles sobre el sector manufacturero se complementan con información sobre el
valor de las ventas extemas de los 200 principales exportadores de América Latina entre 1995 y 1998 (no se
dispone de datos para años anteriores). Según las cifras, el sector manufacturero se mantuvo estable con el
50% del valor de las exportaciones de estas empresas, en tanto la participación de las actividades primarias
declinó de 44.8% a 40.1% y la de los servicios aumentó de 5.3% a 9.2%. Las empresas transnacionales
fueron con creces el componente más dinámico de este grupo de exportadores principales y su participación
en el total de exportaciones trepó de 30.6% a 44.8% durante 1995-1998. Asimismo, la participación de las
empresas transnacionales del sector automotor creció de 15.2% a 20.3%, lo que nuevamente subraya la
importancia de esta industria en la región. Las empresas privadas locales mantuvieron una participación de
alrededor del 33% de las exportaciones, en tanto la de las empresas estatales cayó abruptamente del 37% al
21.9% del total de exportaciones de estas empresas principales. Estas cifras ratifican innegablemente la idea
de que el uso de la estrategia de búsqueda de eficiencia en América Latina para la exportación de
manufacturas ha contribuido muchísimo a fortalecer la presencia de las empresas transnacionales en la
región. El análisis de la industria automotriz mexicana quefigura en el capítulo IV de la edición de 1998, el
examen de la industria del vestido de México y la Cuenca del Caribe contenido en el capítulo IV de la
presente edición y el estudio de caso sobre México que aparece en el capítulo II del informe de 1999
confirman dicha tesis.
El estudio de las ventas consolidadas de las 100 principales empresas transnacionales no financieras
que operaban en seis países de América Latina en 1998 indica que se están produciendo cambios
significativos. Las subsidiarias de las empresas transnacionales con base en Estados Unidos siguen siendo el
grupo más numeroso (43.1% del total de las ventas consolidadas); sin embargo, varias empresas
transnacionales europeas, procedentes de Alemania (10.6%), España (10%), Francia (9.2%) e Italia (5%),
también han establecido una presencia destacable en la región. En el capítulo m del presente informe se
examina en detalle la nueva relevancia de lasfiliales de empresas españolas. La industria automotriz (23.6%
de las ventas consolidadas) es aún la rama de actividad donde se advierte la mayor presencia de las
empresas transnacionales, en tanto el comercio (11.8%), la minería y el petróleo (10%), las
telecomunicaciones (9.9%), la electrónica (9.5%) y la electricidad (8.1%) son dignos de mención en este
sentido. Más del 70% de las ventas consolidadas de las 100 empresas transnacionales principales de la
región correspondió a empresas que figuran entre las 500 clasificadas por la revista Fortune. No obstante,
en términos de los activos extemos, sólo algo más del 50% de las ventas consolidadas de estas 100
empresas transnacionales proviene de las principales empresas transnacionales que figuran en la
clasificación de la UNCTAD, lo que sugiere que las recién llegadas contribuyen en gran medida a los flujos
de IED que ingresan a América Latina. Lasfiliales de empresas europeas, sobre todo españolas, en el sector
de los servicios representan un elemento dinámico de cambio respecto de los principales agentes
económicos transnacionales que operan en la región.
La situación de los bancos extranjeros también ha cambiado drásticamente en América Latina.
Aunque hay relativamente pocos bancos mundiales con redes extensas en la región, los activos del Banco
Sesión 2 -160

Santander Central Hispano (BSCH) en seis países latinoamericanos son casi tan cuantiosos como las redes
de los dos bancos extranjeros que le siguen en tamaño, el Citibank y el Banco Bilbao Vizcaya Argentaría
(BBVA), consideradas juntas. Los bancos españoles representan 32.5% de los activos totales de los 20
bancos extranjeros más grandes de América Latina, superando a los estadounidenses, cuya participación
asciende a 29.7%. Nuevamente, lasfiliales de los bancos españoles actúan como elementos dinámicos de la
expansión de los bancos transnacionales en la región.
En términos de modalidades, la IED que ingresó a la región durante L998 y 1999 mantuvo su
composición de alrededor de 60% para la creación de nuevos activos y 40% para la compra de activos
existentes, si bien el volumen de la IED creció sensiblemente. Los datos de la CEPAL indican que las
fusiones y adquisiciones aumentaron en gran medida durante 1998 y 1999, en tanto el valor de las
privatizaciones bajó considerablemente en el mismo período. Las modalidades de la IED constituyen un
aspecto importante del análisis de este tipo de inversión en la región.
Las fusiones y adquisiciones registradas en 1998 y 1999 en América Latina estuvieron sumamente
concentradas en tres países: Argentina (45.6%), Brasil (26.4%) y Chile (16.5%). La mayor parte de la
inversión extranjera tuvo su origen en España (53%) y algunos otros inversionistas extranjeros: Estados
Unidos (15%), Francia (9%) y los Países Bajos (7.7%). Catorce operaciones en gran escala, por más de 500
millones de dólares cada una, representaron 63.9% del valor total de las fusiones y adquisiciones. De hecho,
las cinco operaciones de mayor envergadura representaron casi la mitad (49%) del total. Sin duda la
operación más cuantiosa fue la adquisición de la empresa argentina Yacimientos Petrolíferos Fiscales (YPF)
por parte de la empresa petrolera española, Repsol, por más de 13 000 millones de dólares en 1999. Otras
transacciones importantes fueron la compra de Endesa Chile y Enersis de Chile por parte de Endesa España,
en dos operaciones separadas. También entre las. cinco operaciones principalesfiguran la adquisición del
Banco Real de Brasil por parte de ABN Amro de los Países Bajos y del Banco Excel de Brasil por parte de
BBVA de España.
La mayoría de las privatizaciones que se realizaron en América Latina en 1998 y 1999 tuvieron
lugar en sólo dos países: Brasil (62.4%) y Argentina (18%). La participación de países inversionistas fue
mucho más diversificada que en el caso de las fusiones y adquisiciones. Los tres principales países
inversionistas fueron los Estados Unidos (14.8%), España (8.7%) y Portugal (8.4%). No obstante, 24
operaciones por más de 500 millones de dólares cada una representaron más del 80% de todas las
transacciones de este tipo. Las cinco privatizaciones más cuantiosas correspondieron al 39.5% del total. La
concesión de aeropuertos en Argentina, otorgada a inversionistas de los Estados Unidos, Italia y Argentina
fue avaluada en 5 134 millones de dólares, en tanto las otras cuatro privatizaciones se registraron en Brasil y
entrañaron operaciones en los sectores de las telecomunicaciones y la electricidad: Telefónica de España y
varios otros inversionistas compraron Telesp; Portugal Telecom adquirió Telesp Cellular; AES y Houston
Energy de los Estados Unidos se unieron a Electricité de France para comprar Eletropaulo Metropolitana de
Eletricidade, y EMBRATEL fue adquirida por MCI Worldcom.
En 1998, dos países -Costa Rica y Bolivia— habían logrado en gran medida elaborar y aplicar
nuevas políticas destinadas a atraer un volumen de ingresos de IED mucho mayor hacia actividades
prioritarias definidas en sus estrategias de desarrollo. En el caso de Costa Rica, cuya afluencia de IED
aumentó a 531 millones de dólares en 1998 (equivalente al 6% del PIB) y donde 64% del acervo de IED
existente había ingresado al país en los años noventa, la nueva política se basó en contactar directamente a
las empresas transnacionales cuyas estrategias consistían en la búsqueda de eficiencia en actividades de alta
tecnología, especialmente electrónica. Previamente Costa Rica había conseguido atraer IED a la industria
del ensamblaje de prendas de vestir ofreciendo incentivos horizontales (fiscales) en zonas de procesamiento
de exportaciones (ZPE); sin embargo, el impacto en términos de desarrollo del éxito de las ZPE fue un
fracaso debido a la falta de progreso suficiente en cuanto a los niveles de los salarios, la capacitación de los
Sesión 2 -161

recursos humanos y la falta de integración de tales actividades a la economía nacional, aparte del hecho de
que hubo otros países de América Central que empezaron a ofrecer menores salarios y otros incentivos que
a la larga hicieron perder dinamismo a la industria de ensamblaje de prendas de vestir de Costa Rica. No
obstante, las autoridades nacionales dieron un golpe maestro al tomar medidas directas y utilizar incentivos
selectivos, con lo cual consiguieron atraer un proyecto de inversión por un valor de 400 a 500 millones de
dólares con el gigante de los microprocesadores, Intel, que instaló una planta de ensamblaje y ensayo en el
país. Esa inversión por sí sola ascendió en 1999 a más del 40% del valor de las exportaciones totales y más
del 90% del valor de las exportaciones de productos electrónicos de Costa Rica. Si bien el solo volumen e
impacto económico de este proyecto de IED es formidable, la nueva política de Costa Rica no debiera
considerarse consolidada hasta tanto se hayan conseguido varias inversiones de este tipo y el conglomerado
nacional de electrónica esté mejor afianzado en el país.
Bolivia atrajo 900 millones de dólares de IED en 1998, lo que equivale a 10% de su PIB. Más del
70% del acervo de IED de ese país ingresó durante la década de 1990. Como parte de la estrategia de
desarrollo denominada Plan de Todos, el Plan de Capitalización, que disponía la venta de activos estatales,
y la Ley Corazón, que preveía beneficios especiales para las exportaciones de energía de zonas de frontera,
se establecieron las bases de una nueva política tendiente a canalizar mayores volúmenes de IED hacia
actividades prioritarias, convirtiendo así a Bolivia en un gran exportador de energía. La nueva política
encaja bien en las estrategias de búsqueda de recursos naturales que aplican las principales empresas
transnacionales, especialmente aquellas interesadas en establecer una red regional de energía. El programa
de privatización captó más del 50% de las entradas de IED de 1995-1998. Lo más novedoso de la política de
Bolivia fue el uso de la IED para la capitalización de empresas mixtas (el 50% de la propiedad quedó en
manos de ciudadanos bolivianos). Entre las grandes empresas estatales que se vendieron figuran la empresa
petrolera, Yacimientos Petrolíferos Fiscales Bolivianos (YPFB), con 55.2% del Plan, la refinería estatal de
metales (Vinto), la Empresa Nacional de Electricidad (ENDE), la Empresa Nacional de
Telecomunicaciones (ENTEL), la Empresa Nacional de Ferrocarriles (ENFE) y la línea aérea estatal, el
Lloyd Aéreo Boliviano. No obstante, el éxito más saliente de la nueva política estuvo en el campo de la
energía donde, además de los nuevos contratos de exploración y operación de pozos de petróleo, la Ley
Corazón ofrecía considerables incentivos para crear condiciones parecidas a una zona de libre comercio
(concesiones por 40 años, importaciones de bienes de capital libres de impuestos, eliminación del impuesto
al valor agregado, etc.), a fin de fomentar las exportaciones de energía a países vecinos, especialmente
Brasil. En 1999, las exportaciones de gas natural a Brasil constituyeron el 26% del valor total de las
exportaciones bolivianas.
El hecho de que Costa Rica y Bolivia hayan logrado canalizar mayores volúmenes de IED hacia
actividades prioritarias indica que la política de IED supone mucho más que la simple liberalización,
desregulación y privatización sumadas a la mayor protección de los inversionistas extranjeros en el marco
de acuerdos de inversión bilaterales, subregionales o multilaterales. Como primera medida, es preciso tener
un plan, es decir, una estrategia de desarrollo en que se definan las actividades prioritarias.

II.
México: fuente de eficiencia para las empresas
transnacionales que operan en el mercado de América del Norte
En el siglo XX las autoridades mexicanas han tenido una relación extremadamente fluctuante con los
inversionistas extranjeros. Durante la mayor parte del período posterior a la Revolución Mexicana, el
Gobierno adoptó una postura nacionalista y restrictiva respecto de la IED, sobre todo cuando se nacionalizó
la industria del petróleo en los años treinta y cuando se promulgó la Ley para Promover la Inversión

Sesión 2 -162

Mexicana y Regular la Inversión Extranjera, en 1973. Dicha postura cambió radicalmente afines de los
años ochenta, cuando en 1989 se agregaron nuevas regulaciones a la mencionada ley y luego, en 1993, se
aprobó una ley completamente modificada. Esta última política daba gran impulso a la afluencia de IED y
estaba en consonancia con el papel más prominente que tenían las operaciones de las empresas
transnacionales como motores del crecimiento de la economía mexicana y la posibilidad de suscribir un
Tratado de Libre Comercio de América del Norte.
El ingreso de México al Acuerdo General sobre Aranceles Aduaneros y Comercio (GATT) en 1986
probablemente marcó el principio del alejamiento del marco vigente de industrialización mediante la
sustitución de importaciones con elfin de abordar una estrategia de crecimiento más orientada hacia afuera,
basada en nuevas exportaciones y en la liberalización y desregulación de la economía mexicana. La IED
habría de constituir un factor importante en esta nueva estrategia. El impacto que esta política tuvo en las
empresas transnacionales se manifestó de tres maneras principales. Primero, las operaciones de maquila
recibieron un fuerte espaldarazo de las políticas reactivadas, que facilitaron el uso más generalizado de
mano de obra barata mexicana en el ensamblaje con insumos importados de productosfinales (sobre todo
autopartes, equipo electrónico y prendas de vestir) destinados al mercado de los Estados Unidos. Segundo,
las subsidiarias existentes en el sector manufacturero que habían formado parte del anterior marco de
política de industrialización mediante la sustitución de importaciones se vieron obligadas a reaccionar,
retirándose del mercado interno cada vez más competitivo, racionalizando sus operaciones de producción en
México (generalmente adoptando una estrategia defensiva de reducir su tamaño) o reestructurando su
sistema de producción general (a menudo haciendo grandes inversiones nuevas para redefinir el papel de
México en sus sistemas de producción internacionalmente integrados). La reestructuración de la industria
automotriz fue un acontecimiento que marcó un hito en este sentido. En tercer lugar, el mayor acceso a los
mercados atrajo a varias empresas transnacionales que constituían nuevos entrantes en el sector de los
servicios de México (comercio al menudeo, telecomunicaciones, serviciosfinancieros, etc.), así como en el
sector manufacturero (alimentos, bebidas, tabaco, etc.). La firma del TLC, en virtud del cual Canadá,
Estados Unidos y México se unieron en una economía de creciente corte continental, consolidó los cambios
que ya se estaban produciendo en México. Durante el período comprendido entre 1990 y 1995, este país fue
el principal receptor de IED de América Latina, y la mayor parte de esos flujos provinieron de sus socios en
el Tratado.
Los resultados fundamentales de la explosión de ingresos de IED hacia México pueden
comprenderse mejor en términos de las principales estrategias empresariales aplicadas por los inversionistas
extranjeros, a saber, la búsqueda de eficiencia y la búsqueda de acceso a los mercados. Frente a la feroz
competencia que planteaban las inportaciones asiáticas de automóviles, computadoras y prendas de vestir
en el mercado estadounidense, las empresas transnacionales de ese país han sido las que en mayor número
han instalado plantas de producción y ensamblaje en México, utilizando la mano de obra nacional más
barata y aprovechando el mecanismo de producción compartida de los Estados Unidos y el TLC con el fin
de reducir los costos de producción y estar mejor preparadas para competir en el mercado de América del
Norte. En consecuencia, México pasó a ser el tercer proveedor en orden de importancia en ese mercado y
las exportaciones de las empresas transnacionales representan más de la mitad del total de exportaciones
mexicanas con tal destino. México se ha especializado en los artículos de mayor dinamismo en el mercado
de América del Norte (manufacturas), aumentando al mismo tiempo su participación en el mercado de
importación. Dos tercios de la IED que recibe provienen de los Estados Unidos y Canadá, lo que revela el
carácter continental de los sistemas de producción internacionalmente integrados que se están
estableciendo. La creación de nuevos activos (plantas y equipos) ha predominado mucho más que la compra
de activos existentes.
La estrategia de búsqueda de acceso a los mercados ha estado vinculada principalmente con la
liberalización y desregulación de algunos sectores de servicios (comercio al menudeo, telecomunicaciones,
Sesión 2-163

servicios financieros, etc.), en que la IED se ha originado en empresas transnacionales europeas, así como
estadounidenses y canadienses, y ha estado concentrada en la compra de activos existentes más que en la
creación de nuevos. Si bien las exportaciones directas fueron escasas en los sectores de los servicios, esta
IED ha acarreado una mejora en el aspecto sistémico de la competitividad internacional de la economía
mexicana.
Algunos de los beneficios evidentes del papel más preponderante que ha tenido la IED en la
economía mexicana se encuentran en su mayor integración a la economía internacional. Las exportaciones
han registrado un fuerte aumento como proporción del PEB y también ha crecido la parte correspondiente a
la IED en la formación bruta de capitalfijo. Varias áreas del sector manufacturero se han visto fortalecidas
y han adquirido competitividad internacional. Así pues, la competitividad sistémica de la economía
mexicana ha mejorado gracias a la IED en los servicios.
No obstante, estos éxitos han estado acompañados porfracasos en otros ámbitos. El dinamismo del
segmento de la economía mexicana dominado por la IED no bastó para aumentar significativamente la tasa
de crecimiento global a largo plazo de la economía, que se polarizó cada vez más debido a la escasa
integración de los sectores modernos y más tradicionales. El éxito de México en términos de lograr la
competitividad internacional ha estado restringido a un solo mercado: América del Norte. El proceso de
industrialización nacional se vio truncado en ciertas instancias porque las empresas transnacionales que
buscaban eficiencia recurrieron sobre todo a insumos físicos importados y el mecanismo de producción
compartida de los Estados Unidos castigaba a los insumos mexicanos. Además, en los sectores de las
telecomunicaciones y los serviciosfinancieros se tropezó con ciertos problemas de tipo regulatorio.
Este análisis indica que los encargados de formular políticas de México podrían aprovechar los
éxitos alcanzados e intentar superar los problemas que existen. Es preciso hacer mayores esfuerzos para
diversificar los mercados de exportación (como Europa y América Latina) y las fuentes de IED (Europa y
Japón). La integración de los sectores modernos y tradicionales de la economía mexicana es vital para
lograr un éxito completo. Este objetivo puede alcanzarse en parte sacando mejor partido de las normas de
origen del TLC que, a diferencia del mecanismo de producción compartida de los Estados Unidos, permiten
la incorporación de insumos mexicanos competitivos en los productos finales destinados a los mercados de
exportación. De tal manera, es necesario aplicar un programa mucho más enérgico para alentar y apoyar a
los proveedores mexicanos, además de aumentar los esfuerzos para mejorar los recursos humanos mediante
la capacitación

III.

España: unos pocos inversionistas audaces dominan algunos de
los principales sectores de servicios de América Latina

La economía española rápidamente se ha puesto a tono en su integración al sistema internacional.
Anteriormente el país hacía las veces de puerta de entrada al mercado europeo, al acoger las plataformas de
exportación de los inversionistas extranjeros, pero ahora ha pasado a ser un inversionista extranjero en sí
mismo, que además concentra cada vez más su IED en América Latina y, más especialmente, en el Mercado
Común del Sur (Mercosur), Chile y los países andinos. La participación que le cupo a América Latina en el
total de la IED española saltó de 29% a 72% entre 1990 y 1998. Una proporción muy considerable de esa
inversión se destinó a los servicios, como las telecomunicaciones, la generación y distribución de
electricidad (electricidad, gas, petróleo) y los serviciosfinancieros. La mayor parte correspondió a unas
pocas empresas transnacionales españolas.

Sesión 2 -164

El proceso de globalización, entendido como el desplazamiento en el largo plazo hacia un único
mercado universal, parece ser el motor de la expansión internacional de estas empresas españolas. Tienen
prisa por establecer sus sistemas internacionales y consideran una oportunidad estratégica el hecho de que
los sectores en que operan estén atravesando por procesos de liberalización y desregulación en el marco del
Acuerdo General sobre el Comercio de Servicios (telecomunicaciones y servicios financieros) o las
Directrices de la Comisión Europea (electricidad) o por reestructuraciones mundiales a fuerza de
megafusiones y megaadquisiciones (petróleo y banca). En estos sectores de servicios tanto los actores
mundiales existentes como los aspirantes a actores utilizan actualmente la IED y las alianzas estratégicas
como medio para posicionarse.
En el caso de las empresas transnacionales españolas que ingresan a la escena internacional, se
han producido cambios muy significativos en las situaciones competitivas de los mercados internacional,
europeo, español y latinoamericano, que han obligado a los aspirantes a actores mundiales a aplicar
estrategias extremadamente audaces de expansión internacional. Algunas de las mayores empresas
españolas en el área de las telecomunicaciones (Telefónica), los serviciosfinancieros (BSCH y BBVA),
la generación y distribución de energía (Endesa España, Iberdrola, Repsol) y otros sectores de servicios
(Aguas de Barcelona, Iberia) están aprovechando los vínculos culturales y lingüísticos para lanzar sus
campañas de expansión internacional principalmente en América Latina. La estrategia empresarial
fundamental que utilizan es obtener acceso al mercado de los servicios que son objeto de liberalización,
desregulación y privatización en la región.
Telefónica, la empresa de telecomunicaciones privatizada de España, ha pasado a ser la mayor
empresa de telecomunicaciones en español del mundo, lo que ha logrado sobre todo al aprovechar la
privatización de las empresas estatales de telecomunicaciones de América Latina. Por ejemplo, su
clasificación entre las 500 empresas mundiales de la revista Fortune pasó del puesto 239 en 1997 al 193 en
1998. Tras invertir más de 10 000 millones de dólares en la región, principalmente mediante fusiones y
adquisiciones, el 31% del total de sus activos es ahora extranjero y cuenta con 49 millones de clientes en
Argentina, Brasil, Chile, Perú, América Central y Puerto Rico. La empresa proyecta invertir otros 20 000
millones de dólares en América Latina afin de consolidar su posición predominante en el mundo de habla
hispana y portuguesa y ha establecido alianzas estratégicas con MCI Worldcom y Portugal Telecom, así
como otras empresas transnacionales españolas (Iberdrola y BBVA).
Los bancos españoles BSCH y BBVA, de por sí productos de fusiones ocurridas durante los años
noventa, han estado a la vanguardia del proceso de consolidación que se está desarrollando en el sector
financiero español. El 45% del total de activos del sector corresponde a los tres bancos más importantes de
España y el 67% a los seis más importantes. En Europa, los bancos nacionales han recurrido al
establecimiento de una sólida red nacional, con lo cual lograron evitar ser adquiridos por instituciones
extranjeras. Estos dos bancos españoles han hecho justamente eso y, además, se han expandido hacia
América Latina, donde junto con el Citicorp poseen las redes más extensas. En la clasificación internacional
de los 1000 bancos principales publicada en The Banker, el BSCH y el BBVA pasaron del puesto 40 al 37 y
del 50 al 47, respectivamente, entre 1997 y 1998. Ambos bancos invirtieron más de 10 000 millones de
dólares en la región y, además de tener vastas redes regionales, con activos que superan los 20 000 millones
de dólares y más de 16 millones de clientes,figuran entre los líderes en la administración de fondos de
pensiones privados. No obstante, cabe señalar que, de los 50 grandes bancos europeos, estos bancos
españoles se ubicaron entre los tres con clasificación deriesgomás alta en febrero de 1999.
Endesa España e Iberdrola han encabezado la campaña de expansión internacional de las empresas
españolas de electricidad. Endesa es la principal empresa eléctrica de España y América Latina yfigura en
el tercer puesto en la Unión Europea. Invirtió más de 8 000 millones de dólares en 12 países durante el
período 1992-1998 y adquirió grandes operaciones en siete países latinoamericanos; en la mayoría de los
Sesión 2 -165

casos llevó a cabo estas transacciones participando en los procesos de privatización locales, aunque también
ha establecido alianzas estratégicas, como fue originalmente el caso de Enersis de Chile. En total, el 40% de
sus utilidades se genera en América Latina, donde posee 25 millones de clientes. Iberdrola, por su parte, ha
invertido más de 2 000 millones de dólares en seis países de América Latina y tiene alrededor de 15
millones de clientes en esa región.
Repsol, la empresa petrolera española privatizada, ha invertido más de 19 000 millones de dólares
en América Latina, siendo su adquisición más reciente la anteriormente privatizada empresa petrolífera
argentina, Yacimientos Petrolíferos Fiscales (YPF). Ha participado en fusiones y adquisiciones de empresas
petroleras o adquirido concesiones para la exploración de petróleo en otros ocho países de la región. En
gran parte gracias a sus activos latinoamericanos, Repsol se está convirtiendo en un grupo petrolífero de
creciente importancia en el mercado internacional, cuya clasificación entre las 500 empresas de Fortune ha
pasado del puesto 257 en 1997 al puesto 224 en 1998.
De modo que, en conjunto, la IED española en América Latina y el Caribe durante la década de
1990 ha sido muy importante y ha tenido considerables repercusiones. Estos seis aspirantes a actores
mundiales, muchos de los cuales son a su vez empresas privatizadas, han aprovechado los procesos de
privatización aplicados en varios sectores de servicios de América Latina para expandirse
intemacionalmente. Han concentrado sus fusiones y adquisiciones en el Mercosur, Chile y los países
andinos y, en algunos casos, han utilizado a su favor el hecho de que el Gobierno español sigue siendo un
importante accionista de las empresas. Si bien en el ámbito local ha surgido cierta preocupación de que esta
acelerada expansión forme parte de la "reconquista española" de sus antiguas colonias, las dificultades más
serias que ha planteado este proceso tienen su origen en la fragilidad e imperfección de los marcos
regulatorios de los países receptores, los efectos que produce en la competencia y el desafío financiero que
entraña una campaña de tan gran escala para estas empresas españolas, dado el pobre desempeño de las
economías en que se desarrolla su actividad.

IV.

Las consecuencias para América Latina de la reestructuración
de la industria del vestido de América del Norte

En general las importaciones de los países industriales han registrado gran dinamismo, y uno de los
componentes más destacables han sido las prendas de vestir, cuyas importaciones al mercado de América
del Norte han aumentado del 2.5% al 4.7% del valor de las importaciones totales de ese mercado entre 1980
y 1996. Actualmente las importaciones representan más de la mitad de las ventas de prendas de vestir en los
Estados Unidos, pese al uso prolongado de cuotas unilaterales por parte de ese país en el marco del Acuerdo
Multifibras y al hecho de que la industria del vestido está entre las que goza de mayor nivel de protección
arancelaria en el mercado estadounidense (15.5% en 1997).
Los países asiáticos han estado a la vanguardia del auge de exportaciones de prendas de vestir en
todo el mundo. Varias de las economías de reciente industrialización de Asia (la región administrativa
especial de Hong Kong, la provincia china de Taiwán, la República de Corea) han podido desarrollar fuertes
industrias textiles y al mismo tiempo desempeñarse como intermediarias de grandes compradores
internacionales que se acostumbraron a recurrir a ellas para producir prendas de vestir "de paquete
completo" sobre la base de los diseños del comprador, especialmente en los segmentos más dinámicos de la
producción del vestido, como la alta costura. Algunas incluso han producido sus propias marcas originales.
Otros países asiáticos han hecho las veces de sitios de ensamblaje para estos sistemas basados en
intermediarios. En 1996 China ostentó una impresionante participación en las importaciones de todos los

Sesión 2 -166

grandes mercados, que ascendió al 17.7% en el mercado de la Organización de Cooperación y Desarrollo
Económicos (OCDE), 8.4% en el de Europa, 15.3% en el de América del Norte, y 58% en el de Japón. Sin
embargo, este desafío asiático provocó una fuerte reacción de los productores del vestido y algunos
compradores de los Estados Unidos, cuya consecuencia más evidente fue que la participación de los países
asiáticos en las importaciones de prendas de vestir estadounidenses bajó de 80% en 1980 a 44% en 1996,
mientras que la de los países de América Latina saltó de 6% a 28% en el mismo período. En efecto, lo que
hicieron los productores y compradores de prendas de vestir de los Estados Unidos fue buscar mayor
eficiencia para sus subsidiarias, filiales, asociados y proveedores de América Latina (principalmente
México y la Cuenca del Caribe), aprovechando los bajos salarios a fin de competir mejor en el mercado de
América del Norte.
Inicialmente las autoridades estadounidenses facilitaron el uso de esta estrategia por parte de las
empresas locales del vestido mediante un mecanismo de producción compartida (partida 807 de la Lista de
Aranceles de los Estados Unidos, posteriormente partida 9802 del Sistema Armonizado de Aranceles) que
permitía la reimportación de productosfinales ensamblados en países con regímenes de salarios más bajos a
partir de componentes fabricados en los Estados Unidos y disponía la aplicación de derechos solamente
sobre el valor agregado fuera de ese país. Los países proveedores de América Latina recibieron acceso
preferencial al mercado estadounidense y los gobiernos de los demás países completaron la nueva situación
competitiva ofreciendo incentivos tributarios y arancelarios en forma de zonas de libre comercio, así como
otras ventajas. Entre 1993 y 1997, el total de importaciones de prendas de vestir que ingresó a los Estados
Unidos gracias al mecanismo de producción conpartida aumentó de 10% a 21% del total de importaciones
de prendas de vestir y un 90% de las importaciones que ingresaron gracias a este mecanismo se originó en
México (36.6% durante 1997) y los países de la Cuenca del Caribe (55.9%). Menos del 1% del valor de
dichas importaciones procedía de los países asiáticos. Más del 80% de todas las prendas de vestir que
México y los países de la Cuenca del Caribe exportaron a los Estados Unidos ingresó mediante el
mecanismo de producción compartida.
Las disposiciones del TLC aplicables a las prendas de vestir produjeron unafisura en términos de
las consecuencias que tuvo para América Latina la reestructuración de la industria del vestido de América
del Norte, dado que el TLC otorga a México ventajas que no comparten los países de la Cuenca del Caribe.
Primero, los derechos arancelarios implícitos con que se gravan las prendas de vestir mexicanas son
notablemente inferiores (0.9%, comparado con 8.5% para los países del Caribe). En segundo lugar, se
inponen menos cuotas para México, que puede realizar ciertos procesos de terminación que aportan valor
agregado (blanqueado, tintura, lavado de desgaste con piedra, planchado permanente, etc.). Tercero, según
las normas de origen, los insumos mexicanos se consideran insumos del TLC, lo que ha dado lugar al
surgimiento de una industria textil y del vestido más integrada. En consecuencia, la participación de México
en las importaciones de prendas de vestir de los Estados Unidos se disparó de 7.4% en 1995 a 15.3% en
junio de 1999, en tanto la de los países de la Cuenca del Caribe sólo aumentó de 15.7% a 17.9% (y sólo
Honduras y El Salvador alcanzaron logros destacables). Así pues, México ha desplazado a China como
principal proveedor de prendas de vestir al mercado de los Estados Unidos. La nueva IED en materia de
textiles y prendas de vestir ha generado un proceso incipiente de formación de cadenas productivas,
tendencia que resulta particularmente notoria en productos como pantalones de algodón para hombre y
mujer, especialmente pantalones vaqueros (rubro en que México tiene una participación superior al 50% en
el mercado de importación de los Estados Unidos). Con las nuevas oportunidades que ofrece la producción
de "paquete completo" se abre la perspectiva de que México se convierta en el "Hong Kong de América
Latina".
Los países de la Cuenca del Caribe han estado atrapados en un modo de entrega al mercado de
los Estados Unidos -basado en los bajos salarios, la producción compartida y las zonas de libre
comercio- que no ofrece ninguna de las nuevas ventajas de que goza México en el marco del TLC. Al
Sesión 2 -167

perder participación en el mercado de importación frente a México, dichos países reclamaron (sin éxito)
al Congreso de los Estados Unidos para obtener la paridad con el TLC. Desgraciadamente, las ventajas
que el TLC otorga a México no es más que uno de los nuevos desafíos que han de enfrentar los
ensambladores de prendas de vestir de los países de la Cuenca del Caribe. El último tramo del Acuerdo
sobre los Textiles y el Vestido, firmado en el seno del GATT en 1994, estará sujeto a una nueva
disciplina comercial (en lugar de las cuotas del Acuerdo Multifibras) a partir del año 2005. A dicho
tramo corresponde la mayor parte de las exportaciones de prendas de vestir de los países de la Cuenca
del Caribe. Debido a tal liberalización, los ensambladores de dichos países deberán competir (sin zonas
de procesamiento de exportaciones ni preferencias de mercado, y con salarios significativamente
superiores) con potencias asiáticas, como China. Mientras México posee una industria textil y del vestido
más integrada, en que están apareciendo opciones de paquete completo (aunque aún queda mucho por
hacer para aprovechar plenamente las normas de origen del TLC), los ensambladores de prendas de vestir
del Caribe muy probablemente enfrentarán una lucha sin cuartel en cuanto a los salarios y los incentivos
de IED en esa industria.

Sesión 2 -168

TAX INCENTIVES FOR FOREIGN INVESTMENT
IN LATIN AMERICA

by
Richard M. Bird and Duanjie Chen
International Tax Program
Joseph L. Rotman School of Management
University of Toronto

A Paper Prepared for the Inter-American Development Bank. December 1999.

Sesión 2-169

Table of Contente
Abstract

171

Executive Summary

171

I.

Introduction

172

II.

Taxes and Tax Incentives in Latin America

175

A. Latín America and Asia Compared
....
175
B. Business Tax Provisions in Selected Countries ..............................................................176
III.

An Effective Tax Rate Analysis

178

A. Calculating Effective Tax Burdens
178
B. Recent trends in FDI in Latín America ..........................................................................182
C. Tax Incentives in Central America and the Caribbean.................................................l83
IV.

Tax Competition: A Few Remarks

V.

Conclusión

185
187

References
Appendices

194

Sesión 2 - 170

Abstract
This paper summarizes the debate on tax incentives over the last decade, analyzes the impact
of taxes on foreign direct investment in Latin America in comparison with other emerging
economies, and discusses the effects of incentives on competitiveness. The paper concludes
that if there are to be any special tax incentives for investment at all, they should be few in
number, simple in structure, and universally available.

Executive Summary
The lengthy debate over the effícacy of tax incentives is far from resolved. Tax incentives are
clearly only; one, and often a minor, factor determining the magnitude of foreign direct investment (FDI)
in any country. More important factors are political stability, a favorable legal framework (trade policy,
labor policy, etc.), and of course economic conditions including natural resources, infrastructure, domestic
market, and cost and quality of labor. When the non-tax environment is favorable, tax incentives may add
to a country's attractiveness. When the environment is unfriendly, they may perhaps offer investors some
needed compensation. In short, although tax incentives do not necessarily or always foster investment, a
"good" tax system may indeed in some instances help attract capital investment.
Relative to Asia, Latin America as a rule has more complicated business tax systems. In
particular, Latin America imposes more non-profit related taxes on business such as asset and payroll
taxes. In addition, unlike the Asian countries in their initial growth period, most larger Latin American
countries currently make little direct use of tax incentives to compete for foreign direct investment. Most
tax incentives appear to be focused more on sectoral development (mining, agriculture) or to exporting or
financial investment.
Detailed analysis of the tax regimes in five larger countries shows that, excluding the tax on
dividends, Chile appears to have the most transparent tax system, with lower taxes and incentives more
focused on new capital investment. Interestingly, Perú also appears to be relatively tax competitive,
owing largely; to its relatively; generous tax depreciation allowance for machinery and equipment. In
contrast, among the countries studied, Argentina and Brazil impose relatively higher taxes, with Brazil
offsetting them with more conditional tax incentives that in principie seem likely to impair economic
efficiency. Finally, México lies in between these two groups in terms of both its general tax level and the
simplicity of its tax system
Recent FDI growth in Latin America is less impressive than in OECD countries owing largely to a
significant decline in such major economies as México and Argentina. Although corporate tax rates in
Latin America - as distinct from all taxes on business - are relatively low, Latin America appears
nonetheless to be the only región in which some important countries (Argentina, Brazil, México) have
recently raised corporate tax rates. These observations of course do not suggest any causal relationship
between tax and FDI developments but they do suggest that further study might be wárranted.
In any event, while academics may argüe about the merits of tax incentives, in the real world
some degree of fiscal competition appears to be an inevitable fact of life from the perspective of most
governments, national and sub-national alike. The question of what tax structure may improve a country's
competitiveness and still be consistent with non-predatory international economic behavior is therefore
always of interest. Broadly, twofiscal approaches to attracting FDI are possible. Thefirst is to apply

Sesión 2 - 171

standardized tax provisions to all investors with relatively low tax rates. The second is to tax different
investment activities differently in order to achieve specific goals of economic policy such as promotion
of exports or development of particular industries or regions. The latter approach is usually implemented
through morefinely-tuned incentives which, experience suggests, even if successful in the short run in
attracting some investment are all too often likely to impinge on allocative efficiency (and to open the
door to significant revenue erosion through evasión), thus providing a less sustainable basis for economic
growth in the long run. On the whole, it thus seems likely; that thefirst - broad-based, low-rate approach
with simple and generally available investment incentives (if any) - seems more suitable for the relatively
more developed countries of Latin America.

I.

Introduction

Many surveys have, over the years, reviewed with varying degrees of detail andrigorthe effects of
tax incentives on investment in developing countries.1 On the whole, most such reviews have not been very
favorable to tax incentives,finding them to be either redundant and ineffective — forgoing revenue and
complicating thefiscal system without adding to capital formation - or distorting and inefficient - directing
investment into less than optimal channels. Indeed in some countries, incentives have been so badly designed
and administered that they have, somewhat miraculously, managed to fall into both of these pitfalls at once.2
Even in some of the fast-growing Asian countries which have made considerable use of tax incentives in the
past, some recent analysts have condemned the tax incentive approach and urged the adoption of a more
neutral system3 Nonetheless, with the exception of Hong Kong, almost all the high-growth Asian countries
have made extensive and prolonged use of specific tax (and other) incentives in the past and indeed continué
to do so.4 Of course, many other factors were also important in explaining the Asian success stories such as
macroeconomic stability, high and growing investment in human capital, and the market-driven (export)
nature of growth (Summers and Pritchett, 1993). Nonetheless, high investment levels clearly played a critical
part (Young, 1993). Are other countries mistaken in hoping for the same linkage?5
The history of investment incentives in the postwar period may be viewed from two quite
different perspectives. One perspective sees this history as a more or less steady trend away from the
extensive reliance on incentives that characterized the initial development efforts in many countries in the
1950s and early 1960s - not least in Latin America — to the nirvana of "incentiveless" tax systems now
commonly recommended by international experts (World Bank, 1991) and, to some extent, actually put
into place in a few countries, notably Indonesia (Gillis, 1985).6 Similar, if less drastic, moves away from
fiscal incentives may be seen in a number of Latin American countries in recent years, including
Argentina, Bolivia, México, and Colombia (Bird, 1992).

2
3
4
5

6

Portions of this introductory section follow closely t e m r detailed discussion in Bird and Chen (1998). Early reviews of
h
oe
i v s m n incentives still worth C n utn include Heller a d K uí a (1963), Lent (1967), Shah a d Toye (1978), a d Gandhi
n et e t
o s li g
n a fm n
n
n
(1987). The m s recent detailed review is S a (1995).
ot
hh
For a example, see the study of Belize by Bird a d Chen (1997).
n
n
See, for instance, Ishi (1993) on J p n a d Choi (1997) on Korea.
aa n
See, for example, Ishi (1993) on J p n Trela and Whalley (1995) o Korea, Chang a d Riew (1994) on Taiwan, a d Asher (1988)
aa,
n
n
n
on Singapore.
Even in Indonesia,frequently cited as a instance in which tax incentives have been abandoned (Gillis, 1985), a recent
n
appraisal concluded t a "generally, t e promotion of investment through the use of incentives has been effective"
ht
h
(Sumantoro, 1993, p. 266).
As Goode (1993) noted, t e attitude to tax incentives in t e 1 5 s a d 1 6 s was markedly different from t a which n w prevails
h
h 90 n 90
ht
o
. In t e earlier period,reportsa d advisersfrequently favored tax incentives to channel investment in one desired direction or
h
n
another. Later, as skepticism grew both a to t e efficacy of incentives a d t e wisdom of interventionist policies in general,
s
h
n h
altitudes changed, a d tax incentives for i v s m n fell into disfavor.
n
n et e t
Sesión 2 - 1 7 2

In this view, the palpable failure of central planning in Eastern Europe and elsewhere, the
accumulated empirical evidence of the relative inefficiency and ineffectiveness of incentives, and the
growth of economic knowledge all point in the same direction — namely, away from the sort of
interventionist policy represented by incentives and towards the ideal of a uniform "level playingfield" in
which market forces, rather than government officials, make investment decisions. The move away from
incentives is thus seen as part of the general move towards market- rather than state-dominated allocative
decisions.
The second perspective on recent changes in attitudes and actions on incentives is, however, quite
different. Which ideas influence policy often depends as much on chance and circumstance as on their
inherent merit. Over time both ideas and practices may change - awayfrom interventionism and towards the
market, for example - as may the country models viewed as exemplary — witness the demise of the Soviet
central planning dream a decade ago. As the recent disillusion with the so-called "East Asian Miracle" (World
Bank, 1993) demonstrates, to the extent that the ideas that domínate at any point in time reflect less
increments in knowledge than changes in fashion, they may well be reversed a few years down the road. The
Asian currency crisis has, for example, clearly reduced the appeal in Asia and perhaps elsewhere in the world
of the incentive approach to growth. Nonetheless,fiscal incentives are not only still to be found in many
countries but recently seem to have become more attractive to some countries.7
The apparently sharp disparity between expert opinion and policy practice leads to several questions.
Are the East Asian countries mistaken in thinking that tax incentives (and other policies fostering investment)
had something to do with their growth success? Does the "Hong Kong model" of low and uniform taxes
offer a surer path to success than what may perhaps be called the JKS (Japan-Korea-Singapore) model of
fiscal (and other) interventionism? Would the East Asian "tigers" have done as well — or better - without the
incentives? And, most important in the present context, does the apparent failure in the past of superficially
similar incentive packages in many Latin American countries8 reflect differences in political or social
"culture,"9 the macroeconomic context,10 or simply differences in the design and implementation of the
incentives?11 To put the point at its simplest, would nothing have worked in Argentina in the 1980s— and
anything have worked in Korea in the same period?
Recent developments in economic theory have seemed to some to east a slightly more favorable light
on incentives. So-called "new trade theory", for instance, with its emphasis on imperfect markets and extemal
economies has to some extent re-opened the case for selective government intervention in trade and factor
markets. The relevance in the real world of these ideas remains doubtful, however, owing both to the unclear
7

8

9
10

11

For example, S a (1995) lists tax holidays in a least 1 developing countries, accelerated depreciation s h m s in a least 28,
hh
t
4
ce e
t
a d o h r i v s m n incentives (ciedits,reinvestmentallowances, etc.) in at least 21. Earlier surveys similarly f u d widespread
n t e n et e t
on
use of a variety of tax incentives in all paits of t e world. Chia a d Whalley (1995) suggest t a not only is t e variety of tax
h
n
ht
h
incentives currently f u d in developing countries greater t a t a in developed countries but t a such incentives are now m r
on
hn h t
ht
oe
c m o in developing countries t a t e were a decade ago.
o mn
hn hy
See, for example, t e detailed studies of Argentina, Ecuador, and t e Dominican Republic summarized in González Cano
h
h
(1994). Particularly notable for detail andrigorare studies of Argentina and México by Sánchez Ugarte, who summarizes his
finding in Sánchez Ugarte (1987).
Ishi (1993, p.156), for example, suggests t a "socio-cultural" factors s e m r likely t a tax incentives to explain high savings
ht
e m oe
hn
rates in J p n
aa.
See, f r example, Weeks (1997), as well as Thirsk (1989) o t e i p r a c of this factor in evaluating incentives. I v s m n
o
n h m ot n e
n et e t
incentives are obviously less likely to be effective in countries with high variábility in inflation and growth rates, just a export
s
incentives are unlikely to be veiy effective in countries with substantial exchange rate instábility.
As a recent s u y of incentives in Belize noted: "Apartfrom t e question of its ef e s on economic incentive, there are two major
td
h
fe t
problems with t e present s se of fiscal incentives in Belize. First, it is not a d c n o be administered well for t r e related
h
yt m
n ant
he
reasons: (i) it is far too complex; (ii) it is completely non-transparent a d non-standardized; and (iii) it is, in t e end, completely
n
h
discretionary. Second, a d perhaps m r easily remediable, it is replete with s a u o y loopholes t a m k it both a revenue "sink"
n
oe
tt tr
ht a e
- into which revenuesflow but nothing necessarily c m s out - a d a potential revenue "bomb" - with t e potential of devastating
o e
n
h
t e n r a revenue s se in t e long run" (Biid a d Chen, 1997)..
h om l
yt m
h
n
Sesión 2 - 173

importance in practice of external economies and the concern that govemments may not be capable of
implementing even beneficial interventions without opening the doors to dubious special interests and
excessive rent-seeking.12 The exacting practical demands of strategic trade theory make it an uncertain basis
on which to base an argument for incentives, although the enthusiasm for such arguments of special interest
groups that stand to gainfrom tax favors seems unlikely to be restrained by such considerations.
Much the same may be said of the new literatura on "endogenous growth." Some scholars have
demonstrated in thisframework how incentives favoring research and development or machinery and
investment may, under certain circumstances, accelerate growth.13 The fact that a few empirical studies
(notably DeLong and Summers, 1991, 1992) lend some support to such arguments makes the apparent case
for interventionist policy stronger, even in the face of the many political and organizalional constraints on
implementing such policies in developing countries (Levy, 1993).
The partial shift in the ruling paradigm, with incentives to some extent moving back into favor, may
be illustrated by contrasting two recent studies done under World Bank auspices. World Bank (1991) reflects
the prevailing conventional wisdom with respect to tax incentives: they erode the tax base, reduce investment
efficiency (because they mostly respond to pressure from special interests), are ineffective and often
inequitable, and facilitate rent-seeking activities and tax evasión.14 In contrast, Shah (1995) concludes his
introduction to an impressive set of empirical studies of tax incentives recently carried out under World Bank
auspices as follows: "Some tax preferences such as expensing, (refundable) investment tax credits for R&D
and machinery and equipment that embody advanced technology have strong theoretical and empirical
support."
Expert attitudes to tax incentives may thus again be changing, whether influenced by changes in
theory (endogenous growth), in empirical results (for example, the production function studies reported in
Shah, 1995), in real world experience (the East Asian "miracle"), in intellectual fashion, or in all these factors.
Whatever the experts may say, however, incentives seem likely to be demanded, and supplied, in the future,
much as they have been in the past. The design, implementation, and evaluation of tax incentives thus remains
an important question.
As the continuing debate illustrates, it is by no means easy to assess the effects of tax incentives on
investment in general, let alone on direct foreign investment. Many factors enter into the investment decisión
- economic conditions (rates of return, exchange rate risks, infrastructure, stability) and non-tax legal
questions (labor policy, trade policy, investment restrictions) - in addition to tax issues. Many aspects of tax
systems - tax structure, international tax rules, treaties, administration - in addition to incentives may affect
investment decisions. Many forms of incentives (and disincentives) may play a role: tariffs, export taxes and
subsidies, transfer pricing policy, price controls and subsidies, limits on repatriation, tax exemptions,
deductions and credits, domestic content requirements, provisión of infrastructure, and even direct cash or inkind grants (for example, of land). Only a limited selection of these issues can be considered in the present
paper. Specifically, our aim is simply to review briefly some recent experience with tax incentives in Latin
America with particular attention to their potential effects on foreign direct investment.

12

13

14

See, for example, t e interesting papers by K u m n (1993), Krueger (1993) a d Slemrod (1995). As K u m n (1993, p.364) p t
h
rg a
n
rg a
us
it: "Free trade [read u io m factor taxation] is a pretty good if n t perfect policy, while a effort to deviale fo it in a
nf r
o
n
rm
sophisticated way will probably end up doing m r h r n t a good."
oe a r h n
See, for example, Barro and Sala-e-Martin (1992). Although there is tío necessary implicañon in these models t a increased
ht
growth results in increased welfare, policy-makers looking for a rationale for doing something t e want to do have never wonied
hy
about such r fn m n s in t e past, a d t e s e unlikely to do so in t e future.
ei e e t
h
n hy em
h
Despite these harsh words, the "reference set" of taxes proposed in World Bank (1991) recognized t a t e e m y still be tax
ht hr
a
preferences in s m cases, although it argued a y incentives should be both limited in duration a d distributed m r evenly across
o e
n
n
oe
sectors a d assets in order to avoid distorting i v s m n decisions.
n
n et e t
Sesión 2 - 1 7 4

Sectíon II of the paper presents a brief review of the major features of the tax and incentive systems in
Latin America compared to those found in Asian countries, followed by a more detailed examination of the
taxatíon of business infive specific countries - Argentina, Brazil, Chile, México, and Peni. Sectíon III then
analyzes and discusses the impact of taxes in thesefive countries on foreign direct investment, utilizing the
METR (marginal effective tax rate) approach. This analysis is coupled with a brief review of recent FDI
flows and some comments on experience with tax incentives in the smaller Central American and Caribbean
countries. Finally, Sectíon IV discusses briefly the "competitiveness" dimensión of incentives, and Sectíon V
concludes the paper with a brief discussion of an appropriate approach to tax incentives for foreign
investment.

II.

Taxes and Tax Incentives in Latin America

In this section, wefirst provide an overview of the major features of the tax and tax incentive
systems in Latin America in comparison with those found in Asian countries.15 We then outline some
specific business tax provisions in Argentina, Brazil, Chile, México and Peni.

A.

Latin America and Asia Compared

Compared with Asia, Latin American taxes and incentives appear to have the following main
features, as illustrated in Table 1.
Corporate income tax rates in Latin America countries are relatively low. The regional
average CIT rate is about 31 percent, which is similar to that in the South East Asia and not
high compared to other regions over the world.
-

It seems common in Latin America to impose non-profit-related taxes other than the property
tax (e.g., tumover tax, and gross- or net-assets-based taxes). As we shall see, when these taxes
are not collected as a minimum tax, they may increase significantly the effective tax burden
on capital investment. Such taxes are uncommon in Asia.
Reflecting past experience, many countries in Latin America take explicit account of the
impact of inflation on taxation by imposing some monetary rule or tax indexation. Such
provisions are rare elsewhere in the world.
Employee profit sharing appears to be popular in some Latin American countries. When this
plan is mandatory and even non-deductible for the income tax purpose, it may significantly
reduce the net-of-tax profit and henee discourage foreign direct investment. Again, such
schemes are unknown in Asia.
Moreover, payroll taxes (including social security contributions payable by the employer) are
rather high conpared with most Asian countries. Although such taxes may not appear to be a
significant burden to foreign investors in view of the relatively low labour cost, they may
nonetheless — particularly when imposed on total payroll without a ceiling — be perceived as
an obstacle to hiring more highly-skilled workers and henee discourage potential foreign
investment that might bring in advanced manufacturing and service techniques.

15

For further discussion of taxes a d tax incentives in Asian countries, see Bird and Chen (1998).
n
Sesión 2 - 1 7 5

Finally, unlike the Asian countries in their early stage of growth, most Latin American
countries have not focused on using tax incentives to compete for foreign direct investment
except that related to exporting andfinancial investment. Most tax incentives found in the
región seem intended primarily to encourage sectoral (e.g., mining, farming, transportation,
etc.) and regional development. That is, they appear to be more instruments of what might
perhaps be called 'industrial policy" than of growth-oriented development policy. The
principal tax incentives that encourage export and appear to be attractive to foreign investors
are exemptions for transaction taxes levied on export-related inputs, which are generally
applicable to free zones and assembly operations. In addition, tax incentives that promote
financial investment often include an income tax reduction or exemption for the investment
income.

B.

Business Tax Provisions in Selected Countries

Below, we highlight the main differences in tax regimes among the five countries selected for
more intensive study. A detailed discussion of the profit tax and incentive regimes in these countries is
provided in Appendix 1, where Table Al provides a summary of the key tax provisions for each of the
five countries and Table A2 outlines their major tax incentives relevant to manufacturing and service
sectors. (Despite the importance of mining and other natural resource sectors with respect to direct
foreign investment in many countries, such investments are often subject to special tax rules that do not
lend themselves to summary here.)
•

•

Tax Depreciation: All five countries adopt the straight-line depreciation method. Again,
Chile and Peni appear to be more generous than the other three countries in allowing
machinery and equipment to be depreciated at about 20% annually, which is double the rate
allowed in other three countries. Perú also grants an accelerated depreciation allowance for
buildings used for lodging in order to promote its tourism industry.

•

Loss carry-overs: Chile allows operating losses to befirst carried back to offset the retained
earnings of previous years18, and then carried forward indefinitely. Other countries allow only
carry forward of tax losses for limited years, and Brazil imposes some restriction on the future
taxable profit that can be used to write off losses.

•

16

Corporate income tax rotes: The tax levied on the corporate income appears to be lowest in
Chile, where the CIT rate is only 15%. However, distributed profit is subject to either the
domestic personal income tax (up to above 33%) or a withholding tax to non-residents (20%
as indicated below). In contrast, the CIT rate is 35% in Argentina and México and 37% in
Brazil16 without differentiating the retained earnings from the distributed. It is noticeable that
all these three higher taxed countries raised their income based tax rate in 199917. Perú has a
CIT of 30%, which, as in Chile, has been stable since the early 1990s.

Inventory Costs: Most countries adopt thefirst-in-first-out method and other inventory
accounting method. Only Argentina allows the last-in-first-out method.

The rate in Chile is a combination of CIT and a social security levy with a taxable basis similar to that for CIT. Neither of
these two taxes is deductible for the purpose of t e other.
h
In the case of México, a portion of t e CIT is deferrable. That is, t e reinvested earnings will be taxed at a lower rate of 3 %
h
h
0
(32% in 1999), and the resultant difference in tax payable will be deferred until the profits are distributed.
In this case, CIT paid in relation to the retained earnings is treated as a advance payment to be set off against future income
n
tax liability or refiinded under general rules.
Sesión 2 - 176

•

Withholding tax on dividends: It has been common in thesefive countries to exempt the aftertax distribution from further taxes, except that Chile imposes a withholding tax (20%) on
dividends received by non-residents. However, effective in January 1999, dividends paid out
of the after-tax profits in México will be subject to a withholding tax of 5%.

•

Property Taxes: Allfive countries levy property taxes based on appraised valué and with tax
rates varying by location.

•

Property Transfer Tax. All countries except Chile collect tax from transfers of immovable
property, with the rate rangingfrom 1.5% in Argentina to 3.3% in México.

•

Social Security Taxes: Argentina collects the highest social security contributions from the
employer based on payrolls (33%) while Chile has the lowest (up to 3.4%). However, there
are also compulsory profit sharing plans in Brazil, Chile, México and Perú, which in effect
impose an additional payroll cost on the employer. In the case of Chile, profit sharing (by
option) may be equivalent to 25% of the total payroll up to 4.75 minimum monthly salaries.

•

Transaction Taxes: All five countries impose the value-added tax at a general rate of around
20%, with Brazil having a complicated dual-VAT system. The main transaction tax that
affects business activity in these countries is the import duty on capital goods and raw
materials. As mentioned below, many tax incentives granted in these countries include import
duty exemption.

•

Tax Incentives: There are no tax incentives specifically designed for encouraging foreign
direct investment in these countries. This report will therefore focus on those tax incentives
targeted on export promotion that exist in allfive countries and a priori appear to be attractive
to foreign investors.
As shown in Table A2, Argentina does not provide any tax incentives to promote specific
sectors (except mining), or regions, or investment projects.
In contrast, Brazil adopts numerous tax incentives to promote export production and
regional development. These incentives vary greatly depending on the location, sector
and the product involved. They are also very specific in terms of the tax to be exempted
or reduced and, in the latter case, the magnitude of reduction.
Chile, besides a generous tax depreciation allowance for all imported or new machinery
and equipment as mentioned above, enacted an investment tax credit in 1999 to encourage
investment in its extreme southem regions. The tax credit ranges from 10% to 40%
depending mainly on the total valué and type of investment.
México refunds import taxes paid by any taxpayers on inputs included in their exports19.
Peni introduced a generous tax incentive package to promote economic development to
its Amazon región. The package includes exemptions for income tax, VAT, excise taxes
and other taxes for a duration of 50 years (Byrne, 1999). Other Peruvian tax incentives for
regional development are divided into two categories - for the Amazon and border

19

México also provides extensive tax incentives for t e in-bound assembly operations (Maquiladoras). These incentives were
h
supposed to be phased out in 1 9 and are not further discussed here: for a outline, see Mexican Investment Board (1999).
99
n
Sesión 2 - 177

regions. Industrial enterprises in the border región are exempted from all taxes except
social security contribution, import duties and some municipal taxes.
In summary, excluding the tax on dividends, Chile appears to provide the most transparent tax system
with lower taxes and rather focused tax incentives aimed at new capital investment. Perú also appears to
be "tax-competitive," providing the most generous tax depreciation allowance for machinery and
equipment In contrast, Argentina and Brazil appear to be on the higher end of business taxation among
these fxve countries, with Brazil providing more conditional tax incentives that in principie impair
economic efficiency. México appears to be between these two groups in terms of both general tax level
and the simplicity of its tax system. In addition to specialfiscal advantages for exporting firms (mainly
import duty exemption for inputs) in all but Argentina, like many other countries in Latin America Bolivia, Uruguay, Paraguay, and Ecuador, for example - several of the countries studied in detail here
(Brazil, Pera, and, to a lesser extent, Chile) provides special tax incentives to businesses operating in socalled "free zones."

III.

An Effective Tax Rate Analysis

In this section, we undertake a quantitative assessment of the impact of the tax regimes in the five
selected countries on foreign direct investment. The methodology used for this assessment is METR or
marginal effective tax rate analysis (see Appendix 4 for a brief technical exposition). Assuming that a US
multinational is the representative foreign investor, we ask the following simple question: How do the tax
systems of the five Latin American countries impact on investment from the perspective of a US firm
considering investing in the región? In this analysis, we consider only the impact of taxes on foreign
capital investment in the manufacturing and service sectors. Despite their importance in terms of foreign
investment in some countries other important sectors, such as resource-based industries and financial
services, are excluded in part because these sectors are often subject to special administrative and tax
regimes.
As mentioned earlier, taxes and tax incentives are of course only one factor affecting FDI inflows.
Other major factors include political stability, transparency of policy and administration, market
efficiency, soundness of infrastructure, quality of labor forcé, and domestic demand. A country which
scores higher in terms of these non-tax factors may attract foreign investors even though it has higher
taxes and fewer tax incentives. (An obvious example is of course the US.) Countries that do not perform
as well in terms of these basic conditions may try to use favorable tax treatment in a sense as
compensation for the potential investors. Of course, even if one confines oneself tofiscal measures there
may be altemative measures that may be preferable- for example, lower and simple taxation combined
with fewer preferential tax- incentives. In addition to the already-cited case of Hong Kong, Ireland and
Hungary to some extent illustrate the latter approach. We shall return briefly to this point in Section V
below.

A. Calculatíng Effective Tax Burdens
To assess the impact of various taxes on investment decisions, we aggregate the impact of the
relevant taxes on the rate of return on capital eamed by the typical investment project. The typical foreign
investment project - which, as mentioned above, is American in this model - is considered to be one in
the manufacturing or service industries. The taxes considered in terms of their direct impact on
investment include corporate income, withholding, gross receipts and other capital-reiated taxes such as
Sesión 2 - 178

import duty, property tax, and property transfer tax. Assets-based taxes are excluded when they are
imposed as a minimum tax (as in Argentina, México and Perú) which can be used to offset income tax
liability and henee has no impact on tax-paying companies. Mandatory profit sharing plan have also been
excluded owing to missing information. If such plans were taken into account, the estímate of effective
tax rate for all but Argentina would be higher than those analyzed below.
The effective tax rate on capital is measured as the amount of tax paid as a proportion of the
gross-of-tax rate of retum on capital, taking into account all tax rates, bases, and incentives in a particular
location. For example, if the gross-of-tax rate of retum on capital is 10 per cent and the effective tax rate
is 50 per cent, then the net-of-tax rate of retum on capital is 5 per cent (50 per cent of 10 per cent). We
estímate the effective tax rate for marginal investment projeets. That is, the net-of-tax rate of retum must
be sufficiently high for the investor to be willing to invest in these Latin American countries compared to
some other American investments in another country or another asset. Thus, in the example above, the
marginal investment would need to earn a net-of-tax rate of retum on capital at least equal to 5 per cent if
project were to go ahead20.
Three additional points about these estimates should perhaps be clarified. First, since our focus is
on real rather thanfinancial investment, only taxes affecting real capital investment decisions are relevant.
Increasingly, some Latin American countries - such as Brazil -have negotiated tax treaties with various
foreign countries. Arguably, the development of a network of such treaties may facilítate foreign
investment both by signalling a country's interest and, more specifically, by removing barriers to profit
repatriation. In some instances (for example, Cañada), a treaty is necessary if foreign investors are to
receive the benefit to tax relief for foreign investment (credits or exemptions) from their countries of
residence. In other instances, such treaties may actually make tax incentives more effective to the extent
the residence country allows what is called "tax sparing" (OECD, 1998). Tax treaties may thus be
important determinants of repatriation decisions in some cases. Nonetheless, they do not need to be taken
explicitly into account in the calculations reported here since they essentially affect onlyfinancial flows.
A second point relates to the less than perfect tax administration that unfortunately still prevails in
some of the countries analyzed here. The METR analysis does not take administrative deficiencies into
account because its intent is to evalúate the impact at the margin of the structural design features of tax
systems. In effect, it estimates the effect of the tax system, if properly administered, on the present valué
of the accumulated cash flow that would be generated by an additional dollar of investment. It does not
assess the efficiency with which tax laws are applied in practice.
Finally, the analysis considers onlyfirms that are profitable and henee should in principie be taxpaying. In principie, it could be extended to loss firms by taking into account such features as loss
canyovers and the interaction of (presumptive) asset taxes and income taxes. This calculation, however,
requires information on the average length of period it takes for a tax-lossfirm to write off its losses and
become able to pay tax, which is not available to our study.
The analysis that we have carried out here begins with a comparison of the effective tax rate on
foreign capital investment in the five countries, assuming no tax incentives for investment (see Table 2,
"regular taxable case" under "manufacturing" and "services"). As Table 2 shows, there are two effective
tax rates for Chile and the service sector of Perú. In the Chilean case, the lower rate corresponds to the
general corporate income tax rate without taking into account the personal income tax or withholding tax
on payable by the recipients which contributes to the higher effective tax rate. In the Peruvian service
sector, the lower rate reflects the fast write-off allowed for lodging buildings used by the tourism industry
20

The mnm m rate of retum needed for i v s m n s is measured using current interest rates and inflation rates for t e various
íi u
n et e t
h
countries. The real rate of retum on capital is obtained by subtracting inflation f o t e nominal interest rates.
rm h
Sesión 2 - 179

while the higher rate is for other service industries. In terms of undistributed profit, Chile appears to be
the lowest taxing country from a foreign investor's perspective. This results mainlyfrom the combination
of the low statutory corporate income tax rate (15%) and the generous tax depreciation allowance for
buildings. However, Chile's tax advantage deteriorates when taxes on profit distribution are taken into
account. In this case, Perú provides the most favorable tax system to foreign investors — particularly in its
tourism industry as a result of having the second lowest statutory corporate income tax rate and the most
generous tax depreciation allowance for machinery and equipment.
In contrast, Brazil appears to provide the least favorable tax system for foreign direct investment
in both manufacturing and service industry. This results from its rather high corporate income tax rate
(37%), less generous tax depreciation allowance, and a rather high turnover tax (over 2.65 per cent).
Argentina shares many of these features with Brazil. It has the second highest corporate rate (35%), a
rather high turnover tax (average 2.55), and the least generous tax depreciation allowance (2% for
buildings). The exemption of capital goodsfrom import duties in Argentina, however, produces a lower
effective tax rate for manufacturing in Argentina than in Brazil. The effective tax rate in México is at the
middle among the five countries, reflecting the combination of its "médium" income tax rate and the
highest property transfer tax (3.3%).
The service sector is taxed more heavily than manufacturing in allfive countries. This is mainly
because buildings and land account for a higher share of capital invested in service sector compared to that
invested in manufacturing sector (79% vs. 26%). Furthermore, property taxes on buildings and land,
which do not apply to other type of assets (machinery and inventory), contribute to the higher effective tax
rate on buildings and land compared to that on machinery and inventory. The higher effective tax rate on
capital in service sector is a combination of these two factors.
For illustrative purposes, we have, as shown in Table 2, simulated the impact of the corporate
income tax alone. .This simulation is done by excluding from our model all the non-profit-based taxes
such as the gross-receipts tax in Argentina and Brazil, the assets tax in Brazil and Chile, the propertyrelated taxes, and import duty on capital goods in all but Argentina. As this simulation shows, the
effective tax rate reflecting the corporate income tax only was, unsurprisingly, lowest in Chile. On the
other hand, the rate attributable to non-profit-based taxes was lowest in México.
Although most attention in such studies is customarily paid to the corporate income tax alone, we
suggest that more attention should be paid to non-profit based business taxes for at least three reasons.
First, in many countries - including such developed countries as Cañada (Mintz Committee, 1998) such
taxes are becoming increasingly important. Second, non-profit taxes behave quite differently over cycles
than profits-based taxes: this may be beneficial to governments from a revenue perspective, but the
reduced cyclical sensitivity of taxes implies that the present valué of future taxes impinging on investment
decisions increases to the extent such taxes are not profits based. Third, METR analysis such as that
carried out herefrequently shows that the most important taxes impinging on business investment
decisions are taxes on asset valúes (property taxes or gross receipts taxes) that may have low nominal rates
but that may constitute major charges on potential profits.
As is well known, inflation may have a significant impact on taxes paid. For example, inflation
increases effective tax rates on capital when assets are depreciated at original cost rather than at the cost
of replacing them, unless assets are re-valued periodically. On the other hand, inflation can also reduce
taxes paid if firms are able to deduct interest expense, unadjusted for inflation, from income. The interest
deduction shelters incomefrom taxation and can more or less offset the impact of inflation on the valué
of fixed assets. This is particularly trae in the case of land and buildings for which the cost of financing
is the dominant factor in determining the cost of capital. Furthermore, for a high-inflationary economy,
writing off inventory on its historical valué can cause inflated tax burden to investors. In the case of the

Sesión 2 - 1 8 0

countries studied here, however, all except Argentina make some tax adjustment for inflation. In the case
of Argentina, the inflation rate is very low (0.5%) and henee has little impact on the effective tax rate.
As a result, the effective tax rates estimated for these countries in Table 2 can be seen as net of countrydififerentiated inflation impact.
In a second case reported in Table 2 (the "tax incentives for exporting" case), we have simulated
the impact of existing tax incentives for export promotion. As described in Section 2 (and Appendix 1),
these tax incentives are similar across countries except Argentina and consist mainly of exemptions from
transaction taxes on export-related inputs. These taxes generally include import duty and value-added
taxes. Since value-added taxes are designed to be refundable as input-tax credits, only import duties
levied on capital and material inputs would affect the effective tax rate endured by investors using
imported inputs. Therefore, compared to the effective tax rates for the "regular taxable case" as analyzed
above, the effective tax rate for the "tax incentives for exporting case" reflects mainly the impact of import
duty exemption for capital goods and raw materials. In Brazil, the incentive for export promotion also
includes an exemption for a turnover-based social contribution of 0.65%. As the table shows, except for
Argentina where no such incentives exist, the effective tax rate for manufacturing industry in all countries
shows a reduction of 5 - 6 percentage points.
The third case reported in Table 2 simulates some tax incentives provided in free zones. "Free
zones" cover a range of possibilities from bonded manufacturing warehouses to offshore financial centres
and export processing zones. In their various guises, this approach has become very popular in Latin
America in recent years and , in some countries (such as Honduras) such zones account for almost all FDI.
Non-tax incentives, such as "one-stop shopping", easier regulatory procedures, and subsidized
infrastructure may be included in some cases, but as a general rule the distinguishing feature of such
zones is their exemption not onlyfrom transaction taxes but alsofrom various direct taxes through devices
such as tax holidays and reduced tax rates. These additional tax incentives include a 40% income tax
reduction in Brazil, a full income tax exemption in Chile, and a complete tax exemption for 15 years in
Pera. (In México, tax incentives in free zones are generally the same as those available to the exporters
outside thefree zones.) As a result, the effective tax rate infree zones may fall by 7, 3 and 16 percentage
points, respectively, in Brazil, Chile and Pera compared to the general tax incentives for export promotion.
With respect to manufacturing investment, following the earlier model of México' s maquiladoras,
the principal form of FDI attracted initially to such zones is almost invariably labor-intensive
manufacturing - for example, with respect to textiles, garments, toys, and sporting goods. Subsequently,
more consumer durable investment (appliances, automobile parts) may be attracted, and in the best of all
cases, as in some Asian countries, even the lower range of "high-tech" investment sectors such as
electronics and precisión engineering sectors. A critical factor in such development, however, appears to
be the availability of high levels of local skills as, for example, in Ireland, Taiwan, and Korea - all
countries which have long invested heavily in educating their populations in the skills and knowledge
needed to compete successfully in the modern world. In the absence of such investment in human capital,
even if the initial impact of FDI tends to raise employment and real wages, the inherently "foot-loose"
lower end of the assembly and "low-tech" industriesfirst attracted to such favored zones are likely to
move on if an even lower-wage tax haven looms somewhere on the world horizon México, for example,
like Indonesia and the Philippines, appears to have gained some industry for this reason as countries such
as Korea and Taiwan began to price themselves out of the low-wage market. Countries relying on tax-free
zones to attract FDI thus to some extent face a dilemma since to develop a sustainable economic base over
time they generally need to spend some of the veryfiscal revenues they are forgoing. There is no easy
answer to this problem, except to emphasize the need for countries utilizing this approach to attracting
FDI to pay considerable attention to the need to forge linkages between "zone" activities and the
development of their economies, and especially the knowledge and skills of their populations, more
generally. In the end, unless the increased economic activity that may be initially attracted by tax

Sesión 2 - 1 8 1

incentives results in increasedfiscal revenues that are then directed to such growth-inducing activities , it
seems unlikely that any initial favorable effect will be long sustained.

B.

Recent trends in FDI in Latin America

Taxes are often considered an important consideration in foreign direct investment and in crossborder portfolio investment decisions but the evidence connecting the two is seldom clear. Foreign direct
investment (FDI) reached new levels in 1998 (UNCTAD, 1999). A recent OECD study (Miyake and
Thomsen, 1999) summarizes FDI trends over the past three years.21 This study shows that, for many of
the OECD member countries, "recent inflows are not simply higher than before, they represent a different
order of magnitude. For ten OECD Members22 in 1998, inflows were roughly twice the level of 1997."
This change was especially notable in the United States and United Kingdom, where the dramatic
increment in FDI inflows was driven by cross-country mergers and acquisitions.
In Southeast Asia, where mergers and acquisitions were lower in the past two years due to the
regionalfinancial crisis, foreign investors were attracted by the growing possibilities offered through
corporate restructuring and by the more liberal environment for such acquisitions. In particular,
traditional intra-Asian investment has been replaced by large purchases of Asian firms by non-Asian
foreign countries. Foreign investment inflow through mergers and acquisitions in 1998 was more than six
times that in 1997 for Japan and 4.5 times for Korea.
In contrast to this phenomenon, Latin America as a whole had a small drop in the valué of crossborder mergers and acquisitions. For example, Mexico's ranking in OECD inflows declined in 1998 as it
participated less than other countries in the wave of cross-border mergers. Tax policies that do not
specifically encourage fast-paced capital replacement may be one obstacle to modernizing industries in the
región. Although there was a significant increase in FDI in Brazil, "the region's biggest target for foreign
buyers,"23 much of this may be related to resource industries.
Table 3 shows FDI inflow as a percentage of GDP in most Latin American countries in
comparison with selected OECD member countries, Asian countries, and some transitional economies in
the Central and East Europe. Of course, the FDI to GDP ratio does not necessarily indícate the magnitude
of a country's success in attracting FDI. For example, while the average FDI to GDP ratio for Latin
America as a whole was 4.8 in 1998, which is more than double that for the United States, the total FDI
inflow for all Latin American countries accounted for only a third of that for the United States. What we
are interested in is the trend in this ratio, which in a sense provides a useful indication of how countries
are performing in the global competition for capital. Three observations can be drawnfrom Table 3.
•

21

23

First, Latin America as a whole caught up with trends in terms of its rapid increment in FDI. Despite
the growing pace of FDI in Latin America (17%), however, its performance was not as impressive as
that in G7 countries (50%) and some other OECD countries such as Sweden and Czech Republic
(90%) owing to a significant decline in some major economies including Argentina, Chile, México
and Venezuela.

It should perhaps be noted that FDI numbers are ot n rather s a y and open to altemative interpretations. It is not
fe
hk
u c m o , for example, for different sources to give quite different numbers, often varying by very large amounts: for
no mn
Poland in 1 9 , for example, estimated FDIflows rangedfrom $ to 1 billion (Transition (World Bank), October 1999, p. 7)
98
5
1
These ten countries are Cañada, Czech Republic, Denmark, Finland, Germany, Korea, Netherlands, Sweden, United Kingdom
and United States.
According to Miyake and Thomsen (1999), Brazil accounted for around 7 percent of total acquisitions in Latin America in
0
1998, compared with just 30 percent in 1 9 .
97
Sesión 2 - 1 8 2

•

Second, the significant increment in FDI inflow in Latin America occurred mainly, apartfrom Brazil,
in some of the less developed economies such as Bolivia, El Salvador, and Guatemala. This recent
performance may reflect a very low starting point in terms of FDI and a rather low level of GDP24. It
may, nevertheless, offer an encouraging sign for future growth in these countries to the extent FDI
inflow is accompanied by the introduction of more advanced technological, operational and
management standards.

•

Finally, among the Latin America countries, the FDI to GDP ratio varíes significantly from a high of
12.9% in Panama to a low of 0.8% in Uruguay. The reasons for this wide variation deserve further
study.

These observations on the FDI trend in Latin American countries compared with other regions do
not suggest any very direct relationship between taxes and FDI. On the other hand, a recent tax rate
survey conducted by KPMG (1999), does seem to suggest that tax reduction may to some extent at least
go hand by hand with the trend in FDI growth. The major relevant findings of the KPMG survey are the
following:
Lower tax rates are emerging in both the Latin American and Asia-Pacific regions
At the same time the trend toward lower corporate tax rates in developed countries continúes
And tax rates everywhere are under pressurefrom increasing business mobility
On the other hand, as we emphasized earlier in our discussion of Table 2, corporate tax rates
only part of thefinancial equation considered by investors - not least in Latin America where
non-profit-based taxes account for over 50% of the effective tax burden in some countries.

C. Tax Incentives in Central America and the Caribbean
No country is an island in today's fiscal world, and an important rationalefrequently offered in
support of tax incentives is that they provide a needed tool in the competition to attract international
investment. Given the growing importance of regional trading arrangements in the región,25 it may be of
interest to summarize briefly some of the results of a recent comparison of fiscal incentives in Belize to
those in CARICOM countries. In addition, since all the countries included in this comparison are
relatively small and make extensive use of "targeted" tax incentives, this section to some extent may offset
the "large country" (and relatively non-incentive intensive) bias of the analysis reported earlier.
As Table 4 illustrates, Belize (which is classified as a less developed country — LDC — in
CARICOM) essentially has in place a somewhat more generous and less transparent versión of the basic
CARICOM LDC incentive package. For example, the main criteria for tax holidays under the general
CARICOM harmonization scheme are export potential and local value-added. Although the same criteria
arefrequently mentioned in Belize, in practice more emphasis appears to be put on job creation and the
total valué of the investment, and in any case the whole process is highly discretionary.26 One of the few
advantages of tax holidays is that they are in principie neutral as between capital and labor, unlike more
directly investment-based incentives such as investment allowances. Relating the length of a tax holiday
to the size of the investment on the one hand and the number of jobs created on the other makes little
sense from this perspective. Questions may be raised about criteria such as local value-added, but in
24
25
26

For example, El Salvador incurred a negative FDI inflow, or withdraws of FDI, in 1 9 and received little FDI in 1 9 . In
96
97
t e meantime, although its GDP grew at a annual rate of 1 % since 1 9 , t e GDP per capita was below $2,000 in 1 9 .
h
n
1
93 h
98
For a interesting recent review of thefiscal systems in t e Mercosur countries, for example, see Macón (1999).
n
h
In contrast, Thirsk (1991), p. 705, notes t a in J m i a t e ratio of domestic value-added to sales a d t e ratio of exports t total
ht
a ac h
n h
o
sales are two of t e t r e m s i p r a t factors used to d t r i e t e length of t e holiday period granted, a d t a t e e is
h he
o t m ot n
e e mn h
h
n ht hr
relatively little discretion exercised with respect to s c decisions.
uh
Sesión 2 - 1 8 3

general they seem both more relevant and easier to monitor than a criterion like job creation in relation to
capital investment for which there can be no standard across industries and regions.27
In other respects too, the current incentives in Belize seem less well-structured than the general
CARICOM package. For instance, the máximum duration of a tax-holiday in other CARICOM countries
is 15 years. In Belize, the period can be as long as 25 years in law and even longer in practice. More
importantly, the duration of import duty exemption in the CARICOM scheme is the same as that of tax
holiday. In Belize, however, these two incentives can be applied separately, which obviously both
complicates the process of application and approval and increases administration and compliance costs.
Another problem arises from the fact that although the duration for carrying forward tax losses after the
tax holiday in other CARICOM countries is limited to five years, it can be indefinite in Belize.28 Finally,
in contrast to the explicit provisión with respect to depreciation allowances in the CARICOM package,
there are no specific investment-related incentives in Belize.29 On the whole, Table 4 certainly suggests
that the existing Belizean system of incentives seems to be at least as generous, if not more so, than that
in most of its neighboring countries. Recent detailed studies of fiscal incentives in Guatemala and Jamaica
generally confirm this impression.
In Guatemala, for example, until recently a wide and varied set of tax incentives was available for
many investors, including investment allowances (deduction of more than 100% of cost of assets), 8-10
year tax holidays for up to 90% of profits, duty exemptions, special exemptions for free trade zones, and a
varied set of benefits for exporting firms. A tax reform in 1993, however, eliminated most of these
incentives — except those for exports - although it "grandfathered" (left in forcé) benefits already granted.
On the other hand, the new tax system introduced a new deduction for reinvested profits of up to 25% of
taxable income.30
Similarly, Jamaica has had a regime of tax holidays and other incentives since at least 1944,
although its current system largely reflects that permitted by the CARICOM agreement of 1974 as set out
in Table 4 above. Apart from a proliferation of minor incentives, the most important change since then
was the adoption of a more transparent and objective system of determining eligibility and the duration
and amount of any benefits granted with the establishment of the Jamaican National Investment
Promotion (JNEP) agency in 1981. Under the previous system - as is still trae in Belize — in effect
decisions were often made at ministerial discretion. Although JNIP recommendations are still subject to
approval by a cabinet committee chaired by the Prime Minister, current practice is reportedly to limit tax
holidays to a máximum of six years. A careful review of the workings of this program concluded that
although it may not have been very effective in generating significant new investment per dollar of
revenue forgone in the past, on the whole the incentives worked fairly well. In particular, it was concluded
that "...Jamaica seems to have strack a reasonable balance between rules and discretion in its incentive
program."31

For example, local value-added is easily defined as ex-factory sales mn s t e purchase of factor servicesfrom non-resident
iu h
sources.
28 As stated in Chapter 46 of Laws of Belize, I c m Tax, Part I, aiticle 15, losses m y be carried forward "as long as t e trade,
no e
a
h
business, profession or vocation is carried on by t e person who incurred t e loss". In c n r s , t e shorter carry-forward period
h
h
o ta t h
m r generally prevalent (in countries which have generally shorter holiday periods in a y case) is presumably intended simply to
oe
n
provide for "hardship" cases in which it m y t k a year or two of full operation before potentially taxable profits are realized. But
a ae
note alsofrom Table 4 t a Belize is not t e only CARICOM c u t y to permit indefinite loss carry-forward.
ht
h
o nr
Of course, t e indefinite loss carry-forward provisión will eventually allow the full write-off of all capital costs incurred during t e
h
h
holiday period, but this is not the s m as a incentive geared specifically to capital outlays s c a a initial or i v s m n
a e
n
uh s n
n et e t
allowance.
3 0 See Bahl, Martínez-Vázquez a d Wallace (1996), pp. 107-08.
n
31 Thirsk (1991), p. 722.
27

Sesión 2 - 184

IV.

Tax Competition: A Few Remarks

As global corporate tax rates decline, the debate on the tax competition heats up. This debate
raises several questions. First, does tax competition help some countries in terms of attracting capital
investment and achieving economic growth? Second, is tax competition harmful to the global economy as
a whole? Andfinally, what would an efficient and competitive tax structure look like?
A recent literature survey of inter-state tax competition in the US suggests that taxes have a small
but statistically significant effect on interregional location behaviour (Wasylenko, 1997). More
specifically, the effect of taxes in a specific jurisdiction depends not only on the tax elásticity of capital
allocation, but also on the extent to which one jurisdiction's overall tax levels are significantly different
from the average of the jurisdictions it competes against. A large deviation from the average tax level,
multiplied by the tax elásticity, will yield a large location, employment, or investment effect. This finding
seems relevant in Latin America given the range of apparent variation in effective tax rates. That is, there
may indeed be some range within which tax influences may affect a country's attractiveness to foreign
investment and henee its growth rate.
In 1998, OECD published a report by its Committee on Fiscal Affairs (CFA) on "harmful tax
competition". This report has been critically received by business. For example, the Business and Industry
Advisory Committee (BIAC) to the OECD released its response to the CFA report in June 1999 arguing
that international tax competition is generally a healthy phenomenon from the viewpoints of both
govemment and business. Its states: "Tax competition tends to keep tax burdens lower, which creates
pressure for less wasteful, and, therefore, more efficient use of public fiinds. In addition, it fosters
increased efficiency in the allocation of scarce resources. Lower tax burdens also translates into lower
cost for multinationals operating within the territory and internationally." These conflicting views
illustrate the very different ways in which the issue of tax competition, whether within or between states,
may be viewed.32
Often, such competition has been attacked as wasteful and distorting, leading to such undesirable
outeomes as "tax jungles" which impose high compliance costs on society and a "race to the bottom" as
competing jurisdictions continually lower tax rates in an effort to retain tax base. Lessfrequently, fiscal
competition between jurisdictions has been said to provide both a useful check on the propensity of
governments to expand and a stimulus to more efficient use of scarcefiscal resources.
Despite
the
fervor with which proponents of both "harmful" and "beneficial" intergovernmental competition often state
their positions, in reality neither the theoretical ñor the empirical literature on this issue is conducive to such
certainty. While there is much still to be learned about intergovernmental competition between governments
at the same level (horizontal competition) as well as between governments at different levels (vertical
competition), what we know to date does not lend clear support to either extreme position in this debate.
Neither the position that all intergovernmental competition is bad ñor the position that all intergovernmental
competition is good dominates. How one assesses these arguments depends upon a variety of factors that
need to be specified carefully with respect to each particular setting in which the question is considered.
Whatever philosophers may argüe, however, in the real world some degree of fiscal competition
seems an inevitable fact of life. The question is then what tax structure would both improve a country's
competitiveness and do so in a manner consistent with non-predatory international economic behavior?
As noted earlier, there are broadly two fiscal approaches with respect to attracting foreign direct
investment. One is to apply standardized tax provisions to all investors combined with low tax rates. The
other is to tax various investment activities differentíy to achieve economic policy goals such as sectoral
development or export promotion. The latter approach is typically implemented through fine-tuned
32

For a elassie example of these two views, see t e debate in Buchanan a d Musgrave (1999).
h
n
Sesión 2 - 1 8 5

incentives which often cause severe tax distortion among different investor groups and impinge on
economic efficiency. Although the second approach may attract foreign investment in the short run,
particularly at an early stage of opening-up an economy to the world, on the whole it seems likely that the
first approach is more conducive to sustainable FDI inflow in the long run and henee more suitable for the
relatively more developed economies of Latin America.
If all taxes were levied on a strictly "benefit" basis, so that the taxes imposed on residents paid for the
benefits they receivedfrom public services and any taxes "exported" to non-residents were similarly offset by
cost-reducing benefits to such non-residents, both horizontal and vertical spillovers would be minimized.33 In
the real world, however, such perfection is not readily attained, and both forms of spillovers are often found.
When there are horizontal spillovers between jurisdictions at the same level (whether between or within
countries), the result may be that taxes will be unduly high, since non-residents (= non-voters) in effect end up
paying for services enjoyed by residents (=voters) to the extent "excess" business taxes are exported.
Alternatively, the result may be too low a level of taxation for fear of loss of tax base to other jurisdictions.
When govemments can inpose taxes that are, in effect, borne by other govemments or nonresidents,
the economic cost of taxation is lower than it should be, and the result is likely to be excessive govemment
spending. On the other hand, if spillovers result in tax base moving to other jurisdictions, tax competition
may make the perceived economic cost of taxation too high, thus resulting in "too little" govemment
expenditure — although, as already mentioned, some have suggested that such competition may provide a
useful constraint on the taxing power of govemments. Those who argüe this way have generally been
concemed about the possibility of excessive spending as a result of intergovernmental spillovers.
Experience supports the conclusión that intergovernmental tax competition exists and can in some
instances constitute a real problem. These arguments are even more persuasive at the subnational than at the
national level. Numerous empirical studies in the United States, for example, provide conflicting evidence on
the extent to which differential state-local tax regimes affect competitive behavior as well as on the
effectiveness of specific subnational tax incentives. With respect to the latter, as one recent balanced survey
plausibly concluded, "studies exist to buttress almost any case about tax incentives" (Wasylenko, 1996). This
confusing literature suggests few firm conclusions, but one that emerges clearly is thatfiscal competitiveness
is more likely to affect location choices within smaller areas than within larger areas. This suggests that fiscal
differentials are likely to be more important factors in affecting location decisions within than between
nations.
Tax levels and tax incentives are part of the "fiscal climate" which prospective investors have to take
into account. At the subnational, as at the national, level a well-designed tax system - one that levies a
relatively uniform, stable, and "reasonable" burden on business - seems more likely to encourage investment
than an erratic and biased set of specific incentives. Nonetheless, the pressure on both national govemments
in the face of global economic pressures and also on the increasingly independent subnational govemments in
many countries to engage in incentive competition is likely to increase over time.

33Some

of the following is based on Bird a d Mintz (1999). Detailed references to t e relevant literature m y be f u d in this source.
n
h
a
on
Sesión 2 - 1 8 6

V.

Conclusión

The key to productive governmental competition, whether within or between nations, lies in
making the relevant decision-makers accountable for their decisions. In tum, the key to effective
accountability is to set out the rules clearly and to make relevant comparative information publicly
available. At base, the ultímate mechanism driving "good" competition is, on one hand, the ability of
citizens to compare governments in terms of the services they provide and the tax-prices they charge and,
on the other, their ability to affect and alter the decisions of government. Democracy without good
information is not enough to ensure that intergovemmental competition will be beneficial rather than
harmful. Ñor, of course, is information without democracy. Much of the recent decentralization in Latin
America, for example, has been closely associated with the revival of democratic institutions throughout
most of the región (IDB, 1997). But less attention has so far been paid to the need for good information
systems in order to ensure that this democracy is exercised in a responsive and responsible way (Bird,
1999). Indeed, even in countries which do not have well-established democratic institutions good
information is essential to ensuring enhanced service outcomes. Transparency is thus important in
ensuring that the ultímate outcome of anyfiscal incentives is socially beneficial
With respect to sub-nationalfiscal incentives, for example, a sound approach is to lay down some
basic "rules of the game" and to enforce them by public accountability of the sort sketched above. Much
the same seems true with respect to tax competition between nations, and one way of interpreting the
OECD's recent initiative with respect to "harmful" competition is precisely along these lines.34 In any
case, one important rule with respect to tax incentives is simply: Keep it Simple. Complex provisions
attempting to "fine tune" incentives to achieve detailed policy goals are likely to prove costly to
administer and unlikely to produce the desired results at a reasonable cost. Few countries, and even fewer
sub-national jurisdictions, have sufficient information to design narrowly-focused incentives in any case.
If there are to be any tax incentives, they should therefore probably be few in number and as simple as
possible in structure - not least to reduce the problems of enforcing tax systems effectively that are bound
to be exacerbated in countries with less than perfect tax administration by the presence of, in effect "onshore tax havens" in the form of "holiday" companies.
In this connection, it is of interest to note the evolution of investment incentives over the last
decade or so in eastem and central Europe, for example. Initially, almost all the former communist
countries introduced extensive tax holidays and other investment incentives. More recently, however,
countries such as Estonia and Hungary, which have been among the most successful in attracting foreign
investment have eliminated tax holidays and instead imposed relatively low and broad-based tax rates (for
example, an 18% CTT in Hungary). Other countries in the región are similarly considering introducing
more transparent and automatic forms of incentives such as accelerated depreciation and more generous
loss offset provisions.35
Whatever the incentive chosen, the second rule for ensuring transparency and accountability is
equally simple: Keep Records. Who gets what incentives? For how long? At what estimated cost in
revenue forgone? And with what results in terms of investment, employment or whatever else is of
interest? There will inevitably be problems in interpreting some of this information. Nonetheless, in the
absence of information incentives are even less likely than usual to play a useful role in terms of
promoting economic growth. Instead, they are likely to turn out to be at best a costly means of advertising

34
35

A similar interpretation of the concern of t e World Trade Organization (WTO) with certain aspects of members' tax policy
h
in connection with WTO's Trade Policy Review Mechanism m y be f u d in Daly (1999).
a
on
See the discussion in Coolidge (1999).
Sesión 2-187

for investment, or, more pejoratively, stealing it from one's neighbors, and at worst a way of rewarding
those considered politically worthy of receiving such favors.
Thirdly, and combining some aspects of the first two rules, any sound incentives should be
required to undergo periodic evaluations in as rigorous a way as possible. If they fail such tests they
should be terminated. If a jurisdiction is unable, or unwilling, to put in place such a "sunset" test of the
efñcacy of incentives, it should probably not be in the incentive business at all. And incentives that
cannot pass such tests should of course not exist.
Successful compliance with these apparently simple rules is clearly not easy, as experience around
the world suggests. The complex legislation, poor implementation, and complete lack of accountability
that presently characterizes most incentives in most national, regional, and local jurisdictions reflect
political and institutional factors that are unlikely to be easily, or soon, changed in many countries.
Nonetheless, in the absence of such changes, any new incentives that are put into place in response to the
cost-benefit calculus of local politicians will likely soon come to resemble instead the sad and familiar
world of most existingfiscal incentives rather than the Brave New World some proponents of incentives
seem to have glimpsed in some of the recent empirical and theoretical literature on endogeneous growth,
increasing retums, and all that.
Even if the present fiscal incentives in some countries were administered perfectly, they suffer
from a number of potentially fatal structural faults. Belize, for instance, is one of the relatively few
countries in the world that allows tax-holidayfirms carry tax losses incurred in the holiday period forward
after the holiday until they are used up.36 For example, a firm may incur an operating loss during its
holiday period and utilize this loss to offset otherwise taxable profits realized after the holiday has expired.
Even if firms realize profits during the holiday and henee benefit oncefrom the tax concession, they may
"double dip" in the sense that investment carried out during the holiday may generate depreciation
deductions that may be carried forward without limit and offset against subsequent profits. Other problems
arise from the lack of clarity with respect to many critical points - for example, when, precisely, does
production "start." In the absence of transparent and objective rules on matters which may significantly
affect thefiscal benefits available, it is hard to believe that the present complex, discretionary, and opaque
system of fiscal incentives found in countries such as Belize are optimalfrom any point of view.
Even if fiscal incentives in a country are well-designed and well-administered, experience
suggests that they are seldom effective and often costly in terms of undermining the creation and operation
of an effective, modern business tax system. As a rule, it often seems possible, in principie, to obtain
more or less the same degree of stimulus to investment whether foreign or domestic through simpler
measures that focus attention where it is most needed, namely, on strengthening, simplifying, and
stabilizing, the basic tax system. Three examples - none of which, it may be noted, would distinguish
between foreign and other investment37— may be suggested:
1. Cut the corporate tax rate. While obviously simple, and likely to be seen as strong sign that a
country is "open for business," this policy is likely to be both costly in revenue terms and to
provide less stimulus to investment than either of the next two altematives.

2. Eliminóte import dudes on capital goods and, perhaps, raw materíals (at least for expor
The introduction of the VAT has of course already moved most Latin American countries well
along this path, as has the generalized reduction of import tariffs. Nonetheless, it is far from clear
36
37

Note t a u d r t e general CARICOM incentive s h m (see above) tax holidayfirms are allowed to carry forward losses, but
ht n e h
ce e
only for up tofive years after t e expiration of t e holiday.
h
h
One argument against such a distinction is t e impossibility of telling what investors are really "foreign."
h
Sesión 2 - 188

that confining the elimination of all input import taxes tofirms operating in so-called "free zones"
as is currently the rule in some countries is either a wise or a necessary policy. In most of the
larger countries, the revenue effects of fixll elimination of duties on inputs are unlikely to be
serious, but of course actions along these lines are as much a matter of trade as of tax policy and
cannot be decided without a much closer examination of trade and development policy than is
possible here..

3. To the extent there is to be any special fiscal incentive for investment as such, it should li
take the form of a system offull or partial "expensing" of new capital investment, that is,
initial write-off of 100% (or some smaller fraction) of the cost of depreciable capital. Although
the rationale for this approach cannot be fully spelled out here, it is well-established in the
literature.38 In brief, such a system would (i) be relatively easy to administer, (ii) would remove all
(or much) of the tax deterrent to such investment, and (iii) would provide more stimulus to
investment per dollar of revenue forgone than either tax holidays or general reductions of
corporate tax rates. Nonetheless, in at least two respects this approach too may give rise to
questions, namely, its relative favoritism to capital-intensive investment and the greater burden it
imposes on the administration of the corporate income tax. Moreover, this incentive may not
benefit the new.firms which often incur operating losses and henee cannot utilize such tax
allowances during their early years. Countries that wish to entice newfirms to bring in capital
intensive business may therefore find a reasonable investment tax credit to be a better choice.
Despite these problems, if some form of tax incentive to investment is considered necessary, on the whole
some form of accelerated depreciation, preferably combined with improvement of the basic tax structure
and administration in the direction of an effectively lower-rate and broader-based system, would appear to
provide an adequate economic replacement for any benefits in terms of increased investment, growth, and
job creation that may now be being realized from the various specialfiscal incentives currently found in
most countries of Latín America.

38

For a useful early discussion, see Musgrave (1981).
Sesión 2 - 1 8 9

Table 1
Main Features of the Latin American Business Tax System
CIT rate
1999

(%)

Non-profitbased
taxes 3 9

Argentina

35

Yes

No

No

33

Yes

Yes

Bolivia*

25

NA

NA

NA

NA

NA

NA

Brazil

27-37

Yes

Yes

Yes

25.8

Yes

Yes

Chile

15-35

Yes

Yes

Yes

25.9

Yes

Yes

35

Yes

NA

NA

27.1

NA

NA

30-10

Yes

Yes

No

23.8

Yes

No

Dominican Rep.

25

NA

NA

NA

7.0

Yes

NA

Ecuador

25

NA

Yes

NA

NA

Yes

NA

El Salvador

25

Minor

No

No

14.5

Yes

No

Guatemala

25

Yes

NA

NA

12.0

NA

NA

Honduras

40.25

NA

NA

NA

NA

NA

NA

México

32-35

Yes

Yes

Yes

20.5

Yes

Yes

Panama

30

NA

No

NA

12.3

Yes

NA

Paraguay

30

NA

NA

NA

16.5

NA

NA

Perú

30

Yes

Yes

Yes

Up to 26.2

Yes

No

Uruguay

30

Yes

No

NA

18.6

NA

NA

15-34

Yes

Yes

NA

Up to 15.5

NA

NA

Colombia
Costa Rica

Venezuela

Inflation
profit
adjustment sharing

Note: NA indicares information unavailable.

39

Excluding the property tax.
Sesión 2 - 1 9 0

Payrofltax Incentives Incentives
for
for financial
Rate
Exporting investment
(%)

Table 2
Effective Tax Rate on the Foreign Investment (%)
Argentina

Brazil

Chile 4 0

México

Perú 41

Impact of CIT only

31.3
20.8

35.5
19.1

22.3/26.9
4.0/10.6

28.2
19.3

24.1
11.5

Effect of non-profit-based taxes

10.5

16.4

18.3/16.3

8.9

12.6

2. Tax incentive for exporting

NA

29.3

15.1/20.4

21.8

16.7

3. Free zones

NA

22.3

12.0/17.0

21.8

0.9

38.8

39.1

Impact of CIT only

22.7

17.1

28.0/31.8
2.6/9.4

29.2
17.6

24.5/29.2
6.9/13.7

Effect of non-profit-based taxes

16.1

22.0

15.4/22.4

11.6

17.6/15.5

2. Tax incentive for exporting

NA

NA

NA

NA

NA

3. Free zones

NA

NA

NA

NA

NA

A. Manufacturing
1. Regular taxable case

B. Services
1. Regular taxable case

40
41

The higher effective tax rate includes t e withholding tax applicable only to t e dividends received by non-residents.
h
h
The lower effective tax rate is for t e tourism sector where buildings used for lodging can be written off 1 % annually for t e
h
0
h
first 5 operating years.

Sesión 2 - 1 9 1

Table 3
Foreign Direct Investment as a percentage of GOP (%)
1996

1997

1998

1996

1997

1998

Argentina

2.4

2.8

1.9

Cañada

1.5

1.8

2.7

Bolivia

6.4
1.4

9.2
2.4

10.2
4.1

France
Germany

1.4
0.2

1.6
0.5

1.9
0.8

6.8

6.9

U.K.

2.2

2.8

4.5

NA

6.4
NA

Colombia

3.6

U.S.

1.2

1.3

2.3

Costa Rica

4.7

0.6

NA

Italy

0.3

0.3

0.2

Dominican Rep.

0.7

2.8

NA

Japan

0.0

0.1

0.1

Ecuador

2.3

2.9

NA

Group average

1.0

1.2

1.8

El Salvador

-0.05

0.0

7.4

Korea

0.5

0.6

1.7

Guatemala

0.5

0.5

3.5

Malaysia

5.1

5.2

5.2

Honduras

2.2

2.6

1.6

Philippine

1.8

1.5

2.6

México

2.8

3.2

2.5

Singapore

7.9

10.1

8.5

Panama

4.1
NA

14.7
NA

12.9
NA

Thailand

0.1

6.2

Group average

3.1

1.6
3.8

3.8

2.7

3.1

3.7

NA

0.6

0.8

Ireland
Sweden

3.8

Uruguay

5.3
0.7

2.2

4.5

8.6

Venezuela
Group average

3.1
2.9

5.8
4.1

4.0
4.8

Czech Rep
Hungary

2.5
4.4

2.5
4.5

4.7
NA

Poland
Group average

3.3

3.6

NA

3.3

3.8

6.7

Brazil
Chile

Paraguay
Perú

Note: The average of the group is estimated excluding the countries for which the data is unavailable. Source: Based o
n
International Monetary Fund, International Financial Statistics, October 1999.

Sesión 2 - 192

Table 4
Incentives in Belize Compared to Central American and Caribbean Countries
CIT

CCA

Other

Loss

WHT

Belize

35.0

no

Yes

Indef

0.0

Antigua

40.0

yes

IBC

6

Barbados

40.0

yes

Export

9

Costa Rica

30.0

no

Reinvest

3-5

Dominican
Republic

26.0

no

No

3

El Salvador

25.0

no

Export

0

Guatemala

34.0

no

Export

4

Guyana

35.0

yes

Export

Indef

Honduras

35.0

no

Export

3

Jamaica

33.3

yes

No

Indef

Panama

34.0

no

Export

St. Lucia

33.3

yes

6

45.0

yes

Indef

Trinidad
Tobago

and

0-15
18

15-22.5
5-15
15-20

NotesrCIT - standard statutory corporate income tax rate (1995)
CCA - s m f r of accelerated depreciation, initial allowance or investment allowance: Antigua
o e om
has a 2% initial allowance on buildings, Barbados a 4 initial allowance on industrial buildings, St.
%
Lucia a 20% initial allowance on industrial buildings and machinery, Guyana a 70% investment
allowance on equipment, and Jamaica a 1 0 write-off of "market equipment cost" over time.
2%
Other - various special incentives for export industries; for Costa Rica, 5 % of reinvested profits are
0
deductible f o income; IBC = international base company; s m minor incentives for specific
rm
o e
industries and locations are n t indicated.
o
Loss - Number of years loss-carryforward permitted; "indef' = indefinitely.
WHT -Withholding tax rate on dividends paid to treaty country; often different rates for royalties
and interest, a d generally higher rates for all f r s of payments to nonresidents in nontreaty countries.
n
om
Source: Boadway a d Shah (1995).
n

Sesión 2 - 193

References
Asher, Mukul (1988) "Fiscal System and Practices in Singapore," in Mukul G. Asher, ed., Fiscal Systems and
Practices in ASEAN: Trends, Impact, and Evaluation (Singapore: ASEAN Economic Research Unit, Institute
of Southeast Asian Studies).
Bahl, Roy, Jorge Martinez-Vazquez, and Sally Wallace (1996J. The Guatemalan Tax Reform (Boulder:
Westview Press).
Barro, Robert J. and Xavier Sala-i-Martin (1992) "Public Finance in Models of Economic Growth," Review of
Economic Studies, 59: 645-61
Bird, Richard M. (1992) "Tax Reform In Latin America: A Review of Some Recent Experience," Latin
American Research Review, 27: 7-34.
Bird, Richard M. (1999) "Fiscal Decentralization and Governmental Competition," in Gianluigi Galeotti,
Pierre Salmón, and Ronald Wintrobe, eds., Competition and Structure: Essays in Honor ofAlbert Bretón
(Cambridge University Press).
Bird, Richard M. and Duanjie Chen (1997) "Fiscal Incentives in Belize: An Evaluation," Report prepared
for the World Bank, Washington.
Bird, Richard M. and Duanjie Chen (1998) 'The Fiscal Framework for Business in Asia," in Wendy
Dobson, ed., Fiscal Frameworks and Financial Systems in East Asia (Institute for International Business,
University of Toronto).
Bird, Richard M. and Jack M. Mintz (1999) "Tax Assignment in Cañada: A Modest Proposal," Paper for
Institute of Intergovemmental Relations, Queen's University, Kingston.
Robin W. Boadway and Anwar Shah, "Perspectives on the Role of Investment Incentives in Developing
Countries," in Anwar Shah, ed., Fiscal Incentives for Investment and Innovation (New York: Published
for the World Bank, Oxford University Press, 1995), pp. 105-25.
Boadway, R., Bruce, N. and Mintz, J. (1984) "Taxation, Inflation, And the Effective Marginal Tax Rate In
Cañada," Canadian Journal of Economics, 27: 286-99.
Buchanan, James M. and Richard A. Musgrave (1999) Public Finance and Public Choice (Cambridge,
Mass.: The MIT Press).
(BIAC) Business and Industry Advisory Committee to the OECD, "A Business View on Tax
Competition" (1999) Tax Notes International, 19: 281-287.
Byrne, Peter D. (1999) "Financial Transactions Taxes and Other Revenue Grabs in Latin America", Tax
Notes International, 19: 301 - 306.
Chang, Ching-huei, and John Riew (1994) "Tax Policy and Business Investment - The Case of Taiwan's
Manufacturing Industry," in R. Musgrave, C. Chang, and J. Riew, eds., Taxation and Economic Development
among Pacific Asian Countries (Boulder Westview Press).

Sesión 2 - 194

Chia, Ngee Choon, and John Whalley (1995) "Patterns in Investment Tax Incentives among Developing
Countries," in Anwar Shah, ed., Fiscal Incentives for Investment in Developing Countries (New York: Oxford
University Press).
Choi, Kwang (1997) "Tax Policy and Tax Reforms in Korea" in Wayne Thirsk, ed., Tax Reform in
Developing Countries (Washington: World Bank).
Coolidge, Jacqueline (1999) 'The Art of Attracting Foreign Direct Investment,in Transition Economies,"
Transition, October, pp. 5-7.
Daly, Michael (1999) "Some Taxing Issues for the World Trade Qrganization: From the Standpoint of the
Trade Policy Review Mechanism," Paper prepared for Conference on the Future of Tax Reform, The Adam
Smith Institute, London, December 8-10.
De Long, J. Bradford and Lawrence H. Summers (1991) "Equipment Investment and Economic Growth,"
Quarterly Journal of Economics, 106: 445-502.
De Long, J. Bradford and Lawrence H. Summers (1992) "Equipment Investment and Growth: How Strong is
the Nexus?" Brookings Papers on Economic Activity (Washington: The Brookings Institution).
Ernst & Young, World Wide Corporate Tax Guide, 1999.
Gandhi, Ved P., et al. (1987) Supply-side Tax Policy (Washington: International Monetary Fund).
Gillis, Malcolm (1985) "Microeconomics and Macroeconomics of Tax Reform: Indonesia," Journal of
Development Economics, 19 : 221-54
González Cano, Hugo (1994) 'The Role of Tax Incentives in Promoting Savings and Investment in Latín
American Countries", Buenos Aires, July.
Goode, Richard (1993) "Tax Advice to Developing Countries: An Historical Survey," World Development,
21: 37-53
Inter-American Development Bank (1997) Latin America After a Decade of Reforms (Washington)
International Bureau of Fiscal Documentation, Taxation in Latin America, Sections B and C (for selected
countries by corresponding authors
Ishi, Hiromitsu (1993) The Japanese Tax System (rev. ed.; London: Oxford University Press)
KPMG Corporate Tax Rate Survey - Januaiy 1999, KPMG web site.
Krueger, Anne O. (1993) "Virtuous and Vicious Circles in Economic Development," American Economic
Review, Papers and Proceedings, 83: 351-55
Krugman, Paul (1993) "The Narrow and Broad Arguments for Free Trade," American Economic Review,
Papers and Proceedings, 83: 362-66
Larrain, León and Ignacio García (1999) "Chile Enacts Tax Incentives to Promote South's Economic
Development", Tax Notes International, 1 vol. 8, No. 25, January 18, 1999,

Sesión 2 - 195

Lent, George E. (1967) "Tax Incentives for Investment in Developing Countries," International Monetary
Fund Staff Papers, 14: 399-417
Levy, Brian (1993) "An Institutional Analysis of the Design and Sequence of Trade and Investment Policy
Reform," World Bank Economic Review, 7: 247-62
Macón, Jorge (1998) Armonizaciónfiscal en el cono sur ((Buenos Aires: Consejo Federal de Inversiones).
McLees, John A. and Cristian E. Rosso Alba (1999) "Argentina Enacts New Tax Reform Law", Tax Notes
International,. 18: 265 - 285.
Mexican Investment Board (1999) México's Maquiladora Export Industry, August.
Mintz, Jack M. (1995) "Tax Holidays and Investment," in Anwar Shah, ed, Fiscal Incentives for Investment
in Developing Countries (New York: Oxford University Press).
Mintz Committee (1998) Technical Committee on Business Taxation in Cañada (Ottawa).
Miyake, Maiko and Stephen Thomsen (1999) "Recent Trends in Foreign Direct Investment", Financial
Market Trends, No. 73, June.
OECD (1998) Curbing Harmful Tax Competition - Recommendation by the Committee on Fiscal Ajfairs
(París ).
OECD (1998a) Tax Sparing: A Reconsideration (París).
Sanchez-Ugarte, Femando (1987) "Rationality of Income Tax Incentives in Developing Countries: A SupplySide Look," in V.P. Gandhi (1987) Supply-side Tax Policy: Its Relevance to Developing Countries
(Washington: International Monetary Fund)
Shah, Anwar, ed. (1995) Fiscal Incentives for Investment in Developing Countries (New York: Oxford
University Press).
Shah, S.M.S. and John Toye (1978) "Fiscal Incentives for Firms in Some Developing Countries: Survey and
Critique," in John Toye, ed., Taxation and Economic Development (London: Frank Cass)
Slemrod, Joel (1995) "Free Trade and Protectionist Taxation," International Tax and Public Finance,
Silvani, Carlos and Paulo dos Santos (1995) "Administrative Aspects of Brazil's Consumption Tax
Reform", VATMonitor, vol. 7, No. 3, May/June 1995.
Sumantoro (1993) "Investment Incentives and Controls," in Euston Quah and William Neilson, eds., Law and
Economic Development: Cases and Materials from Southeast Asia (Singapore: Longman).
Summers, Lawrence and L.H.Pritchett (1993) 'The Structural Adjustment Debate," American Economic
Review Papers and Proceedings, 83: 383-89.
Thirsk, Wayne R. (1991) "Jamaican Tax Incentives," in Roy Bahl, ed, The Jamaican Tax Reform
(Cambridge, Mass.: Lincoln Institute of Land Policy)

Sesión 2 - 196

Trela, Irene and John Whalley (1995) "Taxes, Outward Orientation and Growth Performance in the Republic
of Korea," in Anwar Shah, ed., Fiscal Incentives for Investment in Developing Countries (New York: Oxford
University Press).
Wasylenko, Michael (1996) "The Role of Fiscal Incentives in Economic Development: How Ohio Stands
Relative to its Competitor States," in Roy Bahl, ed., Taxation and Economic Development: A Blueprintfor
Tax Reform in Ohio (Columbus: Batelle Press).
Wasylenko, Michael, "Taxation and Economic Development: The State of the Economic Literature", New
England Economic Review, March/April 1997, p.p. 37 - 52.
Weeks, John (1997) "Latin America and the 'High Performing Asian Economies': How Great were the
Differences and How do we Account for them?" Working Paper Series No. 70, Department of Economics,
School of Oriental and African Studies, University of London.
Weiss, Nelio and Joseph Wolf (1999) "New Brazilian Rules Increase Effective Corporate Tax Rate", Tax
Notes International, 18: 1884 - 1888.
World Bank (1991) Lessons ofTax Reform (Washington)
World Bank (1993) The East Asian Miracle (Washington).
Young, Alwyn (1993) "Lessonsfrom the East Asian NICs: A Contrarían View," Working Paper No. 4482,
National Bureau of Economic Research, Cambridge, Mass.
Zuniga, Omar (1999) "Mexico's new dividend withholding tax may viólate Mexico's treaty obligations",
Tax Notes International,. 19: 223 - 227.

Sesión 2 - 197

Appendix 1
Business Tax Provisions by Country: an Overview
This appendix this appendix provides an overview of business taxation in each of five Latin American
countries under our study: Argentina, Brazil, Chile, México, and Perú. The business taxation means taxes
that may affect business activities, particularly the real capital investment. The major business taxes
include capital taxes, payroll taxes and transaction taxes on business inputs. The capital taxes in our
context include the corporate income tax, personal income taxes on investment income42, and the property
tax on immovable properties and, in some countries, movable properties. A more comparative discussion
is provided in the main text of the report. The description presented in this appendix is based on the
publication of International Bureau of Fiscal Documentation, the 1999 Worldwide Corporate Tax Guide
published by Ernst & Young, and various issues of Tax Note International published in 1999.
Table Al summarizes the main features of each country's corporate tax system.

1. ARGENTINA
In response to its urgent need for additional revenue to balance its fiscal accounts, Argentina enacted a
wide-ranging tax reform at the end of 1998. The reform, as stated by the government, was designed to
improve vertical and horizontal equity in the tax system by means of a reduction of tax costs on labor and
an increase in the tax burden of business43. The reform package contains a two-percentage-point increase
in the corporate income tax rate, the elimination of certain exemptions, new thin-capitalization rules, new
transfer-pricing rules, and increased effective withholding tax rates. It also includes changes to the valueadded tax, a new minimum presumed income tax, a new tax on interest and otherfinancial charges, and
reductions in social security contributions

The capital tax provisions
The corporate income tax rate. The corporate income tax (CIT) rate is 35% in 1999. In early 1990s, the
CIT rate was 20 and 36 percent respectively applicable to the domestic and foreign corporations. This
discriminating dual rate system has been removed since 1992 and all corporations, regardless of the
nationality, have been taxed identically. However, the CIT rate has been raised gradually from 30% in
1992 to 35% today. This trend appears to be in a direction opposite to most other countries that encourage
the foreign direct investments by reducing the CIT rate.
There is also a minimum tax of 1% on assets held at the end of the tax period by the taxpayer. However,
such minimum tax paid in a given period may be credited against the income tax payable for the same tax
period. For a tax-loss company, or a company whose income tax payable is below the minimum tax paid
for a giver period, the unused minimum tax credit may be carried forward for up to 4 years.
The thin capitalization rule. The 1998 tax reform established a thin capitalization rule for legal entities
other thanfinancial institutions. The rule limits the interest deductibility to the higher of a debt to equity
ratio of 2.5 to 1 and 50% of taxable income (before the interest deducted). However, three types of

43

As illustrated by the effective tax rate analysis, taxes on any personal investment income could affect the cost of capital
investment through financing.
McLees and Rosso Alba (1999).
Sesión 2 - 198

interests are excluded from this limitation. They are interest paid to non-residents subject to a 35% final
withholding tax, interest on loans from resident individuáis, and 40% of any other interest.
The tax depreciation rule. The tax depreciation is based on the straight-line method. The depreciation rate
is 2% for buildings and 10% for machinery and equipment.
Inventory accounting method. The valuation of inventory is based on the market cost at the end of the
year, or last-in-first-out (LIFO).
Loss carrv-overs. The loss may be carried forward for up tofive years but may not be carried back.
The withholdine tax rate on dividends. There is no withholding tax on dividends distributed from the
after-tax profits. The 1998 tax reform introduced a new equalization tax on dividends. This tax is applied
to those dividends for which the commercial profits as the distributing source exceeds the taxed profits44,
and the tax rate is 35%, equivalent to the corporate income tax rate.
The turnover tax. A turnover tax is levied at the provincial level. The tax rate varíes by province and
rangesfrom 1% to 5%. To avoid the cascading effect, some provinces and the city of Buenos Aires have
enacted exemptionsfrom the provincial turnover taxes for activities including manufacturing and certain
intermediary services.
The property tax. The property tax is levied at the provincial level. Both taxable base and tax rate vary by
location but subject to the federal government regulation. For example, under the Federal Fiscal Pact, the
taxable base evaluated annually by local government is limited to 80% of the market valué and tax rates
are limited to 1.20% for rural property, 1.35% for suburban and 1.50% for urban.

The transaction taxes
There are mainly two transaction taxes that would affect the capital investment and business activities in
Argentina. One is the property transfer tax of 1.5% levied on the transfer price of immovable properties.
The other is the import duty on raw materials and parts ranged from 0 to 15%. New capital goods are
exemptedfrom the import duty.

The social security contributions
The social security contribution payable by an employer amounts to a total 33% of the employee's salary.
However, there is a ceiling for certain contributions. For example, the taxable base for contribution to the
retirement system and pensión fund is limited to the salary under ARS 4,800 for 1999. The contribution
rate may also be reduced depending on the location of the employment and the compliance record of the
employer.

Tax incentives
As a general rule, most of the tax incentive schemes, especially the industrial promotion regimes have
been suspended or abolished. In particular, there is no tax incentives designed for encouraging foreign
direct investment.

44

The "taxed profit" equals to t e taxable i c m minus the income tax paid and plus the dividends (and other non-taxable
h
no e
profits) received by t e distributing company.
h
Sesión 2 - 1 9 9

2. BRAZIL
Brazil has introduced, recently, several new laws to increase certain tax rates aimed at collecting more tax
revenues for coping with its budget crisis45. It should be noticed that Brazil uses a tax unit, know as
UFIR, for the adjustment of certain amounts (7??) 40 in its tax legislation. Currently, the UFIR is adjusted
every 6 months in accordance with changes in the consumer price index.

The capital tax provisions
The corporate income tax rate. The Brazil adopts a two-tier rate in corporate income taxation: a basic
corporate income tax and a corporate income surtax. Since 1996, the basic CIT rate has been 15%, and
the corporate income surtax of 10% is levied on the annual taxable profits exceeding BRL 240,000. In
other words, the marginal CIT rate on the annual taxable profits exceeding BRL 240,000 is 25%.
However, there is also a social contribution tax (SCT) levied on any CIT payers tofinance the federal
social security system. The tax rate for non-financial companies has been increased temporarily to 12%
for 1999 from the original 8%. The tax is levied on the pre-CIT profits and is not deductible for CIT
purposes. As a result, the aggregated tax rate on corporate income is 27% for the annual taxable profit
under BRL 240,000 and 37% otherwise.
The gross-receivts tax. A social assistance contribution (COFINS) is levied on the monthly tumover, and
the rate has been raisedfrom the original 2% to 3% in 199947. The new rules allow Brazilian companies to
credit the additional amount of COFINS (i.e., the newly increased 1 percentage point COFINS) paid
against the social contribution tax (SCT) on income. However, unused COFINS credit may not be cairied
forward to offset future SCT liabilities. An additional 0.65% of monthly turnover (PIS/PASEP) is also
payable by the employer for a savings programme for employees administered by the Caixa Econlmica
Federal. These levies, if paid and not used to offset the SCT as mentioned, are deductible for the income
tax purpose.
There is also a business tax based on the valué of assets although this tax is not significant (???) 48
The tax depreciation rule. The tax depreciation is based on the straight-line method. The depreciation rate
is 4% for buildings and 10% for most of the industrial machinery and equipment. Furniture and equipment
are also allowed for 10% depreciation annually. The higher depreciation is allowed for companies with
two or more operatíng shifts.
Inventorv accounting method. The valuation of inventory is based on the acquisition cost, or First-in-firstout (FIFO). The weighted-average-cost method is also acceptable to the tax administration but LIFO is
not. However, the acquisition cost or production cost of goods existing at the end of the tax period must
be adjusted to the market valué if this (???)"' is lower.

45
46

47

48

Weiss and Wolf (1999).
We need to know the content of t e "certain amounts". In relation to this question, we would appreciate a m r specific
h
oe
description of the tax unit in t r s of its role in t e tax adjustment for inflation. For example, whether the interest expense is
em
h
deducted with an adjustment for inflation and, if so, how.
The new rules allow Brazilian companies to credit t e additional a o n of COFINS (i.e., t e newly increased 1 percentage
h
mu t
h
point COFINS) paid against the social contribution tax (SCT) on income. However, unused COFINS credit m y not be
a
carried forward to offset future SCT liabilities.
We need to know the tax rate of this assets-based business tax.
This paragraph was adoptedfrom the International Bureau for Fiscal Documentation. It is not clear on whether the word
"this" here m a s the "acquisition cost" or t e "market valué".
en
h
Sesión 2 - 200

Loss carrv-overs. The loss may be cairied forward indefinitely. However, the loss to be written óff is
limited to 30% of the taxable income during a tax period.
The withholding tax rate on dividends. There is no withholding tax on dividends distributed from the
after-tax profits.
The property tax. A municipal real estáte tax is levied on urban land and buildings. The tax base is the
appraised valué of the property, which varíes by location, area, size, age, condition and use. Tax rates,
therefore, vary considerably across municipalities. (???) 50

The transaction taxes
The transaction taxes in Brazil include mainly a dual-VAT system, the import duty, and the taxes on
transfers of immovable properties.
The dual-VAT system consists of the state VAT (ICMS) and the Industrialized Products Tax (IPI) levied
by the federal govemment. The former is levied on all transactions while the latter only industrial goods
at the manufacturing and import stages51.
The import duty ranges from zero to 55%. Besides the full exemption given to the conventional export
processing zones, there are many conditional exemptions up to 80% as incentives for encouraging sectoral
or regional development.
The tax on transfers of immovable properties is imposed at the municipal level with different rate. In the
state of Rio de Janeiro, the rate is 2%.
There is also afinancial transaction tax of 0.38% on all banking transfers of currency.

The payroll taxes
There are several levies based on total payroll and payable by the employer. They includes the 12%
federal social security fund (INSS), 2.5% educational fund, 1.2% for the national industrial training
service and commercial training service fund (SENAI/SENAC), the 1.5% social service fund
(SESI/SESC), the 8% severance payment, and the 0.6% SEBRAE contribution for promoting small
enterprises. All the rates are based on the total payrolls and the aggregated payroll tax rate is 25.8%. As
mentioned above, there are also other social security levies based on taxable profits and total turnover.
The social security levy on the employee is progressive up to 9% and with a taxable ceiling of 516 BRL.
Employees also have the right to share the employer's profit. The profit shared by the employees is
deductible for the CIT purpose. (There is no information about specific rate and regulations.)

Tax incentives
Brazil has made an extensive use of incentives to promote development. The incentive scheme has been
aimed at promoting export, encouraging sectoral and regional development.
50
51

For t e purpose of effective tax rate calculation, we need a national average or a sample rate for a industrial región. In our
h
current effective tax rate calculation, we are using a artificial property tax rate of 1.5%.
n
According to the International Bureau of Fiscal Documentation, IPI is imposed on t e supply a d importation of any type of
h
n
m n f c u e products or goods. The taxable a o n is t e c s o s valué including duties for imported goods, or t e supply
a ua t r d
mu t
h ut m
h
valué including incidental charges and VAT for m n f c u e products. The rates vary inversely with t e necessity of the
a ua t r d
h
productfrom 356% (e.g., for manufactured tobacco and alcoholic beverages) to nil (e.g., for essential products). When the
production phase is concluded, this excise tax ceases to arise a d t e accumulated excise taxes paid are incorporated in the
n h
cost. For detailed discussion on the problems with the Brazilian VAT system, refer to Silvani a d Paul dos Santos (1995).
n
Sesión 2 - 201

Incentive schemes for promoting exports, besides the one applicable to the conventional export processing
zones52, are mainly exemption from certain transaction taxes and payroll taxes. Exporting companies
outside the EPZs are exempted from import duty on capital goods and other inputs for manufacturing
goods for exporting. Furthermore, their export proceeds may be excluded from the gross receipts subject
to the 0.65% tax for the savings programme for employees (PIS/PASEP). Finally, their financial
transactions in relation to the exportation are exempted from thefinancial transaction related taxes.
Incentives for encouraging regional and sectoral development are mainly provided through the companies'
investment in several investment funds and mostly valid until year 2010. More specifically, corporate
enterprises, regardless of the business location, may invest up to 40% of their income tax liability through
the government established investment funds. These funds are under the control and regulation of the
National Securities Commission. Through this arrangement,firms in effect convert 40% of their income
tax liability to their interest in those enterprises established in the regions entitled to use the investment
funds. Furthermore, industrial, farming and construction enterprises operating in the listed regions may
enjoy a 40% income tax reduction provided they invest this tax saving along with 50% of their own
resources in an approved project within the región. There are also other tax incentives granted to
qualifying enterprises established in the listed regions and engaged in the manufacturing and assembling
specific items such as qualifying vehicles. These incentives includes mainly tax exemptions under the
import duty, excise tax,financial transaction taxes, CIT and surtax (AIR).
Other tax incentives related to industrial development are mainly those targeted to the research and
development activities53. These incentives include mainly a tax credit equal to 15% on R&D-related
expenses (which is limited to 8% of the CIT liability), an exemption (with credit) from IPI for purchase of
capital goods including tools, and a double-up in depreciation allowance for new equipment and
instruments used for R&D activities. There are also certain tax credits or reductions in relation to income
generated from the R&D activities and the related transactions.

3. CHILE
The Chilean income tax law contains provisions designed to prevent the taxation of profits created by
inflation. These provisions, known as monetary correction, require taxpayers to revalue certain assets and
liabilities, such asfixed assets, inventory and debtfinancing, annually based on changes in the consumer
price index (CPI) and in foreign exchange rate54.

The capital tax provisions
The corporate income tax rate. The corporate income tax is 15% in Chile. There is no other taxes levied
on the corporate income. Furthermore, income derived from certain regions of Chile (e.g., the extreme
north and south) is exempt from corporate tax.
The foreign tax relief. The only treaty Chile has to avoid double income taxation is with Argentina.
Foreign income tax up to a rate of 15% paid to countries other than Argentina may be credited against
Chilean corporate income tax. Any additional foreign income tax is deductible as an expense.
s2

53
54

Companies operating in export processing zones (EPZs) are exempted from import-related taxes, excise tax, taxes on financial
transactions, and social security contributions. Furthermore, profits of these companies are taxed under either t e rules in
h
forcé when an operation agreement is concluded with t e government or a m r favorable regime established at a later date.
h
oe
These activities include carrying out t e Industrial Technological Development Programmes, developing integrated circuits
h
and investing in research and development of software production technology besides their core business.
For a detailed description of Chilean m n t r correction rule, refer to t e document on Chile published by t e International
o eay
h
h
Bureau of Fiscal Documentation.
Sesión 2 - 202

Other capital based taxes. There is annual business tax, or business Iicense fees levied by the
municipalities. The tax is assessed on the registered capital of 31 December excluding the part invested in
other firms subject to the same tax, taking into account the adjustments of inflation as mentioned above.
The tax rate ranges from 0.25% to 0.5%.
The tax depreciation rule. The tax depreciation is based on the straight-line method and the useful life
fixed by the tax authority. For example, the useful life for machinery in general is 20 years, and that for
factory buildings of a permanent nature is 40 years. A higher depreciation allowance may be authorized
in case of assets in use for two or more shifts. The taxpayer may use an accelerated depreciation system
by directly reducing the useful life of the new or imported assets to one third of that fixed by the tax
authority, provided that the normal useful life of the asset exceeds 5 years. As a result, the useful life for
new machinery in general may be shortened to 6.6 years for the tax depreciation allowance and that for
permanent factory buildings 13.3 years, and the corresponding depreciation allowance is 7.5% and 15%
respectively. The aforementioned monetary correction should be applied to the calculation of tax
depreciation allowance.
Inventory accountins method. For inventory valuation, thefirst-in-first-out (FIFO) method and the
weighted-average-cost method are accepted. However, the corresponding monetary correction must be
added to cost at the end of tax period.
Loss carry-overs. Losses mustfirst be carried back to offset undistributed profits of prior years and then
may be carried forward indefinitely.
The withholding tax rate on dividends. There is no withholding tax on dividends distributed from the
after-tax profits. While dividends received by resident individuáis are subject to personal income tax, the
taxpayer may claim a tax credit equivalent to the corporate income tax paid in relation to the dividends
received.
The property tax. The property tax is collected by the central government, from which the revenue is
allocated to the municipalities. The tax is levied on the cadastral valué of the immovable property, which
must be revised every 3 to 5 years and adjusted for inflation every 6 months. The tax rate is 1.425% and
1.5%, respectively, for urban and rural properties.

The transaction taxes
The transaction taxes include mainly the VAT at a single rate of 18% and the import duty of 11% in
general. There is no tax on transfers of immovable properties (although the related capital gains are
subject to the normal personal income taxation).

The payroll taxes
There are no payroll taxes other than social security charges. Furthermore, social security contributions
are basically payable by workers. Employers generally pay only 0.9% of the payroll to cover work
accidents. This charge may be higher (up to 3.4%) in relation to the degree of risk in working
environment. Furthermore, there is a mandatory profit sharing with employees in any enterprises. An
employer can opt for either of the two methods. One is sharing 30% of adjusted taxable profit with
employees. The other is paying 25% of annual salaries up to 4.75 times the minimum monthly wage at
the end of the taxable period, adjusted for inflation. The total social security contributions payable by
workers amount to about 20% of salaries with 10% for pensions, 7% for health services, and a minimum
2.55% for the disability and life insurance and the management fees for pensión funds.

Sesión 2 - 203

Tax incentives
Chile recently enacted tax incentives to encourage mainly small and medium-sized investments in its
extreme southern regions. The main tax benefit granted under the new law is a tax credit corresponding to
new equipment, machinery or related construction and installation. The tax credit is set as a progressive
percentage, from 10 to 40 percent, of the investment valué in fixed assets, depending on the amount
invested, the type of project and the goods that qualify for the investment. The targeted sectors include
tourism, transportation, and industrial and energy transmission projects. The qualified project invested
through December 31, 2008 is entitled to this tax credit. Any unused credit may be carried forward to
future years until the year 203055.
There is no tax incentives specifically aimed at encouraging the foreign investment56. There are several
incentives designed for export promotion, general investment, and regional development.

4. MEXICO
México adopts certain rules regarding inflation adjustment. The adjustment factor is the proportional
difference in the consumer price index between the starting month and the ending month of a given
period. The income tax law recognizes the effects of inflation on the following items and transactions:
depreciation of fixed assets, cost on sales of fixed assets, sales of capital stock (shares), monetary gains
and losses, and tax loss carried forward.
The capital tax provisions
The corporate income tax rate. The corporate income tax has been increased from 34% to 35% in 1999.
There appears to be a tax deferral of 5 percentage points until the dividends are effectively distributed to
shareholders57. The taxable income for a residential Corporation is its worldwide income from all sources,
while that for a non-residential corporation is its income derived from its Mexican source.
Minimum tax on net assets. There is a minimum tax of 1.8% on the net assets of corporations, which
provides a credit for the CIT payable. Any minimum tax paid in excess of income tax for any tax year
may be carried forward 10 years or back three years to offset CIT liabilities or CIT paid. More
specifically, in the case of carrying back the minimum tax credit, a refund of tax paid in the last 10 years (IBFD) up to that credit (and adjusted for inflation) is permitted.
The tax depreciation rule. The tax depreciation is based on the straight-line method. Depreciation is
computed on original cost of fixed assets, with the amount of depreciation indexed for inflation as
measured by price Índices. The máximum annual deprecation rates are set by law. The rate is 5% for
buildings, 10% for plant, machinery and office equipment, 30% for computers and related equipment,
100% for environmental machinery and equipment.
Inventorv accounting method. For inventory valuation, the basic requirement is the adjustment for
inflation, which is, in effect, equivalent to the average cost method.
Loss carry-overs. Business losses may be carried forward for 10 years.
ss
56
51

Larrain and García (1999).
A general restriction on foreign investment is t e mnm m a o n of USD 1 million although a y foreign investment
h i i u mu t
n
exceeding USD 5 million requires an official approval.
Refer to Zuniga (1999).
Sesión 2 - 204

The withholding tax rate on dividends. There was no withholding tax on dividends distributedfrom the
after-tax profits until 1999. Under the new tax laws effective January 1999, the dividends paid out of the
after-tax profits mustfirst be grossed up by the factor of 1.5385 and then subject to a withholding tax of
5%. As a result, the effective withholding tax rate is 7.7%. (The oíd regime, if the distributing
Corporation does not have sufficient accumulation in its "net tax profit" account to cover the dividend,
then the dividends are taxed at the corporate level at the CIT rate of 34%. In this case, dividends
distributed to foreigners subject to the lower of the treaty rate and the CIT rate. In its treaty with the US,
the withholding tax is 5% or 10% with the lower rate applicable to the receipt owning at least of 10% of
the payer. (E&Y)
The property tax. The property tax is levied at the municipal level. As a result, the tax rate varíes by
location. In the Federal District, the tax rate rangesfrom 0.131 % to 0.647%.
The transaction taxes
The VAT is levied at a general rate of 15% with a lower rate of 10% in border regions. There is also a real
estáte acquisition tax, levied at the local or state level, on the market valué of the transferred property. The
approximate rate is 3.3%.
The payroll taxes
The social security contribution (covering pensión, unemployment insurance, health insurance, etc.) is
levied on salaries up to a specified amount58. A housing fúnd is also payable by the employer at 5% on
salaries with a ceiling. Furthermore, the federal district and states levy a payroll tax on the total
remuneration for dependent personal services at a rate up to 2%. The resultant gross rate payable by an
employer is above 20% and that by the employee is 4%.
There is also a mandatory employee profit sharing plan, which accounts for 10% of the taxable profits
excluding the inflation effect. However, losses of prior years are not deductible in computing profit to be
shared. Furthermore, the portion of profits shared by employees is not deductible for the income tax
purpose. However, new enterprises are exemptfrom profit sharing for thefirst year of operation and those
engaged in manufacturing a new product are exempt for thefirst 2 years of operation.
Tax incentives
The govemment has offered many incentives promoting various economic activities, most of which are
available almost exclusively to Mexican-controlled enterprises or to branches of those foreign companies
that have formally committed themselves to shifting eventually to "Mexicanization." The most popular
incentives available to foreign investors are in-bond-assembly industries, or maquiladoras, although there
is no tax incentive designed specifically to encourage the foreign investment. The main tax incentive
related to maquiladoras is the rather generous safe-harbor rule, which set the minimum taxable income as
5% of the total valué of assets used in the operation. However, there is presently a clear trend towards
treating maquiladoras for tax purposes in the same way as any other Mexican Corporation.

58

The m xm m taxable a o n is defined by specific times the mnm m salaries, which varies f o 1 to 2 mnm m
ái u
mu t
ii u
rm 5
5 ii u
salaries depending on the category of contribution. The m xm m a o n will be set at 2 mnm m salaries for all
ái u mu t
5 ii u
categories of contribution in year 2 0 as some currently levies with lower taxable base being gradually reduced.
07
Sesión 2 - 205

5. PERU
The capital tax provisions
The corporate income tax rate. The corporate income tax is 30% in Perú. The taxable income for a
residential corporation is its worldwide income from all sources, while that for a non-residential
corporation is its income derived from its Peruvian sources.
Mínimum tax on net assets. A mínimum tax of 0.5 % was levied (in 1998) on the net assets of
corporations. The net assets are defined as total assets less depreciation and amortization allowed for the
corporate income tax purpose with certain exclusions and exemptions59. The mínimum tax paid can be
used to as a credit against CIT liability. Unused mínimum tax credit can be carried forward to offset CIT
liability for the year 1999 - 2001.
There is also a mandatory profit sharing with employees. The sharing rate varíes depending on a
company's activities and is based on the pre-tax income. The profit shared with employees under this plan
is deductible as an expense for the CIT purpose.
The tax depreciation rule. The tax depreciation is based on the straight-line method. Depreciation is
computed on acquisition cost of fixed assets plus an inflation adjustment. The annual depreciation rate set
by law is 3% for buildings in general60, 20% for plant and machinery including data processing
equipment, and 10% for other assets. A more generous depreciation rate may be authorized upon
taxpayer's request, taking into account the characteristics of the business (e.g., more than one shift for the
machinery utilization).
Inventory accounting method. Inventories must be carried at cost as adjusted for inflation. Cost may be
determined specifically or by the FIFO, average, retail or basic inventory method, but the LIFO method is
not permitted.
Loss carry-overs. Business losses from Peruvian sources may be carried forward for 4 years, beginning
with the first subsequent year when taxable income arises. A carry-back of losses is not allowed. Losses
carried forward must be adjusted for inflation.
The withholding tax rate on dividends.
The property tax. Urban property tax is levied at the municipal level. The tax rate is progressive from
0.2% to 1.0%.

59
60

For example, a givenfirm's shares or participation in t e capital of other enterprises that are subject to t e mnm m tax are
h
h íi u
excluded f o thisfirm's taxable net assets. Refer to Emst & Young (1999) for m r details.
rm
oe
For buildings and structure of new lodging enterprises, the rate is 10%.
Sesión 2 - 206

The transaction taxes
VAT is levied at a single rate of 16%. A municipal sales tax of 2% is also applicable to any taxable
amount under VAT. Since this tax is not creditable for VAT purposes and follows all the VAT rules, it
virtually forms an additional VAT. As a result, the effective VAT rate is 18%.
A tax of 3% is levied on transfers of immovable property.
Imports are subject to customs duties. The general rate is 12%. Higher rates may apply to specific goods.
The payroll taxes
Social security contribution is charged based on the total salary. The contribution payable by the
employer is 9% for health fund, 1% - 12.2% as a premium for work accidents depending on the degree of
risk related to working condition, and a payroll tax of 5% under the former housing fund. The social
security contribution to the national pensión system is payable by the employee, and the rate is 11% of the
total salary.
There is also a compulsory profit sharing with employees. The sharing basis is the annual pretax profits
excluding previous years' losses. The share is 10% for manufacturing industry and 8% for service sector.
The portion of profits shared with employees is deductible for the income tax purpose.
Tax incentives
Perú introduced a generous tax incentive package, known as "Amazon Law" to promote economic
development to its Amazon región. The package includes exemptions for income tax, VAT, excise taxes
and more. Most importantly, the law is stated to have a duration of 50 years, which most observers do not
believe sustainable . Other Peruvian tax incentives for regional development are for the border región.
Industrial enterprises in the border región are exempted from all taxes except social security contribution,
import duties and some municipal taxes.

61

Byrne (1999).
Sesión 2 - 207

Table A1
Business Tax Provisions Applicable to Manufacturing and Service Industries
Argentina

Brazil

Chile

México

Perú

35

27-37

15

35

30

1-5

2.65(turnover)/

EPS

10 (EPS)

10 (EPS)

Not available

0.25 - 0.5

1.8

0.5

None

None

None

None

1. The Capital Taxes
Corporate

Other

income

tax

on

Assets-based

tax1.0

Thin

Yes
capitalization

Tax

depreciation

Buildinas

2.0 SL

4.0 SL

7.5 SL

4.0 SL

3.0-10.0 SL

Machinen/

10 SL

10 SL

15.2 SL

10 SL

20.0 SL

FIFO

FIFO

average

FIFO/averag

504
years (F)

5 years (F/R)

Unlimited

10 years

4 years (F)

WH tax on dividends

0

0

20

5

0

Urban Propertv taxes

Max. 1.5

Not available

1.425

FD.

1.5

2

None

3.3

3.0

on

None

None

None

None

None

tax

Yes

Yes

Yes

Yes

Yes

21

Dual-VAT

18

15

18

0-15

Up to 55

11 (general)

11 (average)

12 (general)

Inventory

accounting
LIFO

Loss Carry-ovet

Property
Tax

transfer

incentives

Other

0.131-

Progr. 0.2 -

2. The Indirect Tax

VAT
Import

duty

3. The Payroil Taxes (including payroll-based social security

Employer

62

64

25.8

(0.9 - 3.4) +

20.52

15-26.2

Employee

63

33
Up to 9

12.5

19.6-20.4

18.71

9.0

Turnover = t e tumover tax, and EPS = t e m n a o y employee's profit sharing o taxable income.
h
h a dtr
n
As the classification of depreciable assets varíes by country, please refer to the text for details. Also note t a DB = declininght
b al anee method, a d SL = straight-line method.
n
Following t e number of years for loss carry over, t e letters in brackets indícate the following: F = forward, B = backward,
h
h
and R = certain restriction in the valué of loss to be written off. Please refer to t e text for details.
h
See Table A2 for detailed incentives provided by each of these countries.

Sesión 2 - 208

Table A2: Tax Incentives in Relation to Manufacturing and Service Sectors

Target of the incentive
Assembly operation

Description
favorable income tax treatment
exemption from import-related taxes
exemption from the social security contribution

Export promotion

transaction tax exemption for export-related inputs
gross-receipts tax exemption
import duty refund for export-related inputs
refund of charges affecting export-related inputs

Free Zone

Regional development

Tourism promotion

Argentina

complete tax exemption
income tax reduction or exemption
exemption from import-related taxes
reduction in income tax through direct or indirect investment
exemption from income tax and assets tax

México
Yes
Yes
No

Yes
Yes

20% of annual minimum salary for each additional man-year

Perú

Yes
Yes
Yes
1 years
5

by 40%

exemption
Yes
Yes

by 40%
50 years

accelerated depreciation for lodging buildings
exemption from property tax

Investment in new M&E option for tax credit or accelerated depreciation allowance
fast write-off (62.5%) for annual increment in M&E investment
Job creation tax credit

Brazil
Chile
Up to 20yr.
Up to 20yr.
Up to 20yr.

2-5 years
2-5 years
Yes
Yes
Yes

APPENDIX 2
The United States: Tax and Economic Parameters
The combined corporate income tax rate in the United States is 38.9% with 35% at the federal level and an
average 6% at the state level, and the state tax can be deducted for the federal tax purpose.
The expected inflation rate and interest rate is 2.3% and 8.4% respectively. Both are based on the
International Financial Statistics, published by the IMF. The former is estimated based on the consumer
price index for the period of 1995 - 1998, and the latter is based on the 1998 lending rate.
Table A3
The United States' Tax and Economic Parameters (%), 1999

CIT rate

Tax rate
on
interest income

Tax rate
on
equity income

Interest rate

38.9

30

20

8.4

Sesión 2 - 210

inflation rate

2.3

APPENDIX 3
Impact of Non-tax Parameters on the Estímate of Effective Tax Rate
Expected Inflation Rate
The expected inflation rate affects the effective tax rate on capital through its impact on the
nominal interest rate. For a given real interest rate, the higher the inflation rate, the higher the nominal
interest rate will be. When there is no regulation for adjusting the inflation impact, the nominal interest
rate interacts with taxes mainly through the following three channels. Firstly, interest cost is deductible
for income tax purposes at the nominal rate. As a result, the higher the nominal interest rate in relation to
afixed real interest rate, the lower the real after-tax financing cost, and henee the lower the METR. This
effect is particularly favorable for leveraged landfinancing. Secondly, The accumulated present valué of
a given annual tax depreciation allowance decreases as the nominal interest rate rises. Since higher
inflation lowers the present valué of tax depreciation allowance, it increases METR on depreciable assets.
Andfinally, if thefirst-in-first-out method is used for the inventory accounting, it may results in inflated
taxable income and, henee, a higher METR on inventory when prices rise. Since inflation thus affects
METR on different assets in different directions, its net impact will depend upon the capital structure.
However, since all countries but Argentina under our study adopt a rule that is similar to tax indexation,
the inflation impact described here would be largely avoided and henee is not reflected on our estímate of
effective tax rate. The inflation rates shown in Table A3 are based on the consumer price indexes since
1995, published by the International Monetary Fund, International Financial Statistics.
Expected Real Interest Rate
The impact of the real interest rate on the effective tax rate is in part similar to the impact of
inflation since as the real rate rises, so will the nominal rate, thus increasing the effective tax rate on
depreciable assets. For a given debt-asset ratio, however, unless inflation is high, there is unlikely to be
much of a distortion in effective tax rate arising from the deductibility of interest. The expected real
interest rate used for our study is the expected real interest in the US assuming that the foreign investors,
i.e., the US multinationals in our study, will borrow at the same real interest rate everywhere as that
available at home. As shown in Table A4, Appendix 3, the real interest rate corresponding to the nominal
interest rate of 8.4% and the inflation rate of 2.3% in the US is 6.1%.
Debt-asset Ratio
The ratio of debt to assets is sometimes referred to as thefinancing structure. As already noted,
the impact of this ratio on the effective tax rate is related to the expected inflation rate and (real) interest
rate. For a given inflation rate and real interest rate, the higher the debt-asset ratio, the more the potential
benefit from the tax deductibility for debtfinancing cost, or interest expenses. A higher debt-asset ratio
may thus reduce effective tax rate through lowering the real after-tax cost of financing.66
Economic Depreciation
The economic depreciation rate interacts with the tax depreciation allowance to affect the
effective tax rate. Suppose, for example, under our assumption of fully mobile capital and technology
that a given type of machinery is depreciated at the same economic rate everywhere around the world.
For example, Argentina, like m n countries around t e world, imposes a thin capitalization rule to discourage tax-driven
ay
h
borrowing.

66

Sesión 2 - 2 1 1

Countries with faster tax depreciation allowances for this type of machinery will then encourage this type
of capital investment through a lower effective tax rate.
Capital Structure
A real capital investment generally involves two categories of capital: depreciable and nondepreciable assets. These two categories can be further divided into four types: buildings and machinery
(both depreciable) and inventory and land (non-depreciable). Capital investments in different industries
are as a rule structured differently. Moreover, under the same statutory tax rate(s), different types of assets
may incur different effective tax rate due to the various interactions between tax provisión and non-tax
parameters discussed above. In the absence of other information, we use the same estimated capital
structure to aggregate these differentiated effective tax rate on various type of capital for all countries, but
of course different capital structure in different industries nonetheless result in different aggregated
effective tax rate for different industries.

Sesión 2 - 212

Table A4
Non-Tax Parameters
(in per cent)
Argentina

Brazil

Chile

México

Perú

Expected inflation rate

0.5

8.5

6.3

23.4

9.1

Expected real interest rate

6.1

6.1

6.1

6.1

6.1

40.0

40.0

40.0

40.0

40.0

29.0
38.0

29.0
38.0

29.0
38.0

29.0
38.0

29.0
38.0

4.0
18.7

5.3
18.9

3.5
22.7

4.2
21.2

3.6
23.9

5.0
30.0

5.0
35.0

5.0
30.0

5.0
39.0

5.0
31.0

24.0
38.1
35.9
2.0

24.0
38.1
35.9
2.0

24.0
38.1
35.9
2.0

24.0
38.1
35.9
2.0

24.0
38.1
35.9
2.0

60.6
11.7
9.5
18.2

60.6
11.7
9.5
18.2

60.6
11.7
9.5
18.2

60.6
11.7
9.5
18.2

60.6
11.7
9.5
18.2

Debt to assets ratio
Debt raised abroad to home capital
Debt to assets ratio in home country

Manufacturing
Services
Economic depreciation rate
Manufacturing
Buildings
Machinery
Services
Buildings
Machinery
Capital structure by asset type
Manufacturing
Buildings
Machinery
Inventory
Land
Services
Buildings
Machinery
Inventory
Land

Sesión 2 - 2 1 3

APPENDIX 4
Methodology for Estimating Effective Tax Rate
The effective tax rate on capital calculated in our study is an effective corporate tax rate on capital,
which combines all the taxes that would affect the capital investment at the corporate level in stead of
personal level. Personal income taxes, however, may be incorporated into our calculation when they
affect the capital investment at the corporate level. For example, the withholding tax on dividends
certainly affects the cost of capital invested at the corporate level through its impact on the rate of retum
to equity required by shareholders.
Our estimates of the effective corporate tax rate are made mainly for the foreign capital investments,
with the US multinationals as representative foreign direct investors in Latin America. Furthermore,
our estimates are made for two major sectors: the manufacturing and service industries. The simulation
for each sector is undertaken for both regular taxable case and tax-incentive case.
All the estimates of effective tax rates are based on 1999 tax regimes and the latest economic information
available, unless specified otherwise.
1. Theconcept
The effective tax rate measures the impact of a tax system on an incremental unit of capital investment or
business activity. It incorporates the effects of not only statutory tax rates and related tax treatments (e.g.
tax depreciation, tax credit, tax deductibility, tax holidays, etc.) but also various economic factors
interacting with these tax treatments (e.g.financial costs, the inflation rate, and the structure of
investment, etc). In other words, the effective tax rate is a summary indicator of the overall tax burden
imposed by a tax system on an investment within a certain economic environment.
The calculation of effective tax rate is based on the assumption of profit-maximization. Profit-maximizing
firms base their investment/business decisions on the foreseeable incremental net revenue at the present
valué. Taxes reduce the portion of the profits accruing to the investor, while tax allowances mitígate such
a reduction in accrued profits. Due to the interaction between these statutory tax provisions and the actual
economic/industrial conditions (e.g.,financing condition, capital structure, input structure of production,
etc.), the effective tax rates can vary by industry even under the same tax regime. Furthermore, as
mentioned above, for a cross-jurisdiction comparison, the difference in effective tax rate may reflect not
only the variation in tax regimes but also the differentiated economic andfinancial climates.
For profit-maximizing firms, the gross rate of return on capital (net of economic depreciation) must be
equal to thefinancing cost of capital, adjusted for taxes. The size of this adjustment for taxes on a new
investment is the effective tax rate on capital. For example, if the gross-of-tax rate of retum to capital is
20 per cent and the net-of-tax rate of return is 10 per cent, then the effective tax rate on capital is 50 per
cent.
It should be noted that the analysis of effective tax rate provided in this study deals with only 'profitable'
firms. By 'profitable' we mean thosefirms that have taxable income, if not granted a tax holiday. This
assumption is important because that a tax holiday is irrelevant to an unprofitable firm that does not have
to pay income taxes.

Sesión 2 - 2 1 4

2. Methodology
The standard method used to estímate effective tax rates has been extensively documented67. The formula
based on this method has been modified by incorporating some miscellaneous taxes such as capital tax,
property tax, and tax on transfer of property68. The following are general formulas used in this study. It
should be noticed that, for our study, all the formulas apply to foreign investors only.
i) Effective tax rate (t)
The effective tax rate on a given type of capital is defined as the proportional difference between the
gross-of-tax rate of return required by a firm (r G ) and the net-of-tax rate of return required by an investor
(Z1). rG is the difference between the marginal revenue product (or user cost, in equilibrium) and
economic depreciation. The after-tax rate of return is the weighted average of the return to debt and
equity securities held by the investor. Thus, the effective tax rate (t) is defined as
t = (rG - T»)!/3

(1)

ii) The real cost of financing {/)
That is, the cost of financing for an investor is the weighted average cost of financing net of inflation
rate.
For foreign investors, the real cost of financing (/) is defined by
/' = [FI'I'(l

- U') + (1 -FI')p']*(l

-Y Y (I-X) +Y *fi(l-U) - n +%'] - 71'

(2)

withfi' = debt to assets ratio in home country, i' = cost of debt in home country, U' = the statutory
corporate income tax rate in home country, p' = cost of equity in home country, y = the ratio of debt raised
in host country to total investment fund, x = weighted average withholding tax rate in host country, i =
cost of debt in host country, U = statutory corporate income tax rate in host country, n' = inflation rate in
home country, and Jt= inflation rate in host country. That is, the cost of financing to a foreign investor is
the weighted average of costs of its investment fund takenfrom home country and the debt raised
abroad. The former is the weighted average cost of financing at home net of withholding tax payable in
host country, and the latter is the cost of debt in host country adjusted by income tax deductibility and
difference in inflation rate between home and host countries.
iii) The net-of-tax rate of return on capital (/•)
A. Net-of-corporate-tax rate of return on capital
For afinancial investor, or a supplier of the investment fund, the net-of-corporate-tax rate of retum if)
can be estimated as
r» =[fi'i'(l-U')

67
68

+ (1 -fi')p' - n'](1- y ) +y(i-7t)

Boadway, Bruce, and Mintz (1984,1987) and Mintz (1990).
Chen a d Mintz (1993).
n
Sesión 2 - 2 1 5

(3)

This is the net-of-tax rate of return on capital required by fund suppliers including foreign investor itself
and creditors in host countries.
Applying equation (3) to equation (1) results in the effective tax rate on capital for foreign investors.
iv) The gross-of-tax rate of return (r°) on capital
A. Depreciable assets
RG = (1 +tm)(/'+8)(l- k)(l - A)/[(l-U)(l-tg)J

(5)

+tp/(l-tg) - 8

with tm = tax on transfer of property, or sales tax on capital goods where is applicable, 8 = economic
depreciation rate, h = capital risk, k = investment tax credit rate, A = present tax valué of the
accumulated capital cost allowance, a = tax depreciation rate, tp = property tax rate and/or assetsbased tax rate, and tg = gross receipts tax rate.
B. Inventory

rG = (1 +tm)(/+

UjQ/[(1-U)

(1 -tg)]

(6)

with tm = sales tax on inventory where it is applicable, and £= 1 for FIFO accounting method and 0
for LIFO.
C. Land
r G = (l+tm)//[(l-U)(l-tg)]

+ tp/(l-tg)

(7)

v) Aggregation
The effective tax rate for a given industry is the difference between the weighted average of gross-of-tax
rate of return by asset type and the net-of-tax rate of return (which is the same across asset type within the
industry) divided by the gross-of-tax rate of return on capital. That is, the marginal effective tax rate for
industry i (t¡) is calculated as follows:
í, = (Ej rGijWy - rN J/Zj rGijWij

(8)

where j denotes asset type (i.e. investments in buildings, machinery, inventories, and land), wy denotes the
weight of asset type j in industry i.
The above are general format of formulas used in this study. Due to the variance among different
jurisdictions, some variables in the formulas can be zero for some jurisdictions. For example, there is no
gross-receipts tax in Chile, México and Perú and henee tg = 0 in equation (5) - (7) for these countries.

Sesión 2 - 2 1 6

GRAVACIÓN IMPOSITIVA E INCENTIVOS FISCALES
A LA INVERSIÓN EXTRANJERA DIRECTA
LA OPINIÓN DEL SERVICIO DE ASESORÍA
SOBRE INVERSIÓN EXTRANJERA

Joel Bergsman *
Muchos economistas rechazan los incentivos fiscales pues los consideran, en el
mejor de los casos, medidas ineficaces que suponen pérdidas de ingresos para el
erario público y, en el peor, costosas distorsiones que reducen el verdadero valor de
la producción. Por otra parte, a la mayoría de los directores de los organismos de
promoción de inversiones extranjeras prácticamente no hay incentivo que les parezca
mal.
—
"No sirve de mucho desaconsejar el empleo de incentivos con el argumento de que
causan distorsiones en los mercados de factores o productos, cuando estas
distorsiones son, justamente, lo que el cliente busca."
"¡Sr. Bergsman, usted no sabe lo que dice!" (Funcionario del Ministerio de
Finanzas húngaro, discutiendo sobre las treguas tributarias con un asesor del FIAS a
comienzos del decenio de 1990).
* El autor, uno de los creadores del FIAS, dejó el cargo que tenía en ese Servicio para acogerse a los
beneficios de la jubilación en octubre de 1998, y ahora es un consultor independiente en St. Leonard
(Maryland, Estados Unidos). Esta nota es un proyecto común, al que han hecho contribuciones muchos
colegas, del FIAS y de otros organismos, a los que no nombro por temor a olvidarme de alguno. Con todo,
Neda Pirnia merece un agradecimiento especial. Enero el 2000

Sesión 2 - 2 1 7

Indice

I.

Criterios tradicionales: coincidencias y discrepancias

219

II.

Consecuencias para el asesoramiento sobre políticas

222

Bibliografía

227

Sesión 2 - 2 1 8

I.

Criterios tradicionales: coincidencias y discrepancias

¿Cumplen los incentivosfiscales eficazmente sus objetivos, es decir, atraen nuevas inversiones? De
ser así, ¿los beneficios superan a los costos?
Según la experiencia recogida por el FIAS en el asesoramiento a decenas de países en desarrollo y
en transición en los últimos 14 años, los sistemas de incentivos son, en su mayoría, ineficaces. Atraen muy
pocas inversiones adicionales y tienen sus costos: constituyen un drenaje de fondos para las arcas fiscales
de los países que los aplican, en algunos casos resultan contraproducentes porque complican demasiado
los procedimientos para las inversiones y a veces dan origen a una corrupción mucho mayor. El analista
de políticas no necesita más para recomendar algún tipo de cambio.
Una parte fundamental de la labor de asesoramiento del FIAS consiste en ayudar a la mayoría de los
clientes a comprender que los incentivos demasiado generosos o, sencillamente, una escasa presión
tributaria no atraen, por sí solos, inversión extranjera directa. (La cargafiscal elevada y/o la inestabilidad
de las normas pueden ahuyentarla, de manera que si los impuestos pasan de muy elevados a moderados el
ingreso de inversión extranjera directa podría aumentar.) Si alguien pone en duda esta afirmación, puede
considerar el caso de la mayoría de los países de África al sur del Sahara, cuyos incentivosfiscales y
generosas treguas tributarias reducen efectivamente la presión impositiva para los inversionistas
extranjeros a niveles muy bajos y que, sin embargo, reciben muy pocas inversiones extranjeras directas.
Nepal es otro caso detectado recientemente por el FIAS. Los motivos de la ineficacia de los incentivos en
estos países son evidentes: con exclusión de algunas actividades basadas en recursos naturales, no ofrecen
buenas oportunidades de hacer dinero en los tipos de actividades a los que se orienta la inversión
extranjera directa; muchos de estos países también son inestables o presentan dificultades no económicas;
sus mercados internos son pequeños y están alejados de las fuentes de producción y los mercados
extranjeros. Es por ello que ni siquiera la ausencia de impuestos atraerá inversión extranjera directa hacia
estos países o regiones, como no sea hacia los segmentos que despiertan especial interés, los cuales
probablemente la seguirían recibiendo incluso con una moderada presión impositiva.
Estas conclusiones, a las que se ha llegado por simple observación y sentido común, están avaladas
tanto por estudios econométricos como por las numerosas encuestas realizadas entre los inversionistas,
que generalmente coinciden en que ni los incentivosfiscales ni siquiera la presión impositiva, a menos que
sea excesiva, determinan, en la mayoría de los casos, la elección del lugar donde se asentará la inversión
extranjera directa. Algunos de estos estudios y encuestas se mencionan en la bibliografía.
Sin embargo, el mundo no es así de simple. Muchos funcionarios públicos juran que sin incentivos
no podrían atraer inversión extranjera directa. Esto ya era evidente hace casi 40 años, cuando Robinson
(1961) presentó los resultados de una encuesta sobre 365 inversiones en 67 países, efectuada entre 205
compañías. Los resultados son, básicamente, los mismos a los que han llegado las numerosas encuestas
realizadas desde entonces: los funcionarios públicos otorgan a los incentivos un alto grado de prioridad
entre los factores decisivos, a diferencia de los ejecutivos de las empresas multinacionales.
El tema se pudo comprender mejor cuando la pregunta se formuló con mayor precisión. Reuber
(1973) sostuvo que las empresas multinacionales no son todas iguales y reaccionan de distinta forma ante
los diferentes tiposde incentivos, según sus motivaciones. Reuber estudió 80 proyectos de inversión; en
los orientados a la exportación, los incentivosfiscales (incluyendo las treguas tributarias, la desgravación
de derechos y la amortización acelerada) se mencionaron como importantes en el 48% de las decisiones
sobre inversiones, mientras que, respecto de las inversiones dirigidas al mercado interno, en el 56% de las
respuestas se afirmó que el factor de mayor importancia era la protección.
En un importante estudio financiado por el Comité para el Desarrollo y supervisado por la
Corporación Financiera Internacional (CFI), Guisinger y sus colaboradores (1985) llegaron a la misma
conclusión sobre la diferente pertinencia de los incentivosfiscales o la protección para los distintos tipos
Sesión 2 - 2 1 9

de inversión extranjera directa1'. En este último estudio también se presentó otro aspecto importante de la
cuestión: por lo general, para los inversionistas los incentivos no son un factor determinante de la
inversión si corren más o menos parejos con los de los países competidores. En las encuestas menos
complejas simplemente se preguntaba si los incentivos habían influido en las decisiones sobre inversiones;
cuando había incentivos similares en muchos países, la mayoría de los inversionistas respondía que no.
Guisinger formuló una pregunta diferente: ¿habría sido distinto si el país no hubiera ofrecido incentivos y
sí lo hubieran hecho los países competidores? Un número considerable de inversionistas contestó que sí.
Finalmente, fue en el estudio de Guisinger donde quizás por primera vez se enunció de manera clara y
explícita lo que la experiencia parece demostrar: los incentivosfiscales (o, para decirlo en términos más
generales, aquellos encaminados a reducir los costos) suelen ser eficaces en dos situaciones aparentemente
diferentes pero, en el fondo, similares, a saber, cuando la inversión debe decidirse entre:
•
Plataformas que presentan atractivos semejantes para la producción destinada a la exportación a
otros mercados, o
•
Partes, de atractivo similar, integrantes de un gran mercado (como estados de los Estados Unidos o
el Brasil, o países de la Unión Europea, entonces Mercado Común).
En otras palabras, cuando el inversionista está estudiando distintos lugares que ofrecen condiciones
aceptables y bastante similares para la producción con destino a la exportación hacia un mercado mucho
más grande que las jurisdicciones políticas en cuestión, enfrenta la competencia real o potencial de otros
inversionistas que pueden elegir los mismos lugares, de modo que reducir los costos es muy importante.
Los organismos de promoción de inversiones de Tailandia o Carolina del Norte o Irlanda lo saben y por
ello ofrecen generosos incentivosfiscales, subvenciones en efectivo, tierras a precio reducido, incentivos a
la capacitación del personal y otros estímulos, a pesar de que numerosos estudios econométricos,
encuestas de inversionistas y economistas les aconsejan no hacerlo.
Ello también explica por qué la mayoría de los países grandes que despiertan interés no ofrecen
incentivos en el orden nacional: por ejemplo, los gobiernos nacionales de los Estados Unidos, Brasil y
México no proporcionan incentivos a las inversiones extranjeras (si bien Brasil y México, y los Estados
Unidos hace un tiempo, otorgaban gran protección a sus mercados internos, y Brasil no ha abandonado
por completo esta estrategia)27.
Rolfe y colaboradores (1993) señalaron que las preferencias son también función de factores como
la etapa en que se encuentra la inversión (inicial o expansión) y el tipo de proceso (fabricación o
servicios). Lasfirmas que recién se instalan prefieren los incentivos que reducen los gastos iniciales,
mientras que las que se amplían prefieren los relacionados con los impuestos que inciden en las utilidades;
las industrias manufactureras, que necesitan mayores inversiones en activos fijos, suelen inclinarse por los
incentivos vinculados a los activos depreciables más que las empresas de servicios. Bergsman y Edisis
"

La presentación de Guisinger provocó u "debate innecesario" pues el autor subrayó reiteradamente que los resultados de su
n
trabajo, según los cuales en m c a decisiones sobre inversiones influían los incentivos -entendidos como él los definía-,
uhs
contradecían a otros de escaso efecto. En realidad, los resultados a los que arribó Guisinger fueron coherentes con los de los
demás estudios; la diferencia principal consiste e que Guisinger incluyó la protección frente a las importaciones entre los
n
incentivos, mientras que en pocos estudios se adoptó el ms o criterio. Guisinger efectuó u importante aporte a la
im
n
comprensión de los incentivos, pero no modificó radicalmente los conceptos sobre el tema. Véase Wells (1986).

27

China constituye una excepción interesante. Cuando China abrió sus puertas a la inversión extranjera directa e 1979, puso e
n
n
marcha tres medidas importantes que favorecieron a esta en relación con las empresas nacionales (prácticamente todas
estatales): menores impuestos sobre las utilidades, libertad de retener el 1 0 de todos sus ingresos en divisas y mayor
0%
libertad para contratar y despedir trabajadores. Estas dos últimas no hicieron m s que otorgar a las empresas extranjeras el
á
tratamiento internacional corriente, lo cual equivale a admitir que las restricciones aplicadas a las empresas públicas chinas
impedirían casi por completo la inversión extranjera directa. Haciendo un análisis retrospectivo, es probable que las menores
tasas impositivas no generaran grandes inversiones adicionales y, por lo tanto, hayan representado meras pérdidas para la
tesorería.

Sesión 2 - 220

(1988), funcionarios del FIAS, en un estudio acerca de la eficacia de la capitalización de la deuda como
incentivo para los inversionistas en América Latina en los años ochenta, confirmaron la importancia de
reducir los gastos iniciales: el descuento de swaps reducía el costo de la inversión inicial y no dependía de
las ganancias futuras. Este incentivo inmediato y seguro no sólo aumentaba los ingresos previstos sino que
reducía el monto en riesgo; afectó la decisión sobre la inversión en más de la mitad de los casos
estudiados. Este resultado coincide con el éxito logrado por Manda y otros países y regiones que, al
parecer, atrajeron una buena cantidad de inversionistas ofreciendo generosos conjuntos de incentivos que
incluían subvenciones iniciales.
Ahora comprendemos que la eficacia de los incentivos varía según el tipo de inversionista, la clase
de incentivo, el grado de competencia para conseguir la inversión, la calidad del lugar y la reacción de
otros países ante el aumento o la disminución de incentivos de sus competidores. Los incentivosfiscales u
otros que reducen costos suelen ser eficaces:
•

para inversiones sin vinculación permanente, orientadas a la exportación;

•

en países o regiones similares a países o regiones vecinos;

•

en lugares donde los demás aspectos económicos también son favorables, y

•

donde los incentivos inciden en las etapas iniciales del proyecto y presentan mayor grado de
certeza.

La oferta de un mercado protegido también puede resultar muy eficaz para atraer inversiones, en
especial hacia una economía grande y en crecimiento, como lo demuestran claramente los principales
ejemplos de América Latina en los decenios de 1950 y 1960 y China, últimamente. En realidad, la
protección es aún más poderosa, como lo prueba el hecho de que también ha atraído un gran número de
inversiones muy ineficientes hacia economías medianas e, incluso, pequeñas: desde plantas de montaje de
automóviles en Chile (14 millones de habitantes) hasta plantas de laminado de acero en Togo. Pero en la
actualidad esta práctica parece despertar menos interés y está cayendo en desuso en la mayoría de los
lugares, debido a los perjuicios que ocasiona la constante protección contra la competencia y las prácticas
del comercio mundial. Las propias empresas multinacionales se instalan cada vez más en sitios donde
puedan producir según las normas de la competencia mundial, aun en aquellos casos en los cuales el
principal objetivo estratégico en el momento de la inversión sea el mercado local.
Lo mismo se puede expresar diciendo que, al momento de decidir si efectuar o no una inversión, los
inversionistas analizan el conjunto de características pertinentes del lugar. Son contados los casos en los
que los incentivosfiscales o de otro tipo tendientes a reducir costos logran revocar una decisión negativa
adoptada en razón de deficiencias graves del lugar. Los estímulos tales como los diferentes tipos de
desgravación impositiva o las tierras a bajo precio inciden sólo cuando el lugar a) es aceptable y b) es
similar a otros con los que compite. La protección puede tener más influencia si el mercado protegido
parece lo suficientemente rentable.
En realidad, un número importante de compañías multinacionales tiene por norma tomar las
decisiones sobre inversiones haciendo caso omiso de los impuestos y los incentivosfiscales. Sólo toman
en cuenta los factores que consideran más esenciales. Por supuesto, una vez adoptada una decisión
positiva, negocian cuanto sea posible para obtener todos los incentivos a su alcance.
Al estudiar los países en desarrollo y en transición, vemos que todavía hay varios que parecen
ofrecer incentivos cuando, según los argumentos expuestos hasta ahora en este documento, lo más
probable es que este tipo de incentivos atraiga muy poca inversión extranjera directa adicional y
constituya, por consiguiente, una pérdida de dinero para la nación. Hay cuatro hechos que quizás
expliquen, en la mayor parte de los casos, esta aparente anomalía.
En primer lugar, muchos de los funcionarios públicos que están convencidos de la eficacia de los
estímulos para atraer inversiones basan su opinión en su experiencia cotidiana con los inversionistas,

Sesión 2 - 2 2 1

quienes negocian con ahínco para obtener todos los incentivos a su alcance. Estos funcionarios saben que
los inversionistas quieren conseguir todo incentivo permitido por las leyes del país y harán todo lo posible
para lograrlo; lo que ignoran es cuántos inversionistas no optarían por ese país si este les ofreciera una
carga impositiva razonablemente reducida y condiciones relativamente satisfactorias, sin otro incentivo.
Como asesores, creemos saber que, con excepción de los inversionistas orientados a la exportación y sin
vinculación permanente, la falta de estímulos en situaciones que, por lo demás, sean ventajosas no
desanimaría a muchos inversionistas. Pero, cuando el funcionario trata de lograr que una planta de Ford o
de Siemens se instale en su país, se enfrenta con ejecutivos que exigen toda clase de exencionesfiscales y
otros estímulos; la solución de ofrecer condiciones impositivas aceptables, sin incentivos, se encuentra
entre las opciones que nos parecen más convenientes pero no entra en su marco analítico.
En segundo término, algunos de los gobiernos que continúan ofreciendo incentivos de dudosa
eficacia procuran alcanzar un objetivo distinto de aquel al que alude nuestro trabajo. Singapur es el
ejemplo clásico. Nosotros nos referimos a los incentivos para aumentar la cuantía de la inversión
extranjera directa, pero la Junta de Desarrollo Económico de Singapur está tratando de poner en marcha
una política industrial. No procura tanto aumentar el ingreso total de dichas inversiones como atraer
determinadas actividades, incluso determinadas compañías, para modificar la estructura —no el volumen
total— de la actividad económica del país. Si se trata o no de una buena idea es otra cuestión, pero el
argumento de que Singapur obtendría cuantiosas inversiones incluso sin incentivos no es pertinente
porque no se refiere al objetivo del país.
En tercer lugar, los incentivos pueden tener cierta utilidad como señal. La vida es muy compleja y
no se la puede reducir a análisis económicos; incluso en las decisiones sobre inversiones influyen muchos
factores diferentes. Especialmente en el caso de un país que desee transformar una imagen de hostilidad u
oposición a los inversionistas y demostrar que está dispuesto a facilitar la actividad económica privada en
general y la inversión extranjera directa en particular, el anuncio de una serie de incentivos puede resultar
útil aunque más no sea por su efecto como medida de relaciones públicas. Si esos incentivos son la vía
más eficaz en función de los costos para alcanzar dicho objetivo, es otra cuestión, pero su utilidad bien
puede trascender el efecto estrictamente económico que se refleja en el estado de pérdidas y ganancias de
las compañías. Se cree que esafinalidad es la que persigue la República Checa con los incentivos, cada
vez mayores, que viene ofreciendo desde abril de 1998.
Por último, no olvidemos que, en su mayoría, aquellos funcionarios públicos que sostienen que los
incentivos son eficaces y necesarios para su país o provincia son quienes los conceden o están encargados
de atraer inversiones del exterior. Dejan en manos de otros la cuestión de la eficacia en función de los
costos; tienen una misión que cumplir y para ello, naturalmente, desean utilizar todos los instrumentos a
su disposición. Quieren tener atribuciones para otorgar incentivos y gran libertad para hacerlo. En estos
casos es preciso analizar si lo mejor para el organismo de promoción de inversiones es también lo mejor
para el país.

II.

Consecuencias para el asesoramiento sobre políticas

Lo antedicho se puede resumir en la afirmación de que, siempre que los impuestos sean razonables,
los incentivos no serán necesarios para atraer inversiones a) en la extracción de recursos naturales u otras
actividades que generen renta, o b) en los proyectos encaminados a la venta en el mercado interno. El tipo
principal de inversión extranjera directa no incluido en estas dos categorías es la industria manufacturera
sin asientofijo, dedicada a la exportación, que se puede atraer con mano de obra económica (y, en menor
medida, con los cupos de importación, aplicados por los países ricos). Un país pequeño que desee atraer
cuantiosas inversiones extranjeras directas y cuyo principal atractivo sea el que puede ofrecer a esas
industrias manufactureras, seguramente necesitará reducir la presión tributaria (y probablemente también
la tasa impositiva establecida por la ley) más que un país grande con un mercado interno interesante.
Sesión 2 - 222

Hay numerosos conjuntos de medidas impositivas que no ahuyentan a la inversión extranjera
directa. (Pueden variar de un país a otro). Al elegir entre ellos, algunos de los temas que un país —y los
asesores del grupo del Banco Mundial— debe tener presentes son:
•
La historia y la situación actual del país y los cambios que éste encuentra aceptables desde el punto
de vista político y administrativo.
•
Las cuestiones impositivas; no hay que olvidar que el propósito de un sistema tributario no es atraer
inversión extranjera directa sino recaudar impuestos. En este sentido, no hay nada más triste que un país
muy pobre, con unas pocas actividades especializadas de interés para la inversión extranjera directa,
dispuesto a permitir que cualquier inversionista del exterior pague impuestos muy bajos o nulos, con la
esperanza de atraer más inversiones. El resultado es que prácticamente todas las inversiones que lleguen
se concentrarán en esas actividades especializadas que generan renta, y una carga impositiva razonable no
las habría desalentado. Los incentivos casi no darán origen a nuevas inversiones extranjeras directas y
constituirán una pérdida total para las arcas fiscales.
Hay gran cantidad de posibilidades aceptables. ¿Cuáles son sus límites? Actualmente, en muchos
países en desarrollo competitivos las tasas, establecidas por la lev, de los inpuestos sobre las utilidades
superan apenas el 30% o el 35%3/, según comparaciones de la carga inpositiva real efectuadas recientemente
por Jack Mintz, al parecer incluso en países que, por lo demás, resultan atractivos para los inversionistas, las
tasas efectivas de los impuestos sobre las utilidades no deberían superar esos márgenes. Como se trata de
recaudar algo, las tasasfijadas por ley probablemente no deberían ser inferiores al 10% o el 15%. Países
como Brasil, México y (posiblemente) Rusia pueden aplicar las tasas impositivas más elevadas de esta
escala. Los países mucho más pequeños y/o menos atractivos con potencial como plataforma de exportación
necesitan tasas efectivas que no excedan el 10% o el 15%, o quizás aun menos para los exportadores, si
quieren conseguir grandes inversiones extranjeras directas.
Cada opción dentro de esa escala presenta numerosos pros y contras y compensaciones recíprocas. Si
una tasa efectiva próxima al extremo inferior de la escala se considera ventajosa, se podrá llegar a ella
fijando por ley tasas bajas sin incentivos (como se definió anteriormente) o tasas elevadas junto con
incentivos; determinar cuál de estas opciones es la mejor depende en gran medida de la situación particular
del país, su capacidad administrativa, los efectos en la recaudación, etc.4/
Si el cliente utiliza o está analizando la posibilidad de aplicar incentivos, el FIAS siempre aconseja no
conceder treguas tributarias. Suelen estar mal orientadas (porque premian la constitución de nuevas
compañías en lugar del aumento de los bienes de producción, una mejor tecnología o la capacitación laboral,
por ejemplo). Por lo general, no son eficaces en función de los costos (prácticamente en todos los estudios se
señala que generan muy pocas inversiones adicionales). Inciden más que nada en las inversiones sin
vinculación permanentes, a corto plazo, las cuales pueden irse o crear una nueva compañía una vez
terminada la tregua, mientras que, para las inversiones a largo plazo, otros incentivos (véase el próximo
párrafo) dan mejores resultados.
Los incentivos aconsejables serían combinaciones de disposiciones acertadas sobre traslado de
pérdidas a ejercicios futuros, amortizaciones rápidas u otras cancelaciones, exenciones tributarias o créditos
inpositivos. Se pueden aplicar al gasto que el gobierno desee promover (por ejemplo, inversiones en capital
y

4/

En la publicación KPMG Corporate Tax Rate Survey de 1999, obra del KPMG International Tax a d Legal Center, con sede
n
en los Países Bajos, se informa que las tasas, establecidas por la ley, de los impuestos sobre las utilidades se están reduciendo
en todo el m n o debido a la presión de la competencia. En la encuesta se señala que la tasa media entre los países de la
ud
OCDE descendió casi tres puntos porcentuales e los últimos cuatro años, hasta llegar al 34,8% al 1 de enero de 1 9 . Las
n
99
tasas e los países m n s desarrollados son todavía m s bajas; el promedio en la región de Asia y el Pacífico fue del orden del
n
eo
á
3 % y e América Latina, del 28,5%.
2
n
Véase Mintz y Tsiopoulos (1992), donde se encontrará u buen análisis de las ventajas y los inconvenientes mencionados.
n

Sesión 2 - 223

fijo, capacitación laboral) y por ello se focalizan mejor. Pueden incorporarse al código impositivo y actúan
automáticamente; producen beneficios únicamente después de tomada una medida, no ante simples
promesas. Si bien en muchos estudios se cuestiona la eficacia en función de los costos de algunos o de todos
ellos, estos incentivos son preferibles a las treguas tributarias.
Reducir las tasas de interés establecidas por la ley es una práctica cada vez más difundida y parece ser
al menos tan eficaz como los incentivos mencionados. Las tasas "reducidas" son, en la práctica, del orden
del 20% o menos. La desventaja, naturalmente, es que se aplican a la totalidad del sector empresarial, lo que
puede acarrear una fuerte pérdidafiscal. Sin embargo, en algunos países cuyo sistema consiste en imponer
elevadas tasas impositivas, ofrecer incentivos generalizados y luchar poco o nada contra la corrupción, en
realidad casi nadie paga los impuestos sobre las utilidades y la reducción de tasas puede formar parte de un
plan que incluya medidas más acertadas y honestas para hacer cumplir las obligaciones tributarias (que
quizás no funcionarían con las elevadas tasas y los numerosos incentivos discrecionales que imperaban
anteriormente) y que, en la práctica, aumente los ingresosfiscales. Para muchos de los países que utilizan
nuestros servicios este ejemplo optimista no es más que una esperanza para el futuro, pero al menos indica
un buen camino.
Cualquiera sea el incentivo que el cliente insista en ofrecer, es importante que no lo otorgue
únicamente a los inversionistas extranjeros. Excluir a los inversionistas locales no sólo sería cuestionable
políticamente para cualquier gobierno sino que también daría lugar a la constitución de un gran número de
empresas en común con extranjeros, que, en teoría, reúnen los requisitos legales para hacerse acreedoras a
los incentivos pero que, en la práctica, no serían más que las mismas compañías nacionales a cuya lista de
propietarios se ha añadido algún nombre extranjero. Ello ocurre en especial cuando se conceden treguas
tributarias.
Es conveniente recordar a los clientes que son partidarios de los incentivos que la competencia basada
exclusivamente en bajos costos no es el camino apropiado para alcanzar el desarrollo. Los países necesitan
instruir a los trabajadores y mejorar la salud pública y la infraestructura e incluso si se puede recurrir al
financiamiento privado para atender estas necesidades, los ingresosfiscales se precisarán para cubrir otras.
El comportamiento internacional de la inversión extranjera directa indica bastante claramente que si los
impuestos son razonables y la recaudación se utiliza para suministrar servicios, los inversionistas vendrán y
pagarán los impuestos, y la productividad y los ingresos de toda la economía aumentarán con el tiempo. La
inversión extranjera directa no es más que un medio para lograr estefin, y nuestros clientes son, en última
instancia, los habitantes del país y sus descendientes.
Una de las dificultades con que tropezamos a menudo es el deseo de recibir asesoramiento preciso
manifestado por muchos clientes. La diversidad de casos satisfactorios en todo el mundo indica que no hay
una solución única. Sin haber estudiado exhaustivamente (nosotros u otros especialistas) la situación, las
posibilidades y las preferencias de un país, enunciar a secas, por ejemplo, "El impuesto sobre las utilidades
debería ser del 15%" puede ser algo arbitrario. De ser posible, deberíamos ayudar a los clientes a entender
que cuentan con una variedad de opciones aceptables, en términos generales. Deberíamos explicar qué hace
que una opción sea aceptable y otras no lo sean tanto. Tendríamos que señalar las ventajas y los
inconvenientes de las diversas soluciones viables. Al tiempo que nos concentramos en los efectos para la
inversión extranjera directa, también deberíamos señalar todos los otros criterios, incluso más importantes
para los sistemas tributarios, como el aumento de la recaudación, la facilidad de administración, la reducción
de las posibilidades de corrupción y fraude, etc. No olvidemos que el punto de vista de la inversión
extranjera directa es sólo una de las tantas voces que deberían escucharse para diseñar un buen sistema
tributario. Y, por supuesto, deberíamos hacer todo lo que esté a nuestro alcance para entender la situación del
país en cuestión y, en la medida de lo posible, adecuar nuestro asesoramiento a ella.
Otros impuestos. Hasta ahora el análisis ha girado en torno al impuesto sobre las utilidades. A
continuación presentaremos someramente otros impuestos pertinentes:

Sesión 2 - 224

Derechos de importación de maquinaria y equipo: Estos impuestos pueden tener importancia si son
considerables, puesto que representan un aumento de los costos para el inversionista, mientras que su
eliminación los reduce. Es difícil entender la conveniencia de que un país imponga elevados derechos de
importación de maquinaria y equipo, independientemente de que los importadores sean empresas con
inversión extranjera o nuevas compañías que instalan capacidad o antiguas compañías que quieren ampliar o
actualizar sus equipos. Si bien no está referido estrictamente a la manera de atraer inversión extranjera
directa, nuestro consejo para la mayoría de los países, si no todos, es que los impuestos sobre la importación
de maquinaria y equipo probablemente no deberían superar los niveles de "ingreso" de aproximadamente el
5% al 10% como máximo, y que todos, excepto los exportadores, deberían pagarlos. Si en el país existe el
impuesto al valor agregado (IVA), que se aplica tanto a las importaciones como a la producción nacional
(como debería ser), entonces los derechos de importación que superen el IVA sobre maquinaria y equipo
podría, y probablemente debería, ser nulo. Cualquiera sea su nivel, esos derechos y el IVA deberían
reintegrarse (o no cobrarse) a los exportadores. (Naturalmente, se presenta el problema del reintegro de
impuestos sobre bienes de capital a los exportadores indirectos. Nada es perfecto.) En la mayor parte de los
países que recurren a nuestros servicios, la mayoría de los inversionistas no paga, en la práctica, impuesto
alguno a la importación de maquinaria pero, en muchos de ellos, hay que recurrir a halagos o sobornos. En
este tema, como en otros, el consejo de adoptar un sistema simple y automático es, a menudo, uno en los que
más se debe insistir.
IVA o impuestos a los insumos que graven materias primas y bienes intermedios; rebajas para los
exportadores: en estos casos, el consejo estándar es que deben ofrecerse a los exportadores condiciones
semejantes al libre comercio. El tipo de sistema puede ser muy importante, desde el punto de vista
administrativo; los sistemas de reintegro de derechos requieren una administración competente y orientada al
cliente que, en muchos países, no existe. Los países que utilizan nuestros servicios no deben dejar de lado
ningún sistema práctico satisfactorio para los exportadores por aplicar algún otro que es teóricamente óptimo
para un país desarrollado, con capital humano, instituciones y procedimientos tributarios avanzados.
Impuestos y otras cargas sobre los costos salariales: suelen ser altos y varían notablemente de un país
a otro. Cuando son muy elevados, pueden constituir un serio obstáculo para la competitividad de las
exportaciones, a menos que se los traslade por completo a los empleados. En las economías más pequeñas y
orientadas a la exportación, habrá inversión extranjera directa cuando las condiciones sean favorables a la
exportación; para que lleguen al país cuantiosas inversiones de este tipo debemos recomendar sistemas
impositivos viables que faciliten la exportación5'.
También deben analizarse las contribuciones territoriales, los impuestos sobre los activos fijos, los
impuestos sobre la repatriación de utilidades u otros ingresos, etc., pues pueden desalentar la inversión
extranjera directa.
Los beneficios automáticos en comparación con los beneficios discrecionales. La palabra utilizada
normalmente en este trabajo es "incentivoshemos empleado "beneficios" para incluir no sólo los
incentivos impositivos en el sentido más estricto usado anteriormente sino también otros como la reducción
de tasas establecidas por la ley, las subvenciones iniciales, los servicios de contrapartida para los socios y
cualquier otro instrumento que una persona encargada de promoción de inversiones quisiera tener
autorización para otorgar o considerara necesaria para persuadir a un inversionista de efectuar una inversión.
La mayoría de los promotores de inversión extranjera directa quisieran tener atribuciones para otorgar
incentivos u otros beneficios. Sólo dos motivos justifican estas atribuciones, incluso en principio: a) impedir
que los incentivos "se desperdicien" en inversiones que se efectuarían de todos modos, y b) concentrarlos en
y

Los impuestos e incentivos vinculados a los costos de maquinaría o m n de obra pueden provocar sesgos. Los altos
ao
impuestos sobre los salarios pueden inclinar la industria hacia métodos de uso intensivo de capital; lo ms o puede suceder,
im
por ejemplo, con la rapidez de las amortizaciones o los créditos impositivos por inversiones. En las publicaciones sobre el
t m no se señalan efectos apreciables en las proporciones de los factores de estos cambios en los precios relativos; no
e a
revisten importancia e la práctica. No h n sido motivo de gran preocupación en ningún trabajo de asesoría sobre incentivos a
n
a
la inversión extranjera directa en que haya intervenido el FIAS.
Sesión 2 - 225

actividades de especial interés, por ejemplo, poner en marcha una política industrial. El primer motivo no es
estrictamente verdadero; pocos promotores de inversiones tienen los conocimientos especializados o la
integridad y "tacañería" necesarias para aplicarlos si los tuvieran. En el caso de sectores donde los beneficios
son elevados y es obvio que los incentivos son innecesarios, la ley o las normas pueden excluir a la inversión
extranjera directa de los incentivos que se ofrecen en general. El segundo motivo, la aplicación de una
política industrial, puede tener más aplicación práctica. Sin embargo, escapa en parte al objetivo de este
trabajo, pues no se relaciona con la forma de atraer inversión extranjera directa. En realidad, en muchos de
los países que solicitan nuestros servicios, la principal razón por la cual los clientes desean conservar las
atribuciones mencionadas es tener la oportunidad de recibir sobornos.
Las inversiones cuantiosas e interesantes reciben, al parecer, algún tipo de aliciente discrecional
prácticamente en todos los lugares del mundo. Si no lo hace el gobierno nacional (y en ocasiones incluso si
lo hace), las autoridades provinciales y/o municipales pelean por atraer las grandes plantas de automotores o
máquinas herramientas o circuitos integrados para computadoras. Cuando un equipo de Intel llega a un país,
no pasa desapercibido6'. Sin negar esta realidad, podemos recomendar con seguridad un conjunto de a) tasas
inpositivas y normas sobre amortización racionales, dentro de un sistema tributario estable y mínimamente
honesto; b) beneficios automáticos que operen a través del sistema tributario, si alguna otra medida se
considera necesaria; y c) inevitablemente, algún tipo de tasas impositivas muy bajas si se busca que las
plataformas de exportación atraigan gran cantidad de inversión extranjera directa.
Según la experiencia del FIAS, para que las autoridades atiendan sus consejos, el asesor normalmente
debe presentar al menos algún análisis cuantitativo que compare la presión tributaria del país con la de sus
vecinos u otros supuestos competidores. Hasta el gobierno ruso, a comienzos de los años noventa, pidió un
análisis de esta índole, aun cuando el país no competía por obtener inversiones que actuaran como
plataforma para la exportación. La amplia experiencia adquirida nos indica que, en muchos casos, los
argumentos basados exclusivamente en análisis cualitativos se rechazan o ni siquiera se estudian con
seriedad. Probablemente habría que suministrar cifras diferenciadas en función de los sectores y mercados
pertinentes, y también analizar las distintas opciones en materia de políticas con que cuenta el país que
solicita nuestros servicios.
Los beneficios mínimos frente a los beneficios considerables. En la práctica suele haber una
inportante relación de correspondencia entre los diferentes efectos que producen estos dos tipos de
beneficios: los primeros ocasionan menos distorsiones y menos pérdidas en quienes se aprovechan de la
situación sin asumir carga alguna, pero probablemente no alteren ninguna decisión (o no muchas); en un
caso extremo, pueden producir pocos de los efectos buscados, o ninguno, y pueden tener un costo bajo, pero
no nulo. Los beneficios considerables pueden provocar mayores distorsiones (si bien la distorsión quizás sea
exactamente lo que la política busca) y acarrear mayores gastos en quienes se benefician sin contrapartida,
pero alcanzan en gran medida sus objetivos. Ello plantea un dilema para el asesor sobre políticas. Es obvio
que no debe recomendar un sistema complejo de beneficios mínimos, pero si el cliente está estudiando la
posibilidad de aplicar un plan más sencillo y sólido que produzca verdaderamente los resultados que desea,
requerirá una evaluación detenida que permita analizar y sopesar las ventajas y los inconvenientes de las
diferentes opciones, las cuales, probablemente, incluirán resultados inprevisibles ("manzanas y naranjas")
sobre los cuales corresponde al cliente decidir. Para citar un ejemplo extremo (pero real, que le tocó vivir al
FIAS): si un cliente desea diseñar un conjunto de medidas para sustituir la tierra y la mano de obra
actualmente destinadas al cultivo de coca por otras actividades agrícolas, de poco sirve aconsejar no hacerlo
por las distorsiones que provocarían los inportantes beneficios que habría que ofrecer para cumplir este
objetivo.

&

Véase en Spar (1998) u informe acerca del plan empleado por Costa Rica para atraer una planta de Intel sin infringir las
n
normas ni regalar el negocio.

Sesión 2 - 226

Bibliografía
Bergsman, Joel and Wayne Edisis (October 1988), Debt-Equity Swaps and Foreign Investment in Latin
America, IFC Discussion Paper No. 2.
Bond, E.W. and L. Samuelson (1986), "Tax Holidays as Signáis," American Economic Review, 76:4.
Ernst and Young (1994), Investment in Emerging Markets: A Survey of the Strategic Investments of
Global 1000 CompaniesL
Guisinger, S. & Associates (1985). Investment Incentives and Performance Requirements. New York:
Praeger.
Lecraw, D. J. (1991), "Factors Influencing Foreign Direct Investment by Transnational Corporations in
Host Developing Countries," in Buckley and Clegg (eds.), Multinational Enterprises in Less
Developed Countries, St. Martin's.
Lim, D. (1983), "Fiscal Incentives and Direct Foreign Investment in Less Developed Countries," Journal
of Development Studies, 19:2.
Michalet, Charles-Albert (1997), Strategies of Multinationals and Competition for Foreign Direct
Investment, FIAS Occasional Paper No. 10.
Mintz, Jack M. and Thomas Tsiopoulos (1992), Corporate Income Taxation and Foreign Direct
Investment in Central and Eastem Europe, FIAS Occasional Paper No. 4.
Organization of American States (1990), The Impact ofTourism Incentives in the Caribbean Región.
Pirnia, Neda (1996), 'The Role of Investment Incentives in Influencing Investors' Locational Choices: A
Literature Survey," FIAS (processed).
Reuber, G., H. Crookell, M. Emerson and G Gallais-Hamonno (1973). Prívate Foreign Investment in
Development. Oxford, UK: Clarendon Press.
Robinson, HJ. (1961). The Motivation and Flow of Prívate Foreign Investment. Menlo Park, California:
Stanford Research Institute.
Rolfe, R.J., D. Ricks, M. Pointer and M. McCarthy (1993). "Determinants of FDI incentive preferences of
MNEs," Journal of International Business Studies, 24(2), pp. 335-55.
Shah, Anwar (ed.; 1992) Fiscal Incentives for Investment in Developing Countries, World Bank.
Spar, Debra (April 1998), Attracting High Technology Investment: Intel's Costa Rican Plant, FIAS
Occasional Paper No 11.
United Nations Center on Transnational Corporations, (1992). The Determinants of Foreign Direct
Investment: A Survey ofEvidence. New York: United Nations.
United Nations Conference on Trade and Development, (1996).
Investment Geneva.

Incentives and Foreign Direct

Wells, Louis T. Jr. (autumn 1986), "Investment Incentives: An Unnecessary Debate," The CTC Repórter,
United Nations Center on Transnational Corporations.

Sesión 2 - 227

Sesión 2 - 228

SESION 3

ADMINISTRACION DE RIESGOS Y MANEJO
DE PASIVOS CONTINGENTES DEL SECTOR
PUBLICO: ¿ COMO INCORPORAR EL TEMA
DE LAS GARANTIAS GUBERNAMENTALES
-EXPLICITAS E IMPLICITAS-

A LA

PROGRAMACION PRESUPUESTARIA
ANUAL Y PLURIANUAL?

Sesión

- 229

Sesión 3 - 230

GOVERNMENT CONTINGENT LIABILITIES:
A HIDDEN RISK TO FISCAL STABILITY

by
Hana Polackova
The World Bank

The findings, interpretations, and conclusions expressed in this volume are those of the author and should
not be attributed to the World Bank, affiliated organizations, or members of its Board of Directors or the
countries they represent. The author thanks Hafez Ghanem, Sanjay Pradhan, Alien Schick, and Sergei
Shatalov for their helpful comments and suggestions. The paper is a short versión of a forthcoming, more
comprehensive study that further elaborates the policy and institutional aspects of fiscal risks and
includes several country case studies of direct and contingent fiscal risks and the quality of fiscal
adjustment.

Sesión 3 - 2 3 1

Table of Contents

I.

Introduction

233

II.

Possíble Financing Pressures on the Central Government

234

A. The Fiscal Risk Matrix

234

B. Direct Explicit Liabilities......

...

....

......236

C. Direct Implicit Liabilities

237

D. Contingent Explicit Liabilities
E.

Contingent Implicit Liabilities

...
...

.....
....

.....238
....

....239

III. The Increasing Problem of Fiscal Risks

241

A. Trend, Bias, and Moral Hazard.........................................................—............................241
B. Fiscal Risks and the Challenge of Transition and Emerging Markets
C. Hidden Fiscal Risks and the Valué of Transparency and Certainty
D. Fiscal Opportunism Under Debt and Déficit Ceilings

242
......242
243

IV. Understanding, Incentives, and Capacities to Reduce and Control Fiscal Risks
A. Systemic Measures to Reduce Fiscal Risks.

247

B. Before Government Admits an Obligation.......................................................................251
C. When the Government Accepts and Holds an Obligation —
D. After a Liability Falls Due
V.

253
254

Conclusions

255

VI. References

256

VIL Annex

257

Sesión 3 - 232

246

Summary
i.

Governments are facing increasingfiscal risksand uncertainties. Most recently, the Asian crisis
revealed how contingent forms of govemment support create major moral hazard in the markets
and severefiscal instabilities. The main reasons for this situation are: (a) the increasing volumes
and volatility of international prívate capitalflows, (b) a shift in the role of the state from directly
financing and providing services to guaranteeing that the prívate sector will accomplish particular
outcomes, (c) biases in policy decisionmaking under fiscal constraints, and (d) the existence of
moral hazard in the markets because of a perception that the govemment has some residual
responsibilities.

ii.

Since off-budget govemment programs increasingly cause fiscal instabilities, a string of years
with a balanced budget and low public debtfigures does not necessarily suggest either past fiscal
prudence or a goodfiscal outlook. Thus, a study of govemmentfiscal position must also examine
the obligations taken by the govemment outside the budgetary system. The fiscal risks
governments face are of four types: either direct or contingent, both of which in tum are either
explicit or implicit. Governments that want to avoid the danger of sudden fiscal instability and
accomplish their long-term policy objectives must have a good understanding of both their direct
and contingent liabilities and must be able to handle them appropriately. But do governments
have adequate incentives and capacities to deal with not only budgetary but also with all
significant fiscal risks? Assistance and even coercion by international institutions may be
required.

iii.

There are four main ways governments can addressfiscal risks:(a) control the contingent as well
as direct, and implicit as well as explicitfiscal risksand orient policies toward good quality rather
than rapidfiscal adjustment; (b) publicly recognize the limits of the state's responsibilities so as
to deter moral hazard in the markets; (c) ensure that institutional arrangements for public finance
and standards for budgeting, accounting,financial planning, reporting, and auditing address both
contingent and direct liabilities and promotefiscal prudence and equity in all contingent as well
as directlyfinanced public programs; and (d) develop and employ institutional capacities to
evalúate, regúlate, control, and preventfinancial risksin both the public and prívate sectors.

I.

Introduction

Governments are facing increasing fiscal risks and uncertainties. Two of the reasons for this
situation are:first, the international integration of financial markets, which has meant greater volumes
and volatility of cross-border flows of prívate capital; and, second, the privatization of state functions,
accompanied by implicit or explicit state guarantees. State guarantees and insurance schemes, as opposed
to budgetary subsidies and direct provisión and financing of public services, have become a common
method of govemment support. These off-budget programs and obligations involve hidden fiscal costs,
with implicit and contingent liabilities that may result into excessive requirements for publicfinancing in
the médium and long term.
The third reason is that policymakers pursuing a balanced budget or some déficit target tend to
favor off-budget forms of state support that do not require immediate cash and that, at least for some time,
hide the underlyingfiscal cost. Fiscal adjustment that concentrates on déficit reduction may overlook or

Sesión 3 - 233

ele vate thefiscal risksassociated with structural policies affecting pensions and health care. Major fiscal
risks outside the budget derive from explicit promises and implicit expectations that govemment will help
when various failures occur. The subsequent emergence of moral hazard in the markets can exacerbate
these risks, Usually the support govemments offer to large weak banks, enterprises, and subnational
govemments in troubles is outside the budget.
Although it is impossible for govemments in a market environment to avoid allfiscal risk,they
can control and reduce the risks, but only if they recognize and fully consider them in their policy
debates. Whether govemments have the incentives and capacities to reflect fiscal risksin their policy
choices and to carry out appropriate fiscal adjustment is an important question. The incentives will
reflect how well policymakers understand the issues and the pressures they face in dealing with them.
Thefiscal risksbecome apparent only when the institutions conducting fiscal analyses look beyond the
government's budget and debt to include the contingent and implicit liabilities. The extent of the
incentives govemments have to make direct and contingentfiscal risks transparent is linked mainly to the
definition and measurement of internationally recognized fiscal indicators, to the quality of public
awareness, external monitoring, and to the sanctions imposed for concealing relevant data and exposing
the state to excessivefiscal risk.
This paperfirst classifies and analyzes the potential obligations andfiscal risksgovemments face
and their sources. It next outlines the options for reducingfiscal risks in the context of fiscal adjustment,
with particular attention to the typology and analysis of specific fiscal risks,the high risk exposure of
govemments of transition and emerging-market economies, and the quality and bias in govemment
decisionmaking at the time of fiscal adjustment. Several questions are addressed. How can policymakers
be made accountable for recognizing the long-term cost of all forms of govemment activities? How can
the moral hazard induced by govemment interventions be reduced? What standards for public sector
accounting, budgeting, reporting, andriskmanagement would foster soundfiscal performance in the long
term? A forthcoming, more comprehensive study presents a framework for including direct and
contingentfiscal risksin fiscal analysis and applies it to selected countries.

II.

Possible Financing Pressures on the Central Government
A.

The Fiscal Risk Matrix

Govemments face four types of fiscal risk, each of which is a broadly defined liability that
combines two of the following four characteristics: explicit versus implicit and direct versus contingent.1
•

Explicit liabilities are specific obligations of the govemment established by a particular law or
contract. The govemment is legally mandated to settle the obligation when it comes due. Common
examples are the repayment of sovereign debt and repayment of nonperforming loans the state has
guaranteed.

•

Implicit liabilities involve a moral obligation or expected responsibility of the govemment that is not
established by law or contract but instead is based on public expectations, political pressures, and the
overall role of the state as society understands it. Examples of implicit liabilities are future public
The international accounting standards for govemments proposed by the International Federation of Accountants define a
liability as a present obligation of t e govemment that entails a fonn of economic benefits and that arisesfrom past events
h
whose settlement is expected to result in an outflow of g v m e t resources (International Federation of Accountants 1998).
oe mn
Sesión 3 - 234

pensión benefits that are not specified by law, disaster relief for uninsured victims, and default of a
large bank on nonguaranteed obligations.
•

Direct liabilities are obligations that will arise in any event and are therefore certain. They are
predictable based on some specific underlying factors; they do not depend (are not contingent) on any
discrete event. For example, future public pensions specified by law are a direct liability whose size
reflects the expected amount of the benefit, eligibility factors, and future demographic and economic
developments.

•

Contingent liabilities are obligations triggered by a discrete event that may or may not occur.2 The
probability of the contingency occurring and the magnitude of the government outlay required to
settle the ensuing obligation are difficult to forecast. Probability and magnitude depend on some
exogenous conditions, such as the occurrence of a particular event (for example, a natural disaster or
banking crisis) and some endogenous conditions, such as the design of government programs (an
example being the contracts for state guarantees and insurance), as well as on the quality and
enforcement of regulations and supervisión. Thefiscal riskmatrix in table 1 provides a typology of
the sources of the potentialfinancial requirements central governments face. Under each category are
examples of government programs and promises that can create fiscal pressures. Some of the
examples apply across all countries (such as sovereign debt), whereas others are more countryspecific (such as crop insurance).

International accounting standards define a contingency as a condition or situation whose ultímate outcome is determined
only by the occurrence, or nonoccurrence, of one or more future events (International Accounting Standards Committee
1997).

Sesión 3 - 235

THE
Liabilities

8

Contingent

Direct

(obligation

FISCAL

TABLE 1
RISK MATRIX

in any event)

Explicit
Government
liability
is
recognized by
law or contract

Foreign and domestic
sovereign
borrowing
(loans contracted and
securities issued by
the
central
government)
Expenditures
by
budget law
Budget expenditures
legally binding in the
long term (civil service
salaries, civil service
pensions)

Implicit
A
"moral"
obligation
of
the
government
that
mainly
reflects public
expectations
and pressures
by
interest
groups

Future recurrent costs
of public investment
projects
Future public pensions
(as opposed to civil
service pensions) if
not required by law
Social
security
schemes
if
not
required by law
Future health care
financing
if
not
specified by law

(obligation

if

a particular

event

occurs)

State guarantees for nonsovereign borrowing and
obligations issued to subnational governments and
public and private sector entities (development banks)
Umbrella state guarantees for various types of loans
(such as for mortgages, students studying agriculture,
and small businesses)
State guarantees (for trade and the exchange rate,
borrowing by a foreign sovereign state, private
investments)
State insurance schemes (for deposits, minimum
returns from private pensión funds, crops, floods, war
risk)
Default of a subnational government and public or
private entity on nonguaranteed debt and other
liabilities
Cleanup of the liabilities of privatized entities
Bank failure (beyond state insurance)
Investment failure of a nonguaranteed pensión fund,
employment fund, or social security fund (social
protection of small investors)
Default of the central bank on its obligations (foreign
exchange contracts, currency defense, balance of
payments stability)
Bailouts following a reversal in private capital flows
Residual environmental damage, disaster relief,
military financing, and the like

Of fiscal authorities, not the central bank.

B.

Direct Explicit Liabilities

In most countries, the central government commonly recognizes, quantifies, and discloses direct
explicit liabilities. Even so, estimating the size of the government outlays related to these obligations in
the médium term i s not a simple task.
•

Governments usually specify obligations to settle direct foreign and domestic sovereign debt in their
loan contracts and securities. The future financing requirements mainly relate to the maturity,
currencies, and interest rate of the debt instruments. Using these specifications, governments forecast
their debt service profile, simúlate the tradeoff between risk exposure and the cost of borrowing, and
build debt service scenarios for alternative portfolio and macroeconomic developments. Denmark,
Ireland, and the United Kingdom provide excellent examples of how to analyze and disclose
sovereign borrowing risks.

•

Budgetary outlays are normally embedded in an annual budget law, which contains the approved
activities and policies of the government. In principie, the budget is legally binding, and outlays are

Sesión 3 - 236

to comply with the budgeted figures throughout thefiscal year. In practice, the budget is viable only
if it is based on good macroeconomic analysis and if the government employs institutional
mechanisms for fiscal discipline and control.
•

Governments in many countries extend legal entitlements to a salary and pensión at a specified
retirement age to public employees. It is certain these legal entitlements will be a spending item in
future state budgets. Their magnitude is based on forecasts of the numbers of public employees and
their expected remuneration, pensión benefit, and retirement age. Under New Zealand's Fiscal
Responsibility Act, the government is required to analyze and disclose such forecasts in budget
documents. (Where a government plans to downsize the civil service, it may be obligated to pay
redundancy packages, whose overall cost would be contingent on the actual downsizing.)

In contrast to the environment depicted in thefiscal riskmatrix (table 1), countries with legal
provisions that the governmentfinance future social security benefits, such as public pensions, universal
health care, and education, list these as direct explicit rather than implicit items.

C.

Direct Implicit Liabilities

Direct implicit government liabilities often arise as presumed, rather than as legal or contractual,
obligations established by public expenditure policies in the médium term. Only governments that are
committed to transparent medium-term expenditure planning and long-term fiscal discipline recognize
and quantify these obligations. Assuming no policy changes, the implicit cost of demographically driven
public expenditures is what in particular poses a danger to fiscal stability in the long term.
•

The completion of public investment projects and maintenance are only expected, not mandated, by
law. Governments analyze and quantify, and are accountable for, the ex-ante estimates and actual
multiyear investment and ensuing long-term recurrent costs. Countries such as Australia and South
Africa use a medium-term expenditure framework that automatically includes the financing
requirements for operations and maintenance in the fiscal outlook and future budgets. Thus,
government obligations to sustain the benefits of public investments are explicit.

•

In many countries, future public pensión benefits are not grounded in any legal document and
therefore are not explicit but rather are implicit government liabilities. Assuming that a given
pensión policy will continué, it is certain that the overall obligation of the government will occur
(there are economic, social, and political reasons for assuming that a government would not stop
paying the benefits unless itfirst reformed its pensión system). Henee future public pensión benefits
are a direct liability, even though not in the strict accounting sense. Since the provisión of public
pensions is recognized as the most striking problem for fiscal sustainability in aging societies, many
governments have been analyzing the long-termfiscal implications of their pensión policies and of
alternative reforms using long-term fiscal and pensión models. (A point discussed below is that
pensión reforms often encourage prívate sector involvement in saving for retirement because the
government provides indirect forms of support, such as guarantees of mínimum pensión benefits.
These guarantees are an explicit contingent liability of the government.)

•

Similarly, future health care and social securityfinancing can be analyzed as a direct implicit
government liability (even if not accounted as such). Research shows that the dynamics of the
financing requirement for health care in an aging society is often even more explosive than that for
public pensions. Modeling and recognition of the long-term fiscal implications of health care policies
and their reforms are critically important for fiscal stability and equity in the long term.

Sesión 3 - 237

D.

Contingent Explicit Liabilities

Contingent explicit liabilities are a legal obligation of the govemment to make a payment if a
particular event occurs; they are not directly associated with any existing budgetary program. A
govemment's commitment to accept obligations contingent on future events amounts to a hidden subsidy
and may cause immediate distortions in the markets and result in a major unexpected drain on
govemmentfinances in the future.
Although governments recognize each contingent explicit liability in some formal documentation,
many have yet to consolidate all these obligations and their total magnitude in one place and to include
them in their overall fiscal analysis and expenditure planning. In contrast, many corporations,
commercial banks, and insurance companies have made considerable progress in dealing with contingent
liabilities in the past 10 years. Similarly, governments have not yet recognized the importance of good
design, monitoring, and management of their programs to controlfiscal risks. At the policy level, ex-ante
analysis of the risks and future financial implications associated with the contingent forms of govemment
support contribute to better policy choices on equity and long-termfiscal stability.
•

Governments often issue guarantees to cover part or all of theriskthat a borrower will fail to repay a
loan or other guaranteed asset or that an institution will fail to fulfill its obligations. Common
examples include state guarantees of debt and other obligations of subnational governments and
various public and prívate entities, such as budgetary institutions, credit and guarantee funds,
development banks, and enterprises. Guarantees and credit issued through a state-guaranteed
intermediary are particularlyriskybecause they allow the govemment to pursue unannounced policy
decisions, involve a problem of management incentives, and are difficult for governments to monitor
and control. The hidden subsidy to the beneficiary of a guarantee, and the subsequent potential cost
to the govemment, are positively correlated with therisk,size, and duration of the underlying asset.
In addition, the probability of a default may be very high if the guarantee contract does not specify
risk-sharing by both the govemment and the other parties in terms of both thefinancial coverage (part
versus all of the loan) and risk coverage (specific political or commercial, versus all risks).
Government guarantees routinely cover allrisksfully. Such guarantees distort the markets and are
called with high probability. The risk a govemment assumes can be estimated based on the
experience of governments of different capacities, simple rules, and, where appropriate, more
sophisticated methodologies such as actuarial, econometric, loss estimate, and option pricing models.3
Assessment of risksallows governments to reflect the potentialfiscal cost associated with guarantees
in their choices of policies and forms of support and in the design of a guarantee contract. Since
passage of the Credit Reform Act, the United States provides good examples of govemment analysis
and the design of credit guarantees.

•

Governments extend umbrella guarantees to eligible persons or entities borrowing for a specific
purpose, such as university studies, a mortgage, farming, and small business development. The
rationale for these guarantees and the assessment of their risks and potential long-term cost are
similar to those for the individual guarantees discussed above (and are also true for trade and
exchange rate guarantees and guarantees on foreign sovereign borrowing and prívate investments).

•

State insurance schemes often constitute a major risk to future fiscal balances. Common state
insurance programs cover bank deposits, crops, warrisks,minimum returns from pensión funds, and
floods, earthquakes, and other natural disasters. Although most of these programs cover losses that
occur very infrequently, when the losses do occur, their total magnitude may be enormous. The risk
For a detailed discussion of the valuation methodologies for loan guarantees and other contingent liabilities, see Mody and
Patro (1996) and Mody and Lewis (1997).
Sesión 3 - 238

pool under these programs, particularly in small markets, is very limited, one justification for
government's involvement. State insurance schemes rely on net govemmentfinancing from general
taxes, rather than on insurance fees, and thus redistribute wealth. The analysis of risksand potential
fiscal burdens associated with state insurance schemes requires sector data and sophisticated models
(such as the hydrologic model used to estimate the probabilities of floods in a given year), and loss
estimation methodologies and options pricing models to assess the riskiness of the returns of a
pensión fiind. A qualitative analysis of the risk factors is, however, sufficient for the govemment
both to design a sound insurance scheme that would not seriously distort market behaviors and to
make a rough estimate of its potentialfiscal cost. The United States may take the lead in this area as
its govemment adopts the analytical and budgeting method for federal insurance programs proposed
by the General Accounting Office (United States, General Accounting Office 1997).

E.

Contingent Implicit Liabilities

Contingent implicit liabilities are not officially recognized and may be the product of declared
policy objectives. Govemments accept these liabilities only after a failure in the public sector or market
and as a result of pressure by the public, possibly interest groups, or just too high an opportunity cost for
not acting.
Contingent implicit liabilities often pose the greatest fiscal risk to govemments. The event
triggering the liability is uncertain, the valué at risk difficult to evalúate, and the extent of govemment
involvement difficult to predict. In short, it is very hard to identify and estimate the size of contingent
implicit liabilities. They are particularly large if the macroeconomic framework in the country is weak,
the financial sector vulnerable, regulatory and supervisory systems inefficient, and disclosure of
information in the markets limited.
In addition, expectations of govemment involvement generate moral hazard in the markets. The
scope for moral hazard is particularly large in economies in which the govemment significantly
minimizes the pain of past failures of market agents and in which the govemment and investors do not
have a good capacity to monitor the risk exposure of market agents. Govemments can constrain moral
hazard if, in advance and through their actions, they decide on and signal the limits of their potential
intervention to the markets. To reduce rather than expand the moral hazard, the signáis have to make
clear that the govemment will be responsible for minimum public goods only and that there will be
significant pain for agents that fail. The govemment needs to assess the costs and benefits and reveal its
responsibility for each contingent implicit liability separately.
•

In most countries, the financial system is government's most serious contingent implicit liability.
International experience indicates that the markets expect the govemment to helpfinancially if the
stability of thefinancial system is at risk. In case of a failure in thefinancial sector, govemments are
compelled to intervenefinancially far beyond their legal obligation either to secure some critical
functions of thefinancial system or to protect depositors and specific market agents beyond the limits

Sesión 3 - 239

of state insurance schemes.4 Such practice further exacerbates the moral hazard problem in the
financial and corporate sectors.
•

Uncovered losses and defaults on nonguaranteed debt and obligations by a subnational government,
state-owned or large private enterprise, budgetary or extrabudgetary agency, or any other institution
of political significance may induce the government to provide financing. Governments also accept
various obligations of parastatal and public entities subject to privatization. (The government is often
liable for the obligations, including contingent explicit ones, of state-ownedfinancial institutions.)
The contingent implicit government liability associated with both the financial system and
nonguaranteed corporate debt increases with rising amounts and less efficient allocation of private
capital.

•

Depending on social preferences, some critical social and welfare functions, even when the
government has contracted them out, are believed to be the ultímate responsibility of government.
For example, where there is an investment failure of a nonguaranteed pensión, employment, or social
security fund, the government has to finance the social services from the budget. Thus,
nonguaranteed private provisión of social and welfare services poses a contingent implicit financial
risk for governments (Heller 1997).

•

The fiscal authorities also have ultímate responsibility for currency stability and the balance of
payments, and therefore also for unmet obligations of the central bank. Most recently, Thailand and
other Asian countries have shown thatfiscal obligations may arise from afixed exchange rate regime
or the foreign exchange contracts of the central bank and, ultimately, are connected to international
bailout packages. As in most of the instances listed above, theriskescalates with the macroeconomic
vulnerability and moral hazard in the markets.

•

Private capital flows entail contingent implicitrisksto the government in four areas: (a) policy (the
risks include exchange rate overvaluation and sterilization), (b) domestic assets valuation (an asset
bubble for real estáte, productive and intangible assets,financial instruments, and domestic currency),
(c) intermediation (interest rate differential and weakness of the domesticfinancial system), and (d)
borrowing (incentives for corporations, banks, and governments to overborrow and borrow shortterm).

•

Ultimately, environmental damage and natural disasters create a high demand for public monies,
beyond explicit state insurance programs and guarantees. Many countries face thefinancial riskof
operating and dismantling nuclear plants, disposing of nuclear and toxic wastes, and the residual cost
of environmental recovery. In the absence of developed private insurance industries, particularly in
countries with a history of caretaking, disasters such as floods, earthquakes, and droughts create
major political pressures for government action.

4

Standard & Poor's estimated the contingentfiscal cost of the domestic banking system in 1 9 (the level of direct a d indirect
97
n
costs to the government under t e worst-case scenario) in Argentina, Hungary, Italy, Poland, and Sweden to be less t a 1
h
hn 0
percent of gross domestic product (GDP), in Greece, the Philippines, Singapore, the Slovak Republic, the United Kingdom,
and the United States to be about 1 to 20 percent, and in China, the Czech Republic, Hong Kong (China), Japan, the
0
Republic of Korea, Malaysia, Thailand, and Taiwan (China) to be over 3 percent (Standard & Poor's, 1997).
0
Sesión 3 - 240

III.

The Increasing Problem of Fiscal Risks
A.

Trend, Bias, and Moral Hazard

As noted, recent trends show that governments are exposed to expanding fiscal risks. First, the
high volumes and volatility of prívate capital flows and increasing economic dependence of countries on
foreign capital have exacerbated the vulnerability of their domesticfinancial and corporate sectors and,
implicitly, of the government. Particularly in transition and emerging market economies, domestic assets
are subject to the rapídly changing preferences of foreign investors. An interplay of risks—policy(high
exchange and interest rates), valuation of domestic assets (asset bubble), intermediation (interest rate
differential and weak domestic financial system), and borrowing (overborrowing and short-term
borrowing by government and market institutions)—may lead to a sudden dumping of domestic equity,
bonds, and currency by investors. Such instances, which occurred recently in México, Asia, and Russia,
cause a crisis that is solved partly at taxpayers' expense.
Second, states have been transforming their role, moving from directly providing and financing
services to guaranteeing that the prívate sector will accomplish certain outcomes. Privatization of state
responsibilities and attempts to encourage prívate sector initiative through explicit or implicit government
guarantees have left governments with increasing uncertainty about future publicfinancing requirements.
Will a guarantee be called? What will the outlays of state insurance programs be? Will reserve funds be
sufficient to cover the contingent losses? Several governments have incurred expenditures above
envisaged limits following a massive failure of projects covered by state guarantees, busted deposit
insurance schemes, banking crises, and excessive prívate credit. In such instances, many transition and
emerging-market countries have faced capital flight and plunged into fiscal crisis, whereas developed
countries such as France absorbed thefiscal shock by issuing more public debt.
Third, governments may be biased toward off-budget policies, which pose morefinancial riskbut
require less immediate financing. Often, particularly in times of déficit reduction and a short-term
political horizon, policymakers exploit the fact that off-budget commitments and obligations are not
necessarily reported so that the cost of government policies is hidden. Under these conditions,
decisionmakers favor off-budget forms of government support such as state guarantees, direct credit, and
absorption of prívate liabilities and bad assets. There are many examples of fiscal opportunism, for
example, in countries in the European Monetary Union's (EMU) subject to the fiscal ceilings of
Maastricht as well as World Bank and International Monetary Fund (IMF) adjustment programs.5 Such
forms of government support give rise to government contingent fiscal risks whose costs and cash
consequences perhaps will not be seen for many years.
Finally, explicit state guarantees and insurance schemes, or any implicit understanding that a
government will come to the rescue in the case of various market failures, generate serious moral hazard
problems in the markets. Loans and investments with a full guarantee suffer from insufficient analysis
and supervisión by creditors. Beneficiaries of poorly designed state insurance schemes tend to expose
themselves to excessive risks. For instance, in the United States, the generous benefits of the federal
flood insurance program have resulted in excessive construction of houses inflood-prone areas (United
The 1 9 Treaty of Maastricht sets t e following fiscal limits: general government déficit as a net borrowing requirement, 3
92
h
percent of GDP and total gross debt at nominal valué outstanding at the e d of t e year and Consolidated within general
n
h
government, 60 percent of GDP. Both déficit a d debt are calculated according to t e European System of National Accounts
n
h
ESA78. ESA78 only roughly defines t e general government a d does not require the recording of government transactions
h
n
on a accrual basis and of assets at m r e valué. For a description of the opportunisticfiscal behaviors of governments under
n
ak t
fiscal constraints see Easterly (1998) and Forte (1997).
Sesión 3 - 2 4 1

States, General Accounting Office 1997). Given this market behavior, it is more likely the govemment
will be called on to providefinancial support later on.
B.

Fiscal Risks and the Challenge of Transition and Emerging Markets

The implicit and explicitfiscal risksthat both govemments in transition and in emerging- market
economies face are particularly large. Dependence on foreign financing, vague ownership structures in
the economy, underdeveloped regulatory frameworks, and weak enforcement exacerbate the scope for
failure in the financial and corporate sectors. Failures in turn often generate political pressures on
govemments to intervene, ad hoc and ex post, through variousfinancial bailouts. The recent history of
repeated bailouts, coupled with the lengthy tradition of a central plan, has produced enormous moral
hazard in the markets. The Hungarian and Czech Republic govemments were willing to repeat their
bailouts and recapitalize banks while accumulating public liabilities. The public liabilities that the Czech
Republic amassed outside the budgetary system outweighed its low sovereign debt levels (Czech
Republic, Ministry of Finance 1998).
The weak disciplinary effects of the internationalfinancial markets exacerbate thefiscal risks
that govemments in transition and emerging-market economies face. Markets that are small in size and
have a short history and limited disclosure of information limit the understanding that investors have of
therisks.This factor in part explains why in many emerging-market economies creditors have tolerated
excessive risk exposure by domesticfinancial institutions and enterprises before fleeing. As economies
integrate with the international markets, more reliable data become available for statistical analysis, a
shift that enhances the ability of both govemments and investors to estimate risks with standard
methodologies. These countries also have insufficient capacity to manage risk, a capacity that is
expensive to build (for instance, it is costly for govemments to replace low-paid bureaucrats with
financial analysts).

C.

Hidden Fiscal Risks and the Valué of Transparency and Certainty

Government commitments and promises outside the budgetary system blur the analysis of past
fiscal performance and future fiscal developments. Contingentfiscal riskssurface only after a delay and
in the form of unexpected requirements for public financing. Usually govemments lack information on
particularfiscal risksand overallriskexposure. Often they are not held accountable for the outcomes and
cost of their off-budget commitments. As a result, contingentfiscal risksmay accumulate and require
substantial govemmentfinancing in the future. In only a few countries are the govemments required to
assess and compare the full cost of alternative budgetary and off-budget programs and to report all
contingent liabilities and other fiscal risks. The United States and Italy provide some good examples of
risk assessment of state guarantees, and New Zealand and Australia do the same for the reporting of
contingentfiscal risks.
In making policy decisions, govemments often face a tradeoff between directly providing and
financing services or guaranteeing their provisión by the prívate sector. The former requires higher
budget outlays in the short term. Provisión by the prívate sector, with the state guaranteeing certain
outcomes, means minimal budget outlays in the short term but exposure to higher fiscal risks and
uncertainty about future publicfinancing requirements in the longer term. If a govemment is trying to
reduce the déficit and achieve certain short-term results, provisión by the prívate sector looks attractive.
Once a contingent liability falls due and requires govemmentfinancing, however, the govemment has
limited choices: it can increase the déficit, incur additional public liabilities without reporting any
Sesión 3 - 242

increases in the déficit, cut some envisaged expenditures, levy more taxes, sell state assets, default on
some obligations, or engage in some combination of these activities. Each of these actions challenges
govemment performance and credibility, with an attendant reduction in the effectiveness of future
policies, compromises political stability, and impairs future performance and growth in the overall
economy.
In this respect, the positive valué of certainty (the cost of uncertainty) about future public
financing requirements is an important factor for govemment decisionmaking. Alternative forms of
govemment support can be prioritized, not only based on their contribution to the desired policy
objectives and long-term cost, but also to reflect the volatility of the financing requirement and
contribution to the govemment's overallriskexposure.6
Certainty in public financing is particularly valuable to governments that have restricted or
unreliable access to borrowing, low risk management capacities, lowriskpreferences, and strategic cash
and debt management. Contingent liabilities are potentially very harmful for governments that cannot
rely on continued favorable access to borrowing. Large reserve funds may reduce the potential harm
from contingent liabilities when they fall due, but those funds come with an opportunity cost.
Governments with low capacity to analyze and managerisks,and in economies where outcomes are less
predictable and the asymmetry in information is greater, are ill-prepared to cope with the potential moral
hazard andfinancial uncertainties.7 Ideally govemmentriskpreference reflects theriskpreference of the
median voter. Arisk-aversegovemment chooses direct provisión, whose expectedfinancing requirement
is less volatile, over a guarantee, even if both involve equal risk-adjusted net presentfiscal costs and both
would deliver equal policy outcomes. Finally, for governments with sophisticated and efficiently
managed borrowing andfinancing strategies, an ad hocfinancing requirement involves costly disruptions
and efficiency losses.
D.

Fiscal Opportunism Under Debt and Déficit Ceilings

As yet there are no internationally accepted criteria for fiscal performance to address contingent
govemment outlays and encourage truly sound fiscal performance by governments and their fiscal
stability in the long term. Meanwhile, the tradeoff between long-termfiscal stability and the target level
for the budget déficit and debt, and between the quality of fiscal adjustment and the speed of déficit
reduction, may surface throughfiscal opportunism (a bias toward excessive accumulation of contingent
fiscal risks) and nonsustainable policies.
Opportunistic behaviors by countries under IMF and World Bank programs and, more recently,
by countries bidding for European Monetary Union membership indícate that a narrow focus on budget
déficit and debt compels governments to delay structural reforms and investments, conceal the cost and
financing of programs outside their budgets, and raise temporary revenues. Such behaviors generate
6

7

In a multipillar pensión system, the adequacy of govemment guarantees for returnsfrom prívate pensión f n s depends on t e
ud
h
analysis and assessment of therisksof guarantees a d on the govemment's capacity to regúlate and supervise prívate pensión
n
f n s and to cope with the problem of asymmetric information without incurring high transaction costs. Where t e
ud
h
govemment provides pensión benefits directly, it can predict total govemment outlays, and it seeks mainly to balance the size
of t e benefits, retirement age, and contributions to m k the pensión provisiónfiscally sustainable. In t e guaranteed portion
h
ae
h
of t e pensión system, citizens save for their retirement privately. However, the govemment faces a high level of uncertainty
h
about t e a o n s and timing of publicfinancing t a would be required should the pensión guarantees fall due.
h mut
ht
Consistent with t e conclusión in t e 1 9 World Development Report, t a governments should adjust the extent of their
h
h 97
ht
interventions to t e level of their institutional capacities, this paper argües that governments should adjust their overall risk
h
exposure to t e level of theirriskm n g m n capacities.
h
a ae e t
Sesión 3 - 243

uncertainties about future publicfinancing requirements and may endanger future fiscal stability. A focus
on a cash-based budget, déficits, and debt also distorts government decisions about spending priorities
and the timing and form of government support.
Governments may employ a number of opportunistic budgetary and accounting behaviors to meet
their déficit and debt targets (table 2). The behaviors involve any or all of three types of imprudent
actions: assumption of excessive liabilities for cash payment, a running down of public assets, and
excessive use of off-budget support in public policies. Governments doing their accounting on a cash
basis have wide scope to apply thefirst two actions. In cash-based accounting, expenses and liabilities
are accounted not when the obligation is incurred, but only when the government makes the actual cash
transfer. Thus governments collecting a fee for assuming a liability (for example, when it issues a
guarantee or accepts the pensión liability of an enterprise under privatization) report the income as a net
revenue gain. The third action occurs under both cash- and accrual-based accounting standards but is
eschewed in well-designed accrual budgeting rules.
An accrual-based accounting system without accrual budgeting is neither necessary ñor sufficient
to ensure that governments adequately consider contingentfiscal risksin policy. Although this system
encourages governments to prepare a statement of contingent liabilities andfinancial risks,it generally
does not require that the liabilities be included in the balance sheet and that the associated risks be
evaluated and quantified. International accrual accounting standards require that liabilities be accounted
only when the obligation is due with certainty.8
Policymakers are encouraged to make choices consistent with therisk-adjustednet present costs
of altemative policies and forms of government support in an accrual-based budgeting system that is built
on an accrual-based accounting platform. Accrual-based budgeting requires that the net present fiscal
cost associated with various government programs and contingent liabilities be included in budget
documents. This way, contingent liabilities enter the fiscal analyses and public accountability
frameworks from the moment government recognizes them.9

For a discussion of the rules of probability a driskassessment see International Federation of Accountants (1998).
n
Accrual-based accounting in t e public sector is t e trend in countries in the Organisation for Economic Co-operation and
h
h
Development. The International Federation of Accountants has proposed and elaborated accrual accounting standards for t e
h
public sector, and the proposed update of t e IMF's Government Financial Statistics methodology implies accrual-based
h
accounting. New Zealand and Iceland both have implemented accrual-based budgeting, and the United Kingdom, Sweden,
the Netherlands, Cañada, and Australia are doing so.

Sesión 3 - 244

Table 2
Opportunistic Government Behaviors to Meet
the Maastricht Déficit and Debt Ceilings
Behaviors

That increase
future
payables and
liabilities of
the
government

On the Expenditure Side

On the Revenue Side

To meet the déficit
rule:
• Introduce an ad hoc tax to be reimbursed
in the future
• Accept cash for a promise of future
benefits
• Record revenues gross rather than net of
the reimbursements, which are due later
• Exchange some existing public debt
instruments for indexed bonds sold at a
premium

To meet

the debt

rule:

• Transform indebted government agencies
into autonomous legal entities outside the
general government while granting them a
state guarantee
• Enter repurchase contracts with public
debt

That reduce
future
receivables

To meet the déficit
rule:
• Withhold revenues due in the following
fiscal year
• Accept cash in exchange for future tax
exemptions

That
dilute
the valué of
state assets

To meet the déficit
rule:
• Record the capital gains from a sale of
property, possibly with a subsequent
renting or lease-back arrangement
• Charge a dividend from revaluation of the
gold reserves of the central bank
• Charge a higher dividend from public
holdings
To meet the debt rule:
• Sell gold of the central bank
• Sell state assets

To meet the déficit
rule:
• Postpone inescapable
expenditures,
such as infrastructure investment and
maintenance
• Favor off-budget forms of government
support versus direct financing
• Delay legal recognition and financing of
government purchases and transfers
• Postpone
legal
recognition
and
quantification of rebates due taxpayers
• Record subsidies as purchases of (bad)
assets from corporations and banks at
face valué
• Record the déficits of state-owned and
municipal agencies providing nonmarket
public
services
outside
general
government figures
To meet the debt rule:
• Omit the existing net liabilities of public
enterprises and agencies that are outside
the sphere of general government but
that benefit from government guarantees
• Favor trade credit as a form of support
• Exelude contingent liabilities from debt
reports

Sesión 3 - 245

To meet the déficit
rule:
• Cut operations and maintenance
expenditures
• Reduce
expenditures
on
complementary inputs into the service
provided by the asset

IV. Understanding, Incentives, and Capacities to Reduce and
Control Fiscal Risks
Contingentfiscal risksmay significantly affect the results of a country's fiscal analysis. They
may also be an important factor in assessing allocative efficiency in the use of public monies (the implicit
subsidies and risk exposures relative to policy priorities). Finally, contingent and implicit forms of
govemment support may not only berisky,but may also lessen the government's operational efficiency
because they are unnecessarily costly compared to a direct, budgetary provisión.
Afirst necessary condition for fiscally prudent policies is for policymakers to identify, classify,
and understand the fiscal risks facing the govemment. Comprehension of the fiscal risks and their
consequences will at least encourage the govemment to avoid risks that are bound to surface in a
politically meaningful time horizon. Forrisksthat extend beyond that timeframe, achievement of fiscally
sound behavior may depend on coercion. In particular, policymakers are more likely to gravitate to
fiscally sound decisions if the media, the public, investors, credit-rating agencies, and multilateral
institutions understand the government'sfiscal performance in its entirety and if there are sanctions when
the govemment exposes the state to excessive risks and conceals those risks. (See the annex for a
questionnaire to use in evaluating the problem of govemment fiscal risks and risk management in a
country.)
Coercion as a means to discipline a government'sfiscal behavior beyond the budget déficit and
debt can be applied internally and externally. Internally, the principal audit institution can assess the
direct and contingentfiscal risksof each govemment agency and of govemment as a whole and make the
information public. Although voters do not necessarily care about govemment fiscal risk, public
explanation of the fiscal risks by an independent State Audit Office would also empower the external
forces of coercion. To be effective, extemal coercion should be used to ensure that the govemment
applies the international rules for fiscal analysis not only to the budget and debt, but also to its contingent
liabilities, and that it overcome the problem of asymmetric information. Specifically, extemal coercion
would pressure the govemment to meet certain quality standards: the govemment must define, measure,
and monitor fiscal performance in full, using sound indicators and methods as defined by international
authorities such as the IMF, World Bank, European Commission, or sovereign credit rating agencies and
investors, and it must develop adequate public finance institutions and disclose relevant information.
Govemments attempting to conceal data would be subject to sanctions.
The following sections discuss measures that a govemment can take to reduce itsriskexposure
and improve the quality of itsfiscal performance. These measures apply at the policy and institutional
levels, both systemically and at the various stages of govemment decisionmaking.10

10

For policies andfiscal institutions to reduce publicriskssee Schick (1998) and Irwin (1997).
Sesión 3 - 246

A.

Systemic Measures to Reduce Fiscal Risks

The main aim of systemic measures (table 3) to encourage sound fiscal behavior is to improve the
understanding of policymakers, the public, and the markets of thefiscal risks.

SYSTEMIC
Fiscal

Policy

MEASURES
Public

Finance

TABLE 3
TO REDUCE
FISCAL
Institutions

•
•

Identify, classify, and analyze all
fiscal risks in a single portfolio

•

Internalize and disclose the full fiscal
picture

•

1)

Consider ful! fiscal performance
beyond the budget and debt

Determine
the
government's
optimal risk exposure and reserve
policy according to
its risk
preference and risk management
capacity

•

Monitor, regúlate, and disclose the
risks in the public and prívate sectors

Policy

To achieve soundfiscal performance, the govemment needs to include in itsfiscal analysis and
decisionmaking thefiscal risksrelating to future possible obligations of the state and to consider those
fiscal risksin the context of its risk preference, riskfinancing, and risk management capacities. The
following steps are critical.
Consider fiscal performance in full, that is, beyond the budget and debt. Fiscal analysis,
especially of the quality of fiscal adjustment, is complete only if it factors in the cost of the implicit
subsidies in the govemment's contingent support programs. In particular, the govemment cannot
separate the analysis of its fiscal position from the obligations it has undertaken outside the budget
system. The arrears and other obligations of state-owned and guaranteed institutions, for example, may
claim significant public resources in the future. Moreover, the govemment may have mismanaged some
institutions tofinance and implement its policies outside the budget system. A string of years with a
balanced budget and low public debt suggests neither that the govemment has been fiscally prudent ñor
that there will befiscal stability.
In deciding between alternative forms of support, the govemment needs to consider the mediumtermfiscal impact and allocative and operational efficiency of programs pursued outside the budget to the
same extent as with the budget. Medium-term fiscal forecasts, the budget itself, and govemment
financing and borrowing plans are truly viable only if they provide for contingent and otherfiscal risks.
For international institutions, such as the IMF and World Bank, it is time to extend the scope of
their fiscal policy and institutional analysis to address contingent and implicit fiscal risks. Also,
international institutions should assist countries to develop adequate analytical and institutional
capacities, require countries to disclose information regarding theirfiscal risks, and enforce limits on the
countries' exposure tofiscal risksaccording to the countries' analytical and institutional capacities.

Sesión 3 - 247

R

Identify, classify, and anályze all fiscal risks in a single portfolio. To understand and prepare
for the entire range of potentialfiscal pressures, policymakers will have to take stock of all programs and
promises and identify and classify the main sources of fiscal risks,as shown in thefiscal risk matrix
(table 1). For each item of thefiscal riskmatrix and in order of significance, the government needs to
analyze theriskfactors and ways to control and reduce its exposure to the risks. Qualitative analysis of
risks would help the government formúlate and design sound new programs and promises.
The government should consolídate the stock of contingent liabilities into a single portfolio,
along with state debt and other public liabilities, so that it can evalúate correlations, sensitivity to
macroeconomic and policy scenarios, and overall risk exposure. A single portfolio allows the
government to relate its contingent liabilities to its comprehensive risk strategy and guidelines regarding
risk exposure, asset and liability management, hedging, and benchmarking. As an input into the analysis
of riskexposure, the government would also analyze information about budget arrears, state guarantees,
state insurance programs, subnational government borrowing, obligations of state-owned and stateguaranteed institutions, effects of private capital flows, and similar factors. In contrast to the déficit and
debt constraints, indicators reflecting a comprehensive analysis of the government's exposure to fiscal
risks would have greater predictive valué for fiscal stability. (The forthcoming, more detailed versión of
this paper includes afiscal riskanalysis for several countries.)
Determine the government's optimal risk exposure and reserve policy based on its risk
preference and risk management capacity. The government needs to base its risk and reserve strategy
on its overall risk exposure, risk preference, and ability to manage risk and absorb contingent Iosses.
Ideally, the risk strategy would be tied to the risk preference of the median voter. Similarly, the
government would assess new programs based on their marginal impact on overall risk exposure and
fiscal outlook. It would agree to further contingent and implicit forms of financial support only to the
extent it is able to evalúate, regúlate, control, and prevent the risks. If the government has a low capacity
to evalúate and manage risks, the best approach is to favor direct subsidies and provisión of services
rather than guarantees. That is, assuming that the government's intervention in a particular area is
justifiable, the government would opt for budgetaryfinancing of its intervention rather than ensuring that
particular outcomes will be delivered by the private sector. To this end the government has to enact
guidelines for prudent and soundfiscal management (as discussed in the institutional section below).
Reserve funds provide liquidity for guarantees and other contingent liabilities and thereby protect
the government against pressures to increase the déficit and debt, cut envisaged expenditures, or default
on its obligations if a contingent liability falls due. There is a tradeoff between the opportunity cost of
withholding resources instead of spending them or cutting taxes, on the one hand, and the benefits of a
reserve fund in promotingfiscal stability and government credibility, on the other hand. A reserve fund
offers the government morefinancial flexibility to deal with unexpected loss profiles if it sets the whole
portfolio of fiscal riskscentrally, rather than assigning ariskto eachriskseparately.11
2)

Institutions

An institutional framework for publicfinance will encourage the government to pursue sound
fiscal performance only if it encompasses both direct and contingentfiscal risks.A framework for public
finance management that ignores the future fiscal implications of contingent liabilities and other off-

11

Recently, Australia, Cañada, and t e United Kingdom have moved toward a central pool of unallocated, government-wide
h
reserves.
Sesión 3 - 248

budget commitments will only make such forms of government support look inexpensive and politically
attractive.
An adequate institutional framework requires that the government treat any noncash program
involving a contingent fiscal risk as it does other budgetary or debt items from the viewpoint of aggregate
fiscal stability and allocative and technical efficiency, control, public disclosure, and accountability.
Rules for issuing, monitoring, and handling state guarantees and insurance programs and for monitoring
andfinancial management of public, state-guaranteed, and subnational government institutions are also
needed. As the role of the state transforms from direct provisión of services to guarantees against
residual risks, governments need to follow the example of the prívate sector in deepening their capacities
for fiscal analysis and management beyond the state budget and debt.
Internalize and disclose the full fiscal picture. The rules and practices applied in the budget
process,financial management, and public accountability framework determine how muchflexibility the
government has to assume immediate and future direct and contingent nonbudgeted obligations.
Optimally, government choices will reflect qualitative and, where possible, quantitative evaluation of the
future outlays and risks associated with alternative forms of government support, including programs
outside the budget such as guarantees and activities of state-guaranteed agencies.
To address the problem of government accountability and fiscal discipline outside the budget,
public disclosure is more important than full-fledged accrual-based accounting, budgeting, and risk
measurement systems. Particularly for governments with lower institutional capacities, the system should
require them to assessriskfactors, make rough provisión for contingentrisksin the budget, and publish a
statement of contingent liabilities and overall risk exposure. Such a system is more sensible than the
optimal institutional framework, which involves accrual-based accounting and budgeting standards and
sophisticatedriskmeasurement methodologies.
Accrual-based budgeting and accounting standards make the potential fiscal cost and hidden
subsidies of contingent liabilities more transparent ex ante. In this context, quantitative risk analysis
reveáis the difference between the full risk premium, topped up by the cost of evaluating, managing, and
monitoring risks, and the fees the government charges for assuming a particular obligation (for instance,
the pensión liabilities of a privatized enterprise, a guarantee, or state insurance) at the time the coverage
is extended. By bringing off-budget commitments into the budget and recognizing the hidden subsidies
associated with contingent forms of government support, the government better reveáis the long-term cost
and benefits of its commitments and enhances public scrutiny of the potential use of public monies.
Public disclosure of fiscal information extending beyond the budget and direct debt enables the
public and markets to monitor the government's full fiscal performance, including the fiscal risks
accumulated outside the budget. Market agents such as investors and credit rating agencies are then able
to take both direct and contingent fiscal risks into account in their analysis and investment decisions.
Their ability to do so in turn indirectly encourages budgetary and overall fiscal discipline. In addition,
greaterfiscal transparency facilitates parliamentary scrutiny and monitoring by the market, particularly by
investors and sovereign credit rating agencies, and by international institutions such as the IMF, World
Bank, and European Commission.12
Domestically, the government can promote both fiscal transparency and prudent government
decisions by empowering the ministry of finance and principal audit institutions to monitor, control, and
12

The IMF h s outlined the requirements and good practices for fiscal transparency in its Manual on Fiscal Transparency
a
(1998). The m n a briefly discusses the transparency requirements for contingentfiscal risks.
a ul
Sesión 3 - 249

publish the size of contingent liabilities and otherfiscal risks, the extent to which the government's risk
exposures conforms to its proclaimed objectives, and the efficiency of both direct and contingent forms of
govemment support.
Monitor, regúlate, and disclose fiscal risks to the public and privóte sectors. Govemments
reduce the fiscal risks when they strive to prevent market failures and minimize the moral hazard
associated with their programs, commitments, and residual responsibility for market failures. To reduce
moral hazard and failures in the markets, the govemment maintains regulatory and law enforcement
systems, monitors the systemic risks in both the prívate and public sectors, and enforces transparency
about the risk exposure of bothfinancial and nonfinancial institutions in the markets. Well-developed
regulatory and public disclosure systems are particularly important when govemment embarks on
privatization while assuming an explicit or implicit obligation to cover residual liabilities and ensure that
prívate agents achieve particular outcomes.
Prevention of fiscal risks depends on a combination of analytical tools, incentives, and the
capacities of parliamentarians, civil servants, regulators, supervisors, international institutions, and
market agents. Research is being conducted to derive simple rules to indícate the dangers to fiscal
stability, using indicators such as the total face valué of all contingent liabilities, the overallriskassumed
by a govemment, the size and allocation of foreign prívate capital, and the accrual-based budget déficit.
Potentially the best place to develop risk monitoring capacities is in the central bank, given its role in
collecting balance-of-payments data and, in many instances, supervising banks. Specific regulatory and
supervisory agencies such as the securities and exchange commissions may best handle the monitoring of
specific risks. Ultimately, the ministry of finance and the office responsible for public liability
management would handle the monitoring and prevention of the government's overallriskexposure.

Undertake measures to reduce the fiscal risk of individual govemment programs and promi
Whether the government's programs, promises, and exposure tofiscal risks are appropriate depends on
their consistency with govemment policies and actions. The following aspects of consistency in
particular influence a government'sfiscal performance:
•

Consistency of govemment programs and promises with the stated role and strategic priorities of the
state

•

Consistency in the eligibility and management standards applied across govemment programs over
time

•

Consistency of therisksassumed and reserves provisioned under a program with theriskmanagement
capacities of the govemment

•

Consistency between the authority of policymakers to assume contingent fiscal risks and their
accountability.

Table 4 summarizes the steps a govemment needs to take to control itsfiscal risksbefore, when, and
after it announces a program or promise.

Sesión 3 - 250

STEPS

TO

CONTROL

TABLE 4
THE RISK OF INDIVIDUAL
GOVER
PROGRAMS
AND PROMISES

Measures

Fiscal

Before
govemment
admits
obligation
(program,
commitment,
promise
support)

• Assess how the obligation fits the • Evalúate the program risks
announced role and strategic priorities individually and in a single portfolio
along with existing risks, estimate
of the state
• Consider the choices of policies and the potential fiscal cost of the
forms of support with respect to obligation, and set an additional
associated financial
risks
and reserve requirement
govemment
risk
management • Design the program to protect
capacity
the govemment against risks
• Define
and
communicate
the
standards
for
and
limits
of
govemment involvement to minimize
moral hazard

an

of

Policy

Public

Finance

Institutions

When
the obligation is
held

• Stick to the pre-set
govemment responsibility

of

• Budget, account, and disclose
the obligation
• Monitor the program risk factors
and reserve-fund adequacy

After

• Execute the obligation within its preset limits and identify the lessons for
future policy choices
• If implicit, assess whether fulfilling
the obligation coincides with the
state's announced role and promotes
desired behaviors in the markets

• Compare and report the actual
fiscal costs versus the estimates,
evalúate performance, and impose
sanctions for failures

the
obligation
falls due

B.
1)

limits

Before Government Admits an Obligation

Fiscal Policy

Assess how the obligation fits the pronounced role and strategic priorities of the state. W
types of support the govemment decides to offer both outside or through the budget define the actual role
of the state. Therefore, programs outside as well as inside the budget should, in principie, be subject to
the same type of policy analysis and consideration. In the case of contingent support programs such as
guarantees for state institutions and funds, the govemment must consider whether their objectives fit
within its announced role and priorities and whether they justify the potential,risk-adjusted,long-term
fiscal costs.
Consider the choices of policies and forms of support relative to the associated financial
and government's risk preference and risk management capacity. The quality of fiscal performance
benefits when the govemment acknowledges the cost of uncertainty about future public financing
requirements in considering alternative programs and forms of support for particular policy objectives.
As with corporations, an unexpected requirement for financing disruptsfinancial planning and increases

Sesión 3 - 2 5 1

the cost of borrowing or, in a worse case, runs the risk that no creditfinancing is available. Governments
need to evalúate alternative ways to implement their policies not only on the basis of their potential cost
and benefits but also on the extent of the uncertainty they involve for future publicfinancing. In addition,
the government would judge contingent forms of support in terms of the extent of the asymmetric
information and transaction costs. These considerations would be made in the context of the
government's ownriskpreferences andriskmanagement capacities and the reliability of its access to ad
hoc borrowing.

Define and communicate the standards for and limits of government involvement to mini
the moral hazard. It is not so much the budgeted expenditures but the contingent liabilities, particularly
the implicit ones, as understood by the public and markets, that define the outer limits of state
responsibilities and affect the behavior of the public and market agents. The more formally and precisely
the government defines and signáis its responsibilities (its area of commitment), the more distinct are the
explicit liabilities and the smaller are the implicit liabilities. The more credibly the government defines
its responsibilities and the pain market agents will bear in cases of their failure and reliance on
government rescue, the less is the problem of moral hazard. Take the example of a society where the
government has a strong tradition of extensive public services. In such a case the central government
may be expected to take over any obligations of subnational governments in troubles. Such an
expectations raises a scope for moral hazard on the side of subnational governments. The central
government can reduce the moral hazard by signaling that it will only ensure the deiivery of core services
to citizens of insolvent subnational governments. At the same time, it can state that it will not bail
municipalities outfrom their debts and non-core expenditure obligations.
The particular task of government is to signal credibly what actions the markets should not expect
of it in the case of various market failures. If these failures occur, the government will gain needed
credibility and reduce moral hazard in the markets, and so curtail its fiscal risks,if it follows through on
its stated policies and refuses to submit to pressures for alternative actions.
2)

Institutions

Evalúate the risks of programs individually and in a single portfolio that also contains ex
risks, estimóte the potential fiscal cost of each obligation, and set additional reserve requirem
Qualitative analysis of theriskfactors in alternative government programs and estimates of their potential
long-term fiscal costs and hidden government subsidies prior to any commitment helps optimize the
choice and design of government programs. Rough quantification of theriskand potentialfiscal cost of
government contingent liabilities and commitments requires good qualitative analysis of the underlying
risks. Specialized methodologies such as option pricing, actuarial analysis, rate-setting, value-at-loss, and
loss-cost ratio are of great valué in deriving a more precise estímate of the potential costs of a particular
program.13 According to government reserve policy, theriskexposure of a proposed program added to
overall government exposure determines the amount of additional resources that should go into the
government reserve fund.
Design the program well to protect the government against risks. Based on the qualitative risk
analysis, the government needs to identify thoserisksit can control reasonably well, decide whichrisksto
13

Various types of risk,such as sovereign (political, legal, and regulatory),financial credit (foreign exchange rate, interest rate,
and refinancing), and program performance (development, completion, and operating) compound thefiscal risksof
government programs and promises involving both direct and contingent liabilities. For a m r detailed analysis of t e types
oe
h
of risksee Chase Manhattan Bank (1996). For methodologies to estímate potentialfiscal costs see Mody and Lewis (1997),
Mody and Patro (1996), United States, General Accounting Office (1997), and Penacchi (1997).
Sesión 3 - 252

cover under its proposed program of contingent support, and develop effective risk-sharing,regulatory,
and control mechanisms to monitor the performance of the parties under the program Apart from
exogenous risks such as drought, the government faces endogenous risks that are mainly a function of
program design- A poor design can create varying levels of market distortion and moral hazard, whereas
a good design can reduce the potentialfiscal cost of the program. An example is a guarantee contract that
covers only political and not commercialrisks,only 30-50 percent of the valué of the potential loss, and
the last rather than thefirst portion of the loss. Programs that involve implementation by an intermediary
agency that itself must be established, such as a guarantee fund of any sort, are more difficult to design,
particularly in terms of management incentives and performance monitoring by the government.
For some programs, the government may charge arisk-basedpremium, purchase reinsurance
from prívatefirms, or contract out particular risk management functions.14 It may cancel other programs
altogether when it corrects market functioning through its regulatory policies. For example, deregulation
of the insurance markets will encourage foreign insurancefirms to enter the domestic markets and greatly
expand the pooling of somerisksso that the prívate sector can cover them. Risks related to disasters that
are uninsurable in a domestic market because the risk pool is too limited become insurable in the
integrated international market. The government can in turn end such programs as crop and flood
insurance. For the largestfinancial risks,such as major banking and currency crises, it can be argued that
the IMF, World Bank, and other multilateral agencies will provide governments with some kind of
reinsurance.

C.
1)

When the Government Accepts and Holds an Obligation

Policy

Stick to the pre-set limits of government responsibilities. After the government approves a
program or commitment, the main challenge is to ensure that the markets and public do not expect any
state support beyond the announced limits over the life of the obligation. Any indication that the
government might providefinancial support beyond the announced limits will raise the moral hazard for
and distort the behaviors ofthe parties potentially benefiting from the program.
2) Institutions
Budget, account, and disclose the obligation. On the institutional side, the government faces the
challenge of budgeting, accounting and provisioning for, and disclosing the obligation adequately. How
does it ensure that no unknown contingent liability appears only after it is triggered? For instance, the
publicfinance law can state that an obligation is valid only if it was assessed, budgeted, accounted, and,
above all, disclosed at the time of its adoption by government.
Monitor the program risk factors and reserve-fund adequacy. Over the life of an obligation, the
government needs actively to monitor the program'sriskfactors, the performance of the agents under the
program and, in this context, also the adequacy of its reserve funds. Monitoring of intermediary agencies,
14

The ultímate responsibility for projectriskevaluation and program design is probably best placed with t e ministiy of finance,
h
which approves and disapproves any potentialfinancial c m im ns of the government. The office for public debt
o mt e t
m n g m n is likely the best equipped to analyze contingentfiscal risksand integrate t e into a single public liability
a ae e t
hm
portfolio. It is also best placed to decide on hedging a d otherrisk-controlinstruments. For examples of policies to protect
n
t e government against excessiveriskexposure see Schick (1998) and Irwin and others (1997).
h
Sesión 3 - 253

such as banks and various credit and guarantee funds that the state uses to implement its policy objectives
and guarantees, is particularly important. If the govemment lacks a good monitoring capacity, it can
contract this task out for a performance-based fee. The cost of monitoring and administering programs of
contingent support may be relatively high and should be reflected in the ex-ante calculations of the
potentialfiscal cost of a program.
D.
1)

After a Liability Falls Due

Policy

Execute the obligation within its pre-set limits and identify lessons for future policy choice
is critical that the govemment meet an obligation when it falls due within the stated limits, particularly in
terms of the credibility of future announcements and the scope for future moral hazard in the markets.
For instance, paying depositors more than the specified deposit insurance levels tells the markets that the
govemment will submit easily to political pressure, tells depositors that banks offering higher yields are
"safe," and tells the banking sector that excessiverisksare worth taking.
By applying the lessons from its involvement with direct and contingent liabilities, both explicit
and implicit, the govemment is able to adjust its role incrementally, rather than abruptly, in a crisis. A
timely and credible explanation of any adjustment in the state's role that will affect future policy choices
will prompt the public and market agents to adjust their expectations and behavior. For example, by
explaining that the public pensión scheme is not fiscally sustainable and that future govemments will
have to reduce the pensión benefit significantly, the govemment influences the saving behavior of people
in the labor forcé.
If an obligation is implicit, assess whether it coincides with the state's announced role
promotes desired market behaviors. When public interest groups or market agents suddenly cali on the
govemment to extend more support than was originally specified, policymakers need to ask whether
extending that support coincides with its announced role and how it affects future behavior in the
markets. The long-term damage to the govemment of acting upon an ad hoc request may sharply exceed
the potential short-term benefit. Acting upon ad hoc requests may, however, be politically attractive, and
the govemment is often able to find ways to improperly usefinancial institutions and funds outside the
public sector to implement and finance its actions. Thus the public, investors, and international
authorities need to monitor the government's responses to ad hoc claims of an implicit govemment
liability and apply sanctions for fiscally irresponsible choices.
2)

Institutions

Compare and report the estimated and actual cost of govemment support, evalúate
performance, and apply sanctions for failures. The requirement that the govemment report and
compare the ex-ante risk evaluation and actual layouts for a program is critical to govemment
accountability. Performance evaluation applies to govemment departments and officials as well as to the
parties under a program. Sanctions may involve govemment officials (the case where particular interests
distorted the ex-ante risk analysis), the managers of state-guaranteed and intermediary agencies
implementing the government's programs (such as for exposing the govemment to unnecessary and
excessiverisks),and the parties under the program (where they breached an agreement).

Sesión 3 - 254

V.

Conclusions

Governments are exposed to increasingfiscal risks and uncertainties as a result of the increasing
volumes and volatility of prívate capital flows, the changing role of the state, which is shifting from direct
provisión andfinancing of services to guarantees for certain outcomes, biases in policy decisionmaking
underfiscal constraints, and the moral hazard in the markets associated with expected state interventions.
The fiscal risks are particularly large for transition and emerging-market economies, where market
opacity and the danger of market failures are greater.
Governments face four types of fiscal risk:direct and contingent, each of which may be either
explicit or implicit. Most govemment andfiscal analysts concéntrate on direct liabilities (direct explicit,
such as the public debt and govemment budget, and direct implicit such as future pensión and social
security liabilities). Recent international experience, however, indicates that significant fiscal instability
may result from contingent liabilities (contingent explicit such as the obligations of state-guaranteed
institutions and deposit insurance, and contingent implicit such as local govemment obligations, foreign
credit of the domestic corporate andfinancial sectors, and banking failures).
Therefore, a study of govemmentfiscal position cannot be separated from obligations taken by
the govemment outside the budgetary system. Fiscal analysis and medium-term fiscal framework for
countries must factor in the cost of implicit subsidies provided by the govemment in the forms of
contingent support programs. For international institutions, such as the World Bank and IMF, it is time to
(a) extend the scope of their fiscal, policy, and institutional analysis beyond the budget and debt; (b)
require countries to disclose information about their contingent govemmentrisks;and (c) assist countries
to reform their analytical, policy, and institutional publicfinance frameworks to address all major fiscal
risks.
Critically important to long-termfiscal stability and equity is public recognition of the limits of
the state's role and the associated direct and contingent fiscal risks. Public accountability of politicians
and civil servants in areas beyond the state budget must be defined to promote prudent and efficient fiscal
policies and management. Governments need to address the sources of fiscal risksin three ways: (a) by
understanding existing and future fiscal risks and pursuing policies that foster appropriate fiscal
adjustment; (b) by developing an institutional framework that involves adequate public disclosure and
incentives with respect tofiscal risks,and that promotes fiscal prudence and equity in all govemment
programs, including those extending support outside the budget system; and (c) by building and
employing institutional capacities to evalúate, regúlate, control, and preventfinancial risksin both the
public and prívate sectors.
This paper offers an analytical framework to study contingent liabilities and otherfiscal risksof
governments so that analysts can better assess the quality of fiscal adjustment and thefiscal outlook of a
country. A more detailed versión is forthcoming that expands on both the policy and institutional aspects
associated withfiscal risks and uses the analytical framework to assessfiscal riskin several countries.

Sesión 3 - 255

VI.

References

Chase Manhattan Bank. 1996. Risk Management Handbook, Private Sector Participation in Urban
Infrastructure. Washington, D.C.: United States Agency for International Development.
Czech Republic, Ministry of Finance. 1998. Revealing the Hidden Public Debt. Prague: Ministry of
Finance, April.
Easterly, William. 1998. "When Is Fiscal Adjustment an Alusión?" World Bank, Washington, D.C.
Processed.
Forte, Francesco. 1997. "Prudential Accounting and Maastricht Rules." A paper presented at the EU
Accession and Sovereign Debt Management Seminar, Brussels, December 15-16.
Heller, Peter. 1997. Aging in the Asian "Tigers":
International Monetary Fund.
International Accounting Standards Committee.
London.

Challenges for Fiscal Policy. Washington, D.C.:

1997.

International Accounting Standards 1997.

International Federation of Accountants, Public Sector Committee. 1998. "Guideline for Governmental
Financial Reporting." Exposure draft. New York.
IMF (International Monetary Fund). Forthcoming. Manual on Fiscal Transparency. Washington, D.C.:
IMF.
Irwin, Timothy, and others, eds. 1997. Dealing with Public Risk in Private Infrastructure.
Latin American and Caribbean Studies. Washington, D.C.: World Bank.

World Bank

Mody, Ashoka, and Christopher M. Lewis. 1997. "The Management of Contingent Liabilities: A Risk
Management Framework for National Governments." In Timothy Irwin and others, eds. Dealing
with Public Risk in Private Infrastructure. World Bank Latin American and Caribbean Studies.
Washington, D.C.: World Bank.
Mody, Ashoka, and Dilip K. Patro. 1996. "Valuing and Accounting for Loan Guarantees." The World
Bank Research Observer 11(1)( February).
Pennacchi, George. 1997. "Government Guarantees for Oíd Age Income." A paper prepared for the
May 12-13, 1997 Symposium, "Prospects for Social Security Reform," sponsored by the Pensión
Research Council of The Wharton School, University of Pennsylvania, Philadelphia.
Schick, Alien. 1998. A Contemporary Approach to Public Expenditure Management.
Development Institute. Washington, D.C.: World Bank.

Economic

Selowsky, Marcelo. 1997. "Fiscal Déficits and the 'Quality' of Fiscal Adjustment." A paper presented
at the EU Accession and Sovereign Debt Management Seminar, Brussels, December 15-16.
Standard & Poor. 1997. "Criteria: Financial System Stress and Sovereign Credit Risk." Sovereign
Ratings Service (December).
United States, General Accounting Office. 1993. Bank Insurance Fund: Review of Loss Estimation
Methodologies. Washington, D.C.
. 1997. "Budgeting for Federal Insurance Programs." Report to the Chairman, Committee on the
Budget, U.S. Houseof Representatives. GAO/AMID-97-16. Washington, D.C.
World Bank. 1997. World Development Report 1997: The State in a Changing World. New York:
Oxford University Press.

Sesión 3 - 256

VII. Annex
Public Liabilities
how Big a Problem in a Country? A Set of Questions
The Whole Picture: Coverage
1. What are the major risks to future fiscal stability? Fill in the table below with specific items. These
include direct borrowing, guarantees, and institutions that are covered by some type of government
guarantee, state insurance programs, and all government commitments to spend or intervene
financiálly in the future. In classifying the items, think of direct liabilities (obligations of the
government in any event) and contingent liabilities (obligations of the government if a particular
event occurs), each of which can be either explicit (defined by a law or contract) or implicit (broadly
predetermined by public expectations and pressures by interest groups).

Liabilities

8 Direct

(obligation

Contingent
(obligation
in any event)

if

a particular

event

occurs)

Explicit
Government
liability
as
recognized by
law or contract

Foreign and domestic
sovereign
borrowing
(loans contracted and
securities issued by
the
central
government)
Expenditures
by
budget law
Budget expenditures
legaliy binding in the
long term (civil service
salaries, civil service
pensions)

State guarantees for nonsovereign borrowing and
obligations issued to subnational governments and
public and prívate sector entities (development banks)
Umbrella state guarantees for various types of loans
(mortgage loans, student loans, agriculture loans,
small business loans)
Trade and exchange rate guarantees issued by the
state
Guarantees on borrowing by a foreign sovereign state
State guarantees on prívate investments
State insurance schemes (deposit insurance,
mínimum returns from prívate pensión funds, crop
insurance, flood insurance, war-risk insurance)

Implicit
A
"moral"
obligation
of
the
government
that
mainly
reflects public
expectations
and pressures
by
interest
groups

Future recurrent cost
of public investment
projects
Future public pensions
(as opposed to civil
service pensions) if
not required by law
Social
security
schemes
if
not
required by law
Future health care
financing
if
not
specified by law

Umbrella default of a subnational government and
public or prívate entity on nonguaranteed debt and
other liabilities
Liability cleanup in entities under privatization
Banking failure (support beyond state insurance)
Investment failure of a nonguaranteed pensión,
employment, or social security fund (social protection
of small investors)
Default of central bank on its obligations (foreign
exchange contracts, currency defense, balance-ofpayments stability)
Bailouts following a reversal in prívate capital flows
Residual environmental damage, disaster relief,
military financing

Of the fiscal authorities, not the central bank.

Sesión 3 - 257

2. Is there a precise legal delineation of the public sector (for example, in the form of a full list of public
sector agencies) and of govemment responsibilities? If yes, note the definition and reference
appropriate legal documents.
Selected Risks
1. State-guaranteed institutions and directed credit
•
•

List all institutions that fulfill orders of the govemment to extend financing to enterprises, banks,
agencies of any kind, or households. Provide their balance sheets and statements of contingent
liabilities.
What type of govemment support do these institutions receive (for example, privatization
revenues, cheapfinancing via the central bank, state guarantee for borrowings)? Try to draw a
diagram showing the institutions involved in directed credit and thefinancial and crosssupporting flows.

2. Guarantees
• List all govemment guarantees, their issuer (the ministry of finance or other govemment agency),
beneficiaries, creditors, face valúes, the type of risks and their shares covered, currency of
denomination, andriskestimates if any.
3. State-owned enterprises and banks
• List all large state-owned enterprises and provide their audited balance sheets and statements of
contingent liabilities.
• List all large state-owned banks and provide their audited balance sheets, statements of
contingent liabilities, andriskassessment of assets.
Recording and Reporting: Transparency
1. For each type of direct and contingent liabilities you identified in the table above, list the
institutions responsible for final approval, recording, monitoring, and data consolidation.
2. Which institutions can instantaneously retrieve from their databases up-to-datefigures for the
items listed below? Which documents report suchfigures? What is the time lag in reporting:
•
•
•
•
•
•
•
•

Sovereign debt portfolio (break down according to maturities, currencies, and interest rate types)
Debt service profile for the next months and years?
Guarantee portfolio (breakdown according to guaranteed institutions, sectors, currencies)?
Total face valué of all state guarantees?
Total sizes of state insurance schemes?
Total sizes of reserve funds associated with guarantees and state insurance schemes?
Prívate foreign and domestic borrowing?
Sector allocation of foreign credit?

Sesión 3 - 258

2.

Which sources of fiscal risksare, in your view, not reported to the:
•
•
•
•
•
•

Ministry of finance
Cabinet
Central bank
Parliament
Foreign investors
Public.

Institutional Arrangements: Accountability
1. Are there any legal requirements that the govemment estimate, account, and report the future fiscal
costs associated with its budgetary policies and off-budget promises (such as guarantees and other
contingent liabilities)?
No
Yes— in the budget process
when the govemment is called on to pay
when cash is transfeired
other
2. Which of the liabilities that you identified in the table are not regulated by any law and depend fully
on ad hoc govemment decisions?
3. Describe or provide references for:
•

State guarantees: the requirements for their design (the type of risks that can be covered, the
extent of required risk-sharing), issuance (is only the ministry of finance authorized?),
govemment control mechanism (required reports from the creditor and beneficiary, audit and
valuation requirements), and realization mechanism if they fall due
• Subnational governments, public sector agencies and enterprises, and state-guaranteed
institutions: the financial management and reporting requirements and govemment control
mechanism
• Demands on the govemment to extend ad hoc, previously unforeseen financial support: the legal
requirements and practice for deliberation in govemment decisionmaking.
4. Is the govemment legally required to explain the amounts of public liabilities?
No
Yes— to the Parliament
to the public
other
Policy: Practice
1. When considering alternative policy choices and forms of govemment support (such as direct
provisión and financing versus guarantees), do the ministry of finance, cabinet, central bank or
parliament
•
•

Quantify the future fiscal cost of alternative options in a single medium-termfiscal framework?
Describe therisksof alternative options?

Sesión 3 - 259

2. In which areas and under what circumstances do the public or interest groups expect the government
to providefinancial support beyond the budget?
3. List examples when the government withstood political pressure and did not providefinancial support
beyond the budgeted figures (for example, when the government refused to solicitfinancial support
for a failed enterprise or bank).
4. Are public enterprises and banks, state-guaranteed institutions, and creditors and beneficiaries under
state guarantees "rewarded" and "punished" for the quality of their management of risk? Provide
examples.
Risk Management: Capacities
1.

Describe the capacities of the ministry of finance, other government agencies, public sector
institutions and enterprises, and state-guaranteed institutions to evalúate and control the risks of
government programs and contingent liabilities.

2. Describe the process of designing a state guarantee or state insurance program.
3. How is the required size of the government reserve fund determined?
3.

What steps do the ministry of finance and other agencies undertake to preventfiscal risksarising
from the public and private sectors (for example, are any actions taken if enterprise debt or
central bank obligations appear too high)?

Sesión 3 - 260

MANAGING FISCAL RISK IN BULGARIA

by
Hana Polackova, Sergei Shatalov and Leila Zlaoui
The World Bank

The authors thank Patrick Honohan (financial sector) and Esen Ulgenerk (pensions and environment) for
their contributions and Zhicheng Li for her research support. Lubomir Christov, Michael Dooley, Balazs
Horvath, Stella Ilieva, Kathie Krumm, Kyle Peters, lena Roussenova and Salman Zaheer provided helpful
comments and suggestions. Special thanks to Mariella Nenova who coordinated commewnts from several
Government agencies in Bulgaria. Thefindings, interpretations and conclusions expressed in this volume
are those of the authors and should not be attributed to the World Bank, affiliated organizations or
members of its Board of Directors or the countries they represent.

Sesión 3 - 261

TABLE OF CONTENTS
1. Introduction
I. Is BULGARIA EXPOSED TO FISCAL RISKS?

1. Bulgaria 's Fiscal Risk Matrix
Government exposure to risks is not negligible
2. Why Fiscal Vulnerability
Currentfiscal position appears good
Room to accommodatefiscal risk is, however, limited
II. ANALYSIS OF INDIVIDUAL SOURCES OF RISK

1. Direct explicit Risk
Sovereign debt poses medium-termfiscal risk
The proposed reform would bring social security under control
Healthfinancing is likely to grow, either directly or through contingencies
2. Direct implicit Risks
Public investment Program m y generate significant fiscal pressures
a
3. Contingent explicit risks
State guarantees are not large but should be treated with caution
Large environment damages need to be
financed
How large obligations will guaranteed agencies be permitted to accumulate?
4. Contingent implicit risks.
Enterprise obligations m y turn out expensive
a
An implicitfiscal pressurefrom municipalities?
Fiscalriskfrom thefinancial sector is currently low
III. DEVELOPING FISCAL RISK MANAGEMENT FRAMEWORK

1. Individual measures to address selected fiscal risks
Priorities for the short term
Priorities for the médium term
2. A better framework for fiscal management
IV. CONCLUSIONS

2
3
5

5
6
8
9
1
0
12

12
1
2
1
8
1
9
20
2
0
21
2
1
2
5
2
6
27
2
7
2
8
2
9
32

32
3
2
3
3
34
37

V. BlBLIOGRAPHY

39

ANNEX L TABLE ON BULGARIA GENERAL GOVERNMENT
ANNEX 2 TABLE 1: STATE GUARANTEED CREDOS
ANNEX 3 H o w TO BUILD RISK MANAGEMENT SYSTEM FOR A GOVERNMENT?

41
42
44

2

1. Introduction

International evidence has confirmed again and again thatfiscal analysis is incomplete if
it skips over "hidden"fiscal risks,such as obligations made by the government outside the
budget. First, as Kharas and Mishra (1999) explain, analyses of past increases in the stock of
government debt have shown that governments often accumulate debt as a result of "hidden
déficits" rather than reported budget déficits. Hidden déficits mainly arisefrom debt structure, as
currency, maturity or interest raterisksmaterialize, andfrom off-budget government obligations,
such as contingent liabilities that fall due. Most recently, some of thefiscal pressures that have
emerged from the East Asian crisis can be attributed to thefiscal risksdue to the governments'
contingent liabilities (World Bank, 1999).
Second, any changes in the reported budget déficit are an illusion if they are accompanied
by offsetting changes in the valué of public assets and in the countiy's exposure tofiscal risks,
such as government contingent liabilities (Easterly, 1998). Furthermore, as Selowsky (1998)
emphasizes, reported déficit improvements do not necessarily imply "quality" of fiscal
adjustment, which has the dimensión of sustainability as well as efficiency. Narrow interests in
reducing budget déficit may actually increase rather than reduce government exposure to fiscal
risks, and deteriórate rather than improve the prospects of future fiscal performance.
Third, the conventional approach to the analysis of déficit sustainability is limited in two
ways: it looks only at the liability side of the public sector balance sheet and it considere only
direct liabilities, ignoring contingent liabilities, both explicit and implicit.1 In this context, fiscal
vulnerability is defined by Hemming (1999) as a situation where the government is exposed to
the possibility of failure to achieve its broadfiscal policy objectives. It is concerned in particular
with the emergence of unexpectedfiscal risksand policy challenges and with the government's
capacity to respond to them. Fiscal vulnerability takes into account: (a) the initial fiscal position
(including the central government budget, other levels of government, extrabudgetary funds and
quasi-fiscal activities, assets and liabilities, contingent liabilities, fiscal indicators), (b) sensitivity
of thefiscal position to short-termrisks,such as macroeconomic volatility, called contingent
liabilities, and unclear expenditure commitments, (c) médium- and long-termfiscal sustainability
(debt dynamics, baseline projection and stress testing, and long-term pressures from demographic
trends, resource depletion, etc.), and (d) strucutural or institutional weaknesses (expenditure
composition, revenue system, déficit financing, government access to debt markets, institutional
capacity for fiscal management).

Under the conventional approach, the actual déficit is compared with the estimated sustainable déficit level that
will keep the debt to GDP ratio constant for feasible rates of growth, real interest, and inflation. This approach
assumes that keeping a constant ratio of public debt to GDP will ensure public sector solvency and avoid debt
crises in the fiiture. Another, less stringent requirement is to test for the no-Ponzi scheme condition for public
debt, followed up by the neoclassical solvency approach. This methodology checks for pubüc solvency by
comparing the ratio of public debt to GDP with the real interest rate. If the debt ratio systematically grows
faster than the real interest rate, the public sector is considered insolvent. For background on fiscal
sustainability analysis, see Anand (1988,1989, and 1990).

3

Analysis of fiscal risksbecomes increasingly important. Reasons include increasing
volumes and volatility of private capital flows, transformation of the state from financing of
services to guaranteeing particular outcomes, and related to both of these, moral hazards in the
markets, andfiscal opportunism of policy makers. Fiscalrisksbecome particularly threatening in
countries with a limited scope for maneuver in governmentfinancing. Limited access to debt
market and constraints on the use of monetary policy instruments reduce the amount of fiscal
risks that governments "safely" take on.
Bulgaria's ability to absorbfiscal risksis limited. Currency board arrangement, limited
access to debt markets, already high levels of tax revenues, and a large share of nondiscretionary
expenditures severely constrain the scope for government maneuver in cases whenfiscal risks
materialize. The country is recovering from years of macroeconomic instability, high inflation
and currency crisis. Its economy is still fragüe and realization of fiscal risks thus may have very
serious consequences for the country's economic and social development as well as fiscal
performance. The objective of accession to the European Union also demands the government to
maintain allfiscal risksunder a very tight control while accommodating substanüal infrastructure
and environmental investment to achieve high growth and/or meet accession requirements. With
already high level of indebtedness, Bulgaria faces a difficult trade-off. Should Bulgaria adopt a
veiy conservative stance towards debt andfiscal riskat the expense of investment and growth or
rather accept a slower pace of dis-indebtedness and guarantee the resources needed for economic
restructuring and mitigation of the impact on the poor and vulnerable? So far, Bulgaria has been
successful in reducing its debt burden while relying on high levels of fiscal reserves as a main
contingency instrument. Its future investment and developmental agenda, however, cali for a
more optimal mix of fiscal reserve, debt management andriskmitigation strategies.
The objective of this paper is to present a systematic analysis of fiscal risks in Bulgaria
and advise onriskmonitoring, assessment, management and mitigation strategies. The paper is
divided into four sections. After this introduction we present a simple framework, the Fiscal
Risk Matrix, to identify and classify Bulgaria'sfiscal risks.We set our analysis of fiscal risksin
a broader macroeconomic and institutional background, leaning toward assessment of Bulgaria's
overallriskexposure andfiscal vulnerability. Section II offers analysis of the individual sources
of fiscal risk in Bulgaria. We analyze the size, probability and sensitivity of risksarising from
direct and contingent, both explicit and implicit government liabilities. Section ID outlines our
recommendations for Bulgaria to strengthen itsfiscal riskmanagement framework and debt
management policy. We offer a set of immediate and medium-term measures to contain some of
the main sources of fiscal risk.Furthermore, building on country experience with medium-term
expenditure framework (Campos and Pradhan, 1996), we develop a broader medium-term fiscal
framework, as an institutional arrangement for counúy fiscal management. Finally, we
summarize ourfindings in section IV.

4

I. Is Bulgaria Exposed to Fiscal Risks?
1. Bulgaria 's Fiscal Risk Matríx

As in Polackova (1998) we focus onfiscal risksemerging from govemment obligations
of four types: direct explicit, direct implicit, contingent explicit and contingent implicit (table 1).
Government direct explicit liabilities are specific obligations that will fall due with certainty and
are defíned by law or contract. They are the subject of traditionalfiscal analysis and, in Bulgaria,
particularly include sovereign debt service and, in the long term, legally mandated pensión and
health expenditures. Government direct implicit liabilities represent a moral obligation or
political, rather than legal, burden on the govemment that will occur with certainty. They often
arise as a presumed consequence of public expenditure policies in the longer term. In Bulgaria,
the largest are accumulated and expected public investment needs to deliver anticipated public
services and meet key requirements for accession to the European Union. Explicit contingent
liabilities represent govemment's legal obligations to make a payment only if a particular evení
occurs. In Bulgaria, state guarantees for nonsovereign borrowing and obligations for past
environment damages are the main examples of this type of govemment obligation. Implicit
contingent liabilities are those that are not officially recognized until a failure occurs. The
triggering event, the valué atrisk,and the required size of the govemment outlay are uncertain.

5

Table 1

Bulgaria's Fiscal Risk Matrix

Sources of
fiscal risk

Contingent (obligation if a particular event occurs)
Direct (obligation in any event)

Explicit

•

•

•

Implicit
A "moral"
obligation of
the govemment
that mainly
reflects public
expectations
and pressures
by interest
groups

•

•

Future pensión
expenditures required by
law
Health expenditures
required by law

•

Individual state guarantees for nonsovereign
borrowing and obligations

•

Obligation to recover past environment
damages assumed in enterprise privatization
and other environment liabilities

•

Obligations of business promotion bank

•

Obligations of export insurance agency
(insurance policies to cover political and
medium-term commercial risks)

•

Government
obligation is
recognized by
law or contract

Foreign and domestic
sovereign debt (size and
structure)

Obligations of state fund for agriculture

Accumulated and
expected public
investment needs to
sustain delivery of public
services and meet key
requirements for
accession to the EU

•

Environment commitments for still
unknown damages and nuclear and toxic
waste

•

Clean up of enterprise arrears and liabilities

•

Future recurrent costs of
public investment
projects

Default of municipalities on own nonguaranteed debt, own guarantees, and/or
own obligations to provide critical public
services

•

Support to the banking sector in case of
crisis
Obligations listed above refer to thefiscal authorities, not the central bank.

Government exposure to risks is not negligible
Bulgaria's future fiscal position may suffer from short-term shocks arising from
fiscal risks identified in the Fiscal Risk Matrix. Structure and size of Bulgaria 's sovereign debt
are somewhat worrisome. Even after the 1992-94 successful restructuring, Bulgaria remains one
of the most heavily indebted countries of Central and Eastern Europe. High debt levels remain a
significant source of fiscal vulnerability, threatening the country creditworthiness, and impeding
Bulgaria's access to internationalfinancialmarkets.
Of the 83 percent of GDP of public debt, about 90 percent is foreign, nearly a half in
Brady bonds and the rest mostly owed to international financial institutions. Most debt
instruments are long-term, with a floating interest rate, and denominated in US dollars. The
external debt service ratio is projected to remain within a manageable, but rather high range of
20-22 percent of exports of goods and services2. The debt structure is rigid and implies
Traditionally, a debt service ratio of 25 percent or higher is considered to be a fairly reliable predictor of debt
crisis. Bulgaria's debt service ratio is below this crisis warning threshold, but is still rather high; further sections
of this paper suggest that adverse developments could hike the debt service ratio above the 25 percent warning
threshold.

6

significant refinancing risk. Bulgaria has not regained reliable capital market access and faces
volatility of investors' preference. Integrated asset and liability framework indicates that
currency risk is on average naíurally hedged by dollar-denominated exports but is significant
with regards to short-term exchange rate movements. Interest rate risk is also substantial. One
percentage point rise in the international interest rates (Libor) would transíate into an additional
70-80 million US dollars in annual debt service over 2000-2004. Stress-testing interest rate
indicates that a 3 percentage points rise in Libor would bring Bulgaria's debt service ratio above
25 percent of exports - a level that often serves as a crisis warning indicator.
Other risks to fiscal stability in Bulgaria are mainly associated with the remaining
transition process and with the cost of economic and social restructuring. Potential fiscal
pressure may particularly arise from the financial and environmental liabilities of state-owned
enterprises that the govemment may have to assume during restructuring, privatization or
liquidation of enterprises. Part of these liabilities have already been made explicit, others are still
unknown.
Environment, together with the countiy's infrastructure network and energy sector, also
stands for a large amount of investments required in the context of Bulgaria's accession to the
European Union. Overall environmental expenditures are expected to reach about US$8.5
billion, that is 69 percent of 1999 GDP, by 2015, implying an approximate US$0.5 billion per
year in annual public investment expenditure. The Bulgarian govemment public investment
program for 2000-2006 envisages an annual disbursement of US$ 400 (3 percent of GDP in
2000) million for environmental projects. With public investment projected at about 3.5 percent
of GDP in 2000 and the years ahead, the need for investment in infrastructure, institutional
capacity and social welfare is likely to result in significant fiscal pressures. Henee, Bulgaria's
development and progress toward the EU accession will rely on the availability of concessional
financing and on prívate sector participation. Prívate sector participation as well as loans from
Internal Financial Institutions may demand govemment guarantees. Refraining from providing
guarantees would entail developmental costs. On the other hand, generous and imprudent
guarantee policy would generate moral hazard in the markets and fiscal threats to the
govemment.
In the médium term, the govemment faces a trend of increasing pensión and health
spending, possibly averted by the proposed reforms. Both pensión and health reforms, however,
may generate significant transition cost. With reform, the govemment is likely to face around 2
percent of GDP of increased déficits on its pensión account by 2001. Without reform, pensión
déficits are projected up to 2.7 percent of GDP. Total health expenditures are expected to
increase from 4 to above 6 percent of GDP during 1999-2001. With reform, a newly established
Health Insurance Fund is expected to mobilize about 1.5 percent of GDP directly from
individuáis and employers (offset by a reduction in funds collected by the National Social
Security Institute). These reforms, however, will alleviate the state budget only after better
institutional capacities to collect pensión and health contributions are in place. Currently several
institutions are involved in the collection of taxes and social contributions and the govemment
wants to establish by mid-2000 a single revenue collection agency centralizing collection and
control functions.

7

As for contingent liabilities, unlike in most other EU accession countries Bulgaria's risk
exposure is relatively modest, though not unsubstantial. The government has applied prudent
limits and regulations for state guarantees. Presently, the government reports the stock of
guarantees of about 17 percent of GDP. Of these obligations, 9 percent of GDP is de facto direct
debt owed to the IMF. Government obligation for almost 4 percent of GDP of domestic
guaranteed debt expires in April 2000. Calis on the remaining guarantees, which currently
amount to about 4 percent of GDP, however, could cause uncomfortable fiscal losses during
2000-2004. In addition, for private sector development purposes, the government guarantees
obligations of agencies, like the State Fund for Agriculture, the Export Insurance Agency, and
the Business Promotion Bank. These are small so far. As Bulgaria's creditworthiness remains
weak, also the country's private non-guaranteed debt is veiy low.
Pressure on the government to provide guarantees and other forms of off-budget support,
through the various existing and new state-guaranteed agencies, is likely to increase as the
economy recovers, private investors substitute for public investment in agriculture, commercial
banks continué to abstain from providing long-term credit, and foreign creditors fear of
uncertainties.
In thefínancial sector, the 1996-97 hyperinflation, crisis, and subsequent policy actions
by the Bulgarian National Bank have effectively cleaned up the banking sector. The ensuing
reform has successfully capped both explicit and implicit government obligations for lost
deposits and failed banks. Since then, an improved supervisión has controlled exposure of banks
to liquidity and interest rate risks. The quality of bank's loan portfolio and foreign exchange
exposure cause concern but not a significant fiscal risk so far.
Government institutional arrangements for managing fiscal risks are strong in many
aspects but not totally reassuring. Parliamentary scrutiny over contingent liabilities and public
disclosure of their aggregate levels is good. Future fiscal pressures may, however, arise from
legal loopholes or from the government practice to cover all risks under eveiy guarantee, which
insures both creditors and debtors against their possible failures. Finally, the government
capacity to analyze, mitígate and manage risk needs to be enhanced. The government is
presently not fully aware about the size and urgency of its risk exposure. And, without
implementing new approaches, such as public-private risk sharing mechanisms, contingent
liabilities may have negative consequences for the markets as well as for future government
budgets.

2. Why Fiscal Vulnerability
What will be the government's capacity to respond to the realization of fiscal risks in the
future? Building on Hemming (1999), in this section we briefly sketch Bulgaria's fiscal position,
its sensitivity to risks, and sources of risks.

8

Current fiscal position appears good
Bulgaria's current fiscal position has been held in check following the introduction of a
currency board arrangement in 1997 3 . General government budget déficit dropped from 12
percent of GDP in 1996 to 2.5 percent in 1997. In 1998, the budget registered a surplus of 0.9
percent of GDP and in 1999, despite the slow-down of growth due to the Kosovo crisis, déficit is
expected to stay below 1.5 percent of GDP.
Recent revenue performance has been strong and resilient to the economic shocks that
have marked the 1998-99 period (the Russian and Kosovo crises). General government revenues
increased from 31.7 percent of GDP in 1997 to 36.8 percent of GDP in 1998 and an estimated 38
percent of GDP in 1999. General government expenditures, however, are also on the rise. For
1999, they are projected to reach 39.5 percent of GDP, up from 35.8 percent in 1998, recovering
from their collapse during the 1996-97 crisis. Leading this trend, social expenditures have
increased by 1.8 percentage point, while wages and contingency to pay for the cost of structural
reforms by 0.8 and 0.7 percentage points, respectively.
Beyond general government budgets, Bulgaria's fiscal position does not suffer from any
imminent major pressures either. The Currency Board Arrangement (CBA)4 precludes the
National Bank from engaging in quasi-fiscal activities. The number of extrabudgetaiy funds
declined from over 1000 in 1998 to 28 in 1999. We limit our analysis here to the largest ones, the
Pensión Fund, Health Fund and Agricultural Fund. Though these funds are rife with fiscal risks,
which we will discuss in the following section, their positions appear balanced so far.
Medium-term and long-term sustainability is at moderate risk. Bulgaria's fiscal
framework presented in Annex 1 as well as debt dynamics, which we will illustrate later,
suggests relatively stable baseline. Main fiscal pressures come from debt service payments and
transitional costs associated with the implementation of the health and pensión reforms. In the
baseline, debt service payments will remain substantial at about 20-220 percent of exports of
goods and services, growing to about US$1.4 billion per annum by 2004. The pensión déficit

Annex 1 presente Bulgaria'sfiscal framework for 1997-2002, including the main revenue and expenditure
items as well as the projected déficit and its financing.
Introduced after a spell of large déficits and hyperinflation in July 1997, the currency board arrangement carne
as a response to several unsuccessful money-based stabilization attempts and to widespread lack of financial
discipline of large state-owned enterprises andfinancial institutions as well as budgetary agencies. CBA was
meant to stop a vicious circle of government subsidies and soft commercial bank financing, that have kept lossmaking enterprises afloat. The arrangement has fixed the domestic currency, lev (BGL), to the Germán mark,
and prescribed full coverage of the monetary base with foreign reserves. Furthermore, it cut off central bank
credit to both the government and the banking sector. Since January 1,1999, the lev isfixed to the Euro at the
same rate as the DM peg to the Euro. In addition, the original rate BGL1000/DM was changed on July 5, 1999
when three zeros were removed, the denomination became BGN, and the peg 1BGN per 1DM. Under the June
1997 law, the Bulgarian National Bank is not allowed to extend credit to the State or any state agency except
against purchase of special drawingrightsfrom the IMF. Similarly, the Bulgarian National Bank (BNB) is
forbidden to extend credit to banks, except under very strict conditions in its narrowly defined role of lender of
last resort.

9

and health expenditures increase by 2 and 1 percent of GDP, respectively, between 1999 and

2001.
The govemment has committed to a fiscally responsible behavior. Under its IMF
supported program and with EU accession objective, the govemment intends not to run déficits
in excess of 3 percent of GDP and in the médium term reduce its debt to GDP ratio below 60
percent (thus comply with the Maastricht déficit and debt criteria). Furthermore, the govemment
aims at keeping the debt service below 25 percent of exports and 30 percent of fiscal revenues, at
pursuing market risk benchmarks in its debt portfolio, and at cushioning possible future shocks
with large reserves.
In line with the requirements of a currency board arrangement, Bulgaria has maintained
comfortable levels of both external (central bank) and fiscal reserves. External reserves have
exceeded the equivalent of six months of imports of goods and non-factor services. On its Fiscal
Reserves Account (FRA), which consists of the balances of all govemment budgetary and extrabudgetaiy accounts in the banking sector, the govemment maintains certain floor throughout the
fiscal year. This floor which exceeded 8 percent of GDP at the end of 1998 is the main
contingency instrument to address selected fiscal risks. The floor on the FRA can be adjusted to
accommodate larger than expected structural reform-related contingent expenditures, higher than
projected interest payments or shortfall of official financing relative to program projections5.
Large reserves impose, however, opportunity cost of investment and growth. As we will discuss
further, only major improvements in risk mitigation and risk management capacity would reduce
the reserve requirement, opening a way toward a more optimal mix of reserve and hedging
strategy and releasing resources for investment.

Room to accommodate fiscal risk is, however, limit ed
A combination of factors explains why Bulgaria's fiscal position is less resilient to fiscal
risks than it appears from the analysis of the current fiscal position. First, while the CBA is very
effective in achieving fiscal stability, it does by definition reduce the range of options otherwise
available for déficit financing, and therefore the scope for fiscal expansión or for accommodating
sudden expenditure hikes due to the materialization of unaccounted for fiscal risks. Out of the
four possibilities that are available to most countries in financing their public sector déficit
(printing money, running down foreign reserves, and foreign and domestic govemment
borrowing), the CBA rules out the former two. Foreign and domestic borrowing, along with
exceptional proceeds such as privatization revenues, thus remain the only means of déficit
financing and of raising money to face sudden shocks.
Following the CBA adoption, the main source of déficit financing shifted from the
domestic banking system (on a net basis) to
privatization revenues (chart 1). High
Chart 1: General Government Déficit Financing
negative net domestic financing in 1998-99
15.0 -i

In addition, the govemment allocates resourcek
the 1 > d :t for contingent expenditures. Arount 1
ug
percent of GDP in 1999 and 1.2 percent of GDíf
ver! allncated for possible calis on guarantees and
implementation problems of pensión and health
réfoGftis.
ISi

<

0.0
-5.0
1996

I Etra
xenl

1997

•Domestic

1998

1999 1/

0 Privatization

reflecte the amortization of past bonds. Net external financing also shows a negative transfer of
1.3 percent of GDP in 1997 declining to 0.7 percent in 1999. In 1998, the overall balance
registered a surplus equivalent to 0.9 percent of GDP. Privatization receipts contributed to 68
percent of the overall financing while net domestic and external financing were negative. A
déficit of 1.5 percent of GDP is projected for 1999 while privatization receipts are expected to
approximate 2.6 percent of GDP, comfortably covering the financing needs. Privatization
receipts cannot be used for current budgetary expenditures. They enter the fiscal reserve account
and can be used for debt repayments and investment financing. The privatization receipts of the
municipalities can be used for ecological projects, investment debt repayments or writing off
non-performing loans of municipality-owned enterprises.
As the privatization process comes cióse to an end, revenues from the sale of state-owned
enterprises will fall down. The largest and most profitable state-owned enterprises have already
been or are now being privatized. Further sizeable revenues could be expected from the
privatization of BTC (Bulgaria telecommunications company), Bulgartabac (tobáceo company),
Bulbank (the largest state-owned bank) and several power distribution companies in 2000-01,
after which the scope for raising substantial revenues from privatization shrinks. This may raise
questions about the availability of resources for debt payment and investment financing.
Second, in responding to shocks, the government faces a constraint on both the revenue
and expenditure side. With revenues at about 38 percent of GDP in 1999 and a bias toward
payroll tax, further increase in tax rates would damage investment and growth. Actually, the
main fiscal policy objective is to broaden the tax base and strengthen collection in order to lower
tax rates on labor and income.
Social security contributions rates have remained high at (depending on workers
categories), 49.7, 44.7 and 34.7 percent of gross wages, paid by the employers, with additional 1
percent (deductible from the personal income tax) paid by employees. The combined corporate
tax rate arising from profit tax and municipal tax is 34.5 percent (a reduced 26-percent rate
applies to small enterprises). The valué added tax rate is 20 percent. Social insurance
contributions generate about 8.5 percent of GDP, similarly as the value-added tax. Non-tax
revenues, such as municipal fees, various levies, and income from rented properties generate
additional 6 percent of GDP. Reflecting trade liberalization, excise and custom duties account
for less than 5 percent of GDP. Positive institutional developments, such as improvement in tax
administration have contributed to stronger revenue performance in most areas.
The structure of government expenditure is rigid. At around 39 percent of GDP in 1999,
government expenditures are dominated by social protection programs, debt service, and wages.
Pensions and other social outlays account for 13 percent of GDP and wages for over 5 percent of
GDP. Interest payments now exceed 5 percent of GDP. Necessaiy maintenance and operating
expenditure are projected at around 5 percent of GDP. This comes up to about 30 percent of
GDP, leaving about 7 percent of GDP in total for defense and capital expenditure in 1999.
Experience of the last years has shown that in case of emergency the government avails of
instrumente for a temporaiy reduction of spending. The annual budget law provisions impose

11

consistency between spending and revenue flows- in the course of the year only 90 percent
disbursement on expenses is allowed, only if revenues perform as expected, is spending allowed
to increase up to 100 percent of the Budget Law. The law also explicitly states the priorities in
case of lower than expected revenue flows. While these provisions appear effective to reach
déficit targets, expenditure cuts, as a possible response to shocks, or to accommodate higher
investment levels will be difficult to make without a thorough and detailed analysis to prioritize
expenditure and identify efficiency gains.

II. Analysis of Individual Sources of Risk
This section evaluates fiscal risks that emerge in the form of direct and contingent, both
explicit and implicit liabilities of the central govemment. In this context, we define risk broadly,
as elements involving uncertain fiscal cost (henee contingent liabilities for instance) and as
variability in the effective cost of govemment obligations and in the associatedfinancialflows.

1. Direct explicit Risk
Sovereign debt poses medium-term fiscal risk
Trends in govemment debt appear reassuring.
Until recently, Bulgaria's govemment debt was a major
obstacle to growth and investment. Extemal debt
exceeded US$10 billion, or 57 percent of GDP, in 1990.
As the cost of debt service had risen above Bulgaria's
perceived capacity to pay, the govemment announced a
unilateral moratorium on debt service payments. In the
following years, output collapsed, and sovereign debt to
GDP skyrocketed to 150 percent of GDP. Restructuring
of debt to the París Club in 1992 and London Club in
1994 eased the extemal debt burden by 33 percent of
GDP. Domestic debt, however, had been rising swiftly, as the govemment borrowed to prepare
state-owned enterprises for privatization, and to bail out failing banks. Bulgaria was unable to
access international capital markets, and its official lending has been limited ever since. Striving
to retum its debt below the Maastricht threshold of 60 percent of GDP, the govemment generated
sizeable primary surpluses of 6.4 percent of GDP annually over 1994-1999, reducing its debt to
82 percent of GDP in 1998 (chart 2). This level on gross basis, as well as 59 percent of GDP on
net basis is still among the highest of the EU accession countries.
As Bulgaria had to regularize its relations with its extemal creditors, it had to cope with a
negative transfer of resources, that averaged 5 percent of GDP during 1994-1998. Meanwhile,
the budget déficit was financed from domestic sources. Domestic déficit financing averaged 6.7
percent of GDP annually during 1994-1995, peaking at 14.9 percent of GDP in 1996. Investors'
confidence eroded quickly, and average maturity of the of Treasury bilis collapsed from ten
months to mere two weeks at the peak of the crisis.

12

Introductíon of the CBA brought
confidence back.
The burden of interest
payments eased to around 5 percent of GDP, but
foreign debt service is projected to remain
substantial6. In absolute terms it is expected to
exceed US$1 billion in 2001 and stabilize at
about 21-22% of exports of goods and services
(Chart 3). This level is uncomfortably cióse to
the empirical crisis threshold of 25 percent and
will continué to constrain Bulgaria's flexibility
in the use of debt financing to dampen fiscal
shocks.
Since the introductíon of the CBA, the government has been cautious in its borrowing
policies. It prudently refused repeated offers from the commercial creditors in the Eurobond
market in 1998. Aftershocks of the Russian and Kosovo crises that have shifted the yield curve
substantially upwards in 1999 prove that this prudent stance was well grounded. Net financing
from the domestic market has been negative since mid-1998, while the external debt has been
reduced by two successful debt buyback operations in 1998-99 at deep discounts. The total face
valué of debt eliminated via buybacks is estimated at well over US$1 billion. Bulgaria was
rewarded for its responsible borrowing strategy by a 1999 rating upgrade (to B+) and by a steady
increase in the average maturity of the Treasury bilis from 13 to 21 months as of end-1999.
Debt structure has stabilized but remains rígid. The share of foreign debt dropped briefly
after the Paris and London Club restructuring deals. Rapid erosion of the valué of domestic
currency during 1995-1997, however, wiped out the lev-denominated debt and brought the
external debt burden back to unsustainable
level of 90 percent of GDP. The 1998 share
Chart 4 Structure of Sovereign Debt
of domestic debt was 17 percent of the total
public debt, with about two thirds (mainly
associated with the realization of past
contingent
government
liabilities)
denominated in the US dollar. The remaining
one third of the domestic debt is in Treasury
bilis (Chart 4). The share of foreign and
domestic debt denominated in US dollar is
about two thirds of the total debt in 1999.
**
18%
Brady bonds domínate the foreign debt
portfolio followed by debt owed to
international financial institutions. Most debt
instruments have long maturity. Current financing constraint makes it difficult to change the
debt structure and improve the risk structure of the debt portfolio.

Projections include new identified borrowing and gapfill financing. Source: IMF staff estimates.

13

Integrated asset and liability framework reveáis substantial exposure to severa
Integrated asset and liability management seeks to handle in a comprehensive fashion all three
main forms of market risk facing govemment: (a) refinancing risk, (b) interest rate risk, and (c)
foreign exchange risk. In a financial portfolio approach we compare the risk structures of
financial flows and the stocks of financial assets and liabilities on the two sides of the
govemment balance sheet. 7 We define our approach rather narrowly, focusing on govemment
direct debt portfolio and central bank reserves. We do so because the direct debt portfolio is very
large compared to contingent liabilities, and central bank and fiscal reserves are large compared
to other financial, existing and potential, assets. In addition, we take into account the lev's peg
and the country's trade flows since these drive Bulgaria's capital account developments.
We consider refinancing risk first. Refinancing risk can be defined as the volatility of a
sovereign's access to liquidity and to sources of debt financing over a longer-term period. As the
1997-98 emerging-market crisis illustrated, Bulgaria again may face prohibitively expensive
market access in the future. Moreover, Bulgaria's ability to refinance domestic debt is also very
limited, since the domestic capital market remains rather shallow.
In this context, we evalúate refinancing risk by the overall maturity of sovereign debt
portfolio and by the volatility of its repayment profile. Maturity structure seems reassuring.
Brady bonds and debt to international financial institutions mature in 16.5 and 10 years,
respectively, bringing thus average foreign debt maturity to above 13 years. The dollar
denominated domestic debt is also long-term. Maturity of the outstanding Treasury bilis has
reached 21 months by end-1999. However, this maturity structure is rigid. Neither restructured
London Club debt ñor the debt to international financial institutions can be rolled over. Net
Treasury bilí financing has been negative since mid-1998, issuance small, and auction
placements consistently below offer volumes. In addition, as we will discuss further, a
significant amount of Bulgaria's govemment debt has a floating interest rate and thus makes the
repayment profile volatile. Refinancing risks will increase as debt service rises in the médium
term.
Evaluation of the currency risk is simplified under the CBA. The CBA credibility appears
strong. Thus we treat the Euro-denominated debt not much different from domestic debt.
Bulgaria's cross-currency risk boils down to the US dollar/Euro risk, as the share of other foreign
currencies is modest, both for the assets and liabilities. Bulgaria's natural hedges to currency risk
in its govemment debt portfolio include foreign exchange reserves, mainly denominated in the
Euro, and current account cash flows, essentially net exports of goods and services, mainly
denominated in US dollars, with a déficit emerging in the Euro-denominated trade.

This approach skips over possible valuation changes driven by the exchange rate movements in the
government's real assets and liabilities. A MI balance sheet approach, incorporating real assets and liabilities,
is methodologically more attractive, but more difficult to implement. First, there usually exists a large maturity
mismatch between the asset and the liability sides, since government's real assets are normally longer-term than
itsfinancial liabilities. Second, the usefulness of financial instruments to hedge currency risk in the
government's real assets and liabilities is limited since markets to hedge long-term risks are not well
developed. (Hansen, 1998)

14

With respect to foreign exchange reserves, we observe a large mismatch between the bias
toward the Euro in their structure and the bias toward the US dollar in the structure of Bulgaria's
sovereign debt. In stylized terms, we present a simple combined balance sheet of the government
and the central bank in table 2.

Table 2 Simnlif ed balance sheet as o 12/31/98
(bilí BGL)
BNB (issue dep>
MOF
BGÜ USq| EUR BGÜ USDl EUR
258 4162
Currency reserves
-651 -258 -103
651
258 103
Gov. Account
Gov debt
-10709 -176'
-651
0 4059
651 -104^.1 -1658
Central bank's exposure: euro

Net exposure
BGÜ USDl EUF
0
258 416Í
0
0
0
0 -10709 -176'
0 404512400.6

Sovereign debt: US$

Source: World Bank staff estimates

With respect to the current account flows, Bulgaria's large share of foreign and domestic
debt denominated in US dollar broadly satisfies general hedging objectives. This holds, even
though, the lev is pegged to the Euro, making Bulgaria's exposure to currency risk determined
primarily by the US dollar/Euro exchange rate. Indeed, strengthening of the dollar since October
1998 (Chart 5) had so far cost Bulgaria's budget an extra $80-100 million in debt service cost. In
the current account, however, this loss was mirrored by increases in the dollar-denominated
export revenues. Chart 6 compares currency composition of net merchandise exports and debt
service payments in 1998 and shows that the US dollar exposure on the sovereign debt was
broadly matched by the dollar inflows on net exports. The peg to the Euro serves as a natural
hedge to Bulgaria's large déficit in Euro-denominated trade.
Chart 5. Daily exchange rate, Lev/$1

C h a r t 6. Natural h e d g e f o r c u r r e n c y risk, 1998
USD million

1.95

400

1.85

1.75

360.87

USD

BIT
[75.28)

SH
DH
( 7 a 2 9 )

»

^46)

(113.44)

(400)
1.65

(431.73)
(563.44)

Jan-98 Apr-98

Jul-98

Oct-98 Jan-99 Apr-99

Jul-99

Oct-99

• net exports • debt service

(800)

Source: World Bank staff estimates.
The natural hedge on the dollar, however, is not complete and is on the wane. In 1998, the
mismatch was some 20 percent of the dollar-denominated exports of US$70 million; in 1999 it is

15

estimated to over 1/3. Chart 6 exeludes dollar exposure on the dollar-denominated domestic
debt, which constitutes another 10 percent. And there is a large quarter-on-quarter variance in
the currency structure of debt service against foreign trade, exposing the country to short-term
currency risk.
Interest rate risk in public debt portfolio is a more serious reason for concern. The share
of floating rate debt in both external and domestic debt portfolio was over three quarters at end1998. The drop in LIBOR over 1997-1998 made this to govemment's advantage, lowering its
debt service costs. Possible increases in LIBOR in the future, however, will reverse this trend.

Chart 7. Stress testing: the impact of higher interest rates on the ratios of...
debt service-to-exports...
...and debt service-to-budget revenues

For external debt, the risk of losses on higher interest rates is significant. The relationship
between higher interest rates and debt service is exponential, as the interest costs will open wider
the financing gap in the future years. Stress testing the debt service profile with an interest rates
increase of 1%, 2% and 3% (see chart 7), we find that higher interest rates would result in the
debt service to exports ratio to be on average higher than the baseline scenario by 1.5-5.2 percent
in 2000-2004 8 . Any increase of international interest rates in excess of 3% over the current levels
may lead to unsustainable levels of debt service. The fiscal impact of additional interest
payments over 2000-2004 provides more comfort under the baseline scenario which assumes
continued strong fiscal performance. Nevertheless, higher international interest rates would add
from US$ 95 million (for a 1 percent increase in interest rates) to US$ 340 million (for a 3
percent increase) to the annual budget expenditures, or 1.5-6.5 percent increase in a debt service
to revenue ratio in 2000-2004.
Domestic interest payments represent the equivalent of a mere quarter of the external debt
interest payments, but they are more volatile than the latter. As the domestic fixed income
The baseline scenario is based on the Ministiy of Finance assumptions about the future track of international
interest rates.

16

market is shallow, it is extremely sensitive to shocks. In 1999, a combination of factors caused a
rise of yields on domestic debt by 4-5 percentage points, while the effective cost of Bulgaria's
external debt had risen by less than 1 percent. In a not unlikely scenario, an increase in domestic
interest rates in parallel with international rates may raise government debt service costs by 1-2
percent of GDP or 2-6 percent of the 1999 budget revenues. For Bulgaria, the current
predominance of instruments with floating rates is thus quite risky. In the short run not much can
be done about it, as the government continúes with its prudent borrowing policy. Once this policy
bears its fruits in terms of improvement in sovereign risk rating and decline in borrowing costs,
new borrowing decisions should aim to contain the interest risk in the debt portfolio by
contracting debt with fixed interest rates. At that time, a sensible interest-rate benchmark for both
external and domestic components of the debt portfolio should be constructed.
Risk of derívatives, as volatility in the cost of derivative instruments, is negligible.
Similarly to other sovereign borrowers with sub-investment grade ratings, Bulgaria faces
difficulties in accessing derivative markets and finding the needed counterparties. Bulgaria
would benefit particularly from interest rate swaps and currency swaps to smooth the fiscal
impact of exchange and interest rates volatility. But possible availability of derívatives also
makes the development and application of asset and liability management framework for risk
management more urgent.
Institutional and transaction risk refers to possible failures in the debt management
processes. Since 1998, Bulgaria has been systematically strengthening its debt management
capacities. As one of the steps, the Ministry of Finance has Consolidated debt management into a
single Department for Government Debt. The department includes specialized units that manage
and track debt, guarantees, and on-lending.
Bulgarian National Bank (BNB), as the
Government's fiscal agent, conducís primaiy auctions of government securities and aims at
deepening the secondary market. Both the BNB and the Ministry of Finance have been
enhancing their debt tracking systems. The BNB debt monitoring system tracks both external
and domestic debt, but generates only external debt service projections. Debt database of the
Ministry of Finance covers state guarantees as well as the entire government debt, but its
projection and reporting engines are under development.
Both institutions lack tools to analyze debt sustainability and to assess risks of their asset
and liability portfolios. There exists no stractured procedure to assess and compare the risk
exposures, and decisions are thus made with little reference to any risk management objectives.
For example, servicing the dollar-denominated debt from the Euro or lev accounts alters the net
combined exposure of the fiscal authorities and the Bulgarian National Bank, thereby adding to
Bulgaria's fiscal vulnerability. As a hurdle in the development of further institutional capacities,
both the Ministry of Finance and Bulgarian National Bank suffer from high staff turnover.
The system for transaction processing remains complex, with several overlapping
procedures in the Ministry of Finance and the BNB. It has been successfiil in preventing any
payment slippage since 1994, but in some instances, erroneous decisions with regareis to
accounts and source of funds used for debt servicing have increased debt servicing costs.

17

Streamlining the transaction processing within the overall Treasury management framework is
therefore strongly warranted.

The proposed reform would bring social security under control
Bulgaria's govemment legally guarantees an old-age pensión and other social security
programs. Social security guarantees are becoming fiscally expensive. Bulgaria has 2.4 million
pensioners and 3.2 million insured in the mandatory social security system. The oíd dependency
ratio, as the share of population aged 60 plus to those between 15-59, has surpassed 0.43. The
system dependency ratio, as the share of pensioners to the working population, has exceeded
0.75. Social security programs are exerting increasing pressures on the budget, rising from 9
percent of GDP in 1997 to 13 percent in 1999. Social security programs include: (a) retirement
and various social pensions, (b) unemployment and one-off retraining benefits, (c) various social
assistance programs for specific groups (students, families with numerous children, and others),
and (d) social benefits paid through the municipal budgets, but subsidized from the central
budget, for the uninsured and poor.
The govemment spends 9 percent of GDP on old-age and social pensions. The pensión
system suffers from arrears in contributions (200 million BGN, which is about 1 percent of GDP,
as we will discuss in the enterprise sectíon of this paper) but not from any arrears in benefits
payment. With no reform, expected demographic trends threaten to generate cash déficits in the
pensión system of nearly 2.7 percent of GDP by 2002 (table 3). To contain these expected
increases, the govemment proposes to downsize the existing pay-as-you-go, improve options for
voluntary prívate pensión saving, and introduce mandatory defined contribution professional and
employer saving plans. Proposed changes into the present pay-as-you-go system include a
gradual increase in the retirement age and its convergence for men and women , transfer of the
responsibility for early retirement to professional pensión funds, and adoption of universal
mandatory pensión contributions based on salary levels. On the front of other social security
programs, govemment particularly aims at restricting access to short-term social security
payments.

Table 3 Projected Social Security Expenditures

Projected pensión
Projected pensión
Total social security Total social security
déficits: w/o reform
déficits: w/reform
spending w/o reform Spending w/reform
BGN, mil % of GDP BGN, mil % of GDP BGN, mil % of GDP BGN, mil % of GDP
489
2000
2.1
492.9
2.1
2081
8.9
2041
8.7
2001
648
2.6
476.2
1.9
2405
9.5
2176
8.6
2002
744
2.7
524.8
1.9
2700
9.9
2408
8.8
Source: NSSI
Under the reform scenario, projections of govemment pensión expenditures look more
reassuring, with incurred cash déficits around 2 percent of GDP. Projections shown in table 4

The present retirement age (55 for women and 60 for men) will be preserved until 2003. After the year
2003, the retirement age for men will grow by 4 months each year until it reaches the age of 65. The retirement age
for women will grow by 6 months each year until it reaches the age of 63.

18

assume the envisaged increases in the retirement age, reductions in pay-as-you-go pensión
benefits, but also higher compliance, resulting in a cash déficit of around 1 percent of GDP
annually, for the next five years.

Table 4. Projected Public Pensión Expenditures After Reform

Annual Cash Deficit/Surplus
Total Outlays
BGN.mil
Years
% of GDP
BGN, mil
% of GDP
1999
2173
9.5
-114
0.5
2000
2315
10.1
-296
1.3
2001
2432
10.0
-262
1.1
2002
2550
9.8
-240
0.9
2003
2667
9.7
-238
0.9
2004
2743
9.4
-341
1.2
2005
2780
9.0
-261
0.9
2006
2802
8.6
-130
0.4
2007
2798
8.2
45
0.1
2008
2798
7.9
169
0.5
2009
2802
7.5
309
0.8
2010
2848
7.4
420
1.1
2011
2877
7.1
796
2.0
Source: Agency for Economic Analysis and Forecasting
Henee, according to the above calculations provided by two different sources in Bulgaria,
déficit from pensión and other social security benefits could range between 1 and 2 percent of
GDP per year in the next 12 years. The state has provided no guarantee or commitment under the
planned pillar 2 and 3. Therefore, no explicit fiscal liabilities are expected from these programs
in future.10

Health financing is likely to grow, either directly or through contingencies
Similarly to social security, Bulgaria's govemment also legally guarantees its citizens'
access to healthcare. Health care has been available free of charge to the population, costing the
govemment budget around 4 percent of GDP. Since population aging often has a stronger impact
on the cost of health care than pensions, the budget may expect increasing pressures. In this
respect, health-financing obligation of the state is of direct nature. The govemment, however,
has been considering a reform that would reduce direct budgetary health outlays, but introduce
another, contingent portion of public financing in the new health system. Law on Health
Insurance, which has established a National Health Insurance Fund to partly cover health care
cost, formulates an obligation for the state to cover the Fund's possible fiiture déficits as well as
to pay contributions of all of the state employees as their employer. Table 6 shows expected
sources of healthfinancingin the reformed system.

In the early stages of the new program design, we found it impossible to estimate any possible implicit
liabilities that might arise from pillars 2 and 3.

19

Table 6

Health Fund: Expected Revenues

Health insurance contributions
o/w from government
State budget
Total
Source: Ministry of Finance.

1999
BGNmil % ofGDP
257
1.2
97
0.5
806
3.6
1063
4.8

2000
BGNmil % ofGDP
577
2.4
213
0.9
840
4.0
1417
6.4

2001
BGNmil % ofGDP
650
2.5
231
0.5
1,035
4.1
1,685
6.6

Contingent government liability arising from the new arrangement will realize, should the
NHIF face difflculties in collecting health contributions. Collection difficulties may arise from
the complexity of the Fund's contribution structure. Contributions will be levied on the higher
income family member, with additional amounts imposed for dependents and other non-wage
earners. The system wotdd require detailed monitoring of family relationships and incomes.
Collection and enforcement problems are particularly expected vis-a-vis non-cash compensations
and self employed. Furthermore, the Fund is designed to contract health services with health
providers. Thus, government's obligation to cover the Fund's déficits is likely to erode
incentives in the entire health system and to become costly, also for the state budget. Initially
envisaged for mid-1999, actual payments under the new system have been postponed to mid2000 in view of the above-mentioned implementation problems.

2. Direct implicit Risks
Public Investment Program may generate significant fiscal pressures
Bulgaria's Public Investment Program, first developed in 1998 for years 1998-2001,
offers many positive features but also suffer from weaknesses experienced earlier by other
countries. Particularly, Public Investment Program does not include estimates of recurrent and
maintenance cost implications of proposed investment programs. This shortcoming may lead to
unexpected pressures on the budget in the future and, consequently, to disruptions in further
investment programs as well as in the provisión of expected services. In addition, risks
associated with Bulgaria's investment program are three-fold: (a) project and program quality,
(b) availability of expected financing schemes, and (c) counterpart requirements.
Public Investment Program covers planned capital expenditure by the state budget,
budgetary spending agencies, large state-owned enterprises, and funds, such as National
Environmental Protection Fund, Republican Road Network Fund, National Telecom System, and
Air Traffíc Management. The PIP is an integral part of the organic budget law of the current year
and indicative for the following years. Over the last three years public investment expenditure
increased from 1 percent of GDP in 1997 to 3.5 percent in 1999. The Government program for
2000-2006 fixes public investment at 3.5 percent of GDP per year. This limit may come under
significant pressures from competing requirements of
infrastructure rehabilitation,
environmental clean-up, institutional capacity building, restructuring of social services, and
poverty alleviation programs, all key to the country's development and growth and/or EU
accession.

20

Project quality risk mainly emerges from capacities of the sectoral agencies, which are
responsible for economic and financial analysis and for the overall soundness of their proposals.
The agencies' capacities, however, vary greatly and most agencies suffer from tight human
resource limitations. Similar problems are facing the Ministry of Finance, which is responsible
for review and prioritization of proposals prior to submitting them to the Council of Ministers.11
Thus, project proposals are not necessarily well designed and well assessed. Coupled with the
fact that PIP does not include recurrent and maintenance cost implications of investment
programs, this may cause that future pressures on the budget exceed expectations.
Financing risk, risk that pre-identified sources of financing fall short, is also real. The
1998-2001 PIP envisages a US$3.9 billion investment program for the public sector. Out of this,
US$2.6 billion belongs to state-owned enterprises. The state is expected to guarantee credits to
finance over half of the investments. As we will discuss further, many large state-owned
enterprises are incurring losses and accumulating arrears.
Unless these enterprises are
restructured to become commercially and financially viable entities, state guarantees on their
credits may produce significant contingent fiscal risk for future state budgets.
Counterpart requirements mainly relate to Bulgaria's investment-related borrowing from
International Financial Institutions and EU accession funds. EU programs, such as SAPARD
(Special Accession Program for Agriculture and Rural Development), require government to cofinance between 25 and 50 percent of investment cost. These requirements do not produce fiscal
risk by themselves but require attention in order to have the needed room in future budgets.

3. Contingent explicit risks
State guarantees are not large but should be treated with caution
In 1999, Bulgaria has reported about 17 percent of GDP of outstanding state guarantees
(table 7). This amount includes 9 percent of GDP of IMF debt that is intermediated by the
Bulgarian National Bank. There is no doubt that state budget will cover IMF debt repayment.
Thus, in our framework we consider IMF debt as a direct rather than contingent liability of the
fiscal authorities, and thus analyzed it as a part of sovereign debt above.
Table 7

Bulgaria : Publicly Guaranteed Debt Outstanding (end of 1998)
Amount
As % of GDP
External (US$ million)
1596
13.0
IMF
1114
9.1
Others
482
3.9
Domestic (BGN million)
840
3.9
Source: Ministry of Finance a d Bulgarian National Bank (Monthly Bulleün, December 1998)
n

Selection criteria are established by the Council of Ministers to emphasize: (a) compliance with the
governmental development strategy, (b) contribution to economic growth and employment creation, (c)
regional, social and ecological priorities, (d) reduction in the amounts of incomplete construction, (e)
mobilization of external funds, (f) project readiness, and (g) administrative readiness and implementation
capacity. The PIP is now expected to be closely linked with the new, seven-year national strategy.

21

Non-IMF state-guaranteed foreign debt, reaching almost 4 percent of GDP, is dominated
by official creditors on the financing side (99 percent) and infrastructure state-owned enterprises
(92 percent) as the debtors (see Annex 2). The main source (95 percent) of domestic guarantees
are obligations of (that is deposits in) the State Savings Bank. For this Bank, the guarantee fully
covers deposits denominated in domestic currency until April 2000. After this date, these
deposits become subject to the Deposit Insurance Law, which provides for limited coverage
under the Deposit Insurance Fund.
Excluding the IMF debt and
assuming a run on State Savings
Bank's deposits before April 2000
unlikely, the amount of outstanding
guarantees comes to some 4 percent of
GDP. Compared to some other EU
accession countries, such as the Czech
V
Republic (Brixi and Ghanem, 1999),
this amount is not excessive. But most
^
- •
•
,
of these guarantees are risky. Most of
the state-owned enterprises, which are
their main beneficiaries (covered
Sourc:e: World Bank staff estimates
debtors), have been incurring losses
and accumulating arrears in taxes, social security contributions, and/or wages, and/or to their
suppliers.12 Prior to the end of 1998,13 domestic and two external guarantees have been called.
Expanding our ALM approach to cover contingent liabilities, we mainly observe that enterprise
debt covered by state guarantees largely has floating interest rate and is denominated in foreign
currency. An increase in interest rates and appreciation of the dollar vis-á-vis the Euro will make
the guarantees more likely to be called as well as will raise the budgetary cost of each guarantee
called. Chart 8 shows future possible budgetary cost of the outstanding external guarantees
(excluding the IMF guaranteed debt) according to different default risk levels. The maturity of
enterprise debt is relatively short, thus front-loading the fiscal risk of the existing guarantees.
Under a 60-percent default scenario, the fiscal loss would range between 1.2 and 1.4 percent of
budget revenues in 2000-2004. 13

Chart 8 Fiscalriskof guaranteed debt
n f

h n H o p f

r p v p n u p c

2 . 5 % -

2 . 0 %

1 . 5 % '

- • — 2 0 %
- « - 4 0 %

1 . 0 % ]

6 0 %

0 . 5 % -

8 0 %

0 . 0 %

•i~"

1 9 9 9

—

1

2 0 0 0

—r-

—

••

2 0 0 1

i-

1

—

2 0 0 2

,

i

2 0 0 3

1

•i

2 0 0 4

Beneficiaries of state guarantees include such continued loss-makers as the Bulgarian Railways, Bourgas Port,
and Maritza Iztok mines. Other large loss-makers with obligations covered by the state, such as Zarneni Hrani,
are in the process of shutting down.
In Hungary and the Czech Republic, state guarantees covering allrisksunder the contract have been called
with roughly 20 and 40 percent probability.

22

The demand for new guarantees is high emanating both from official donors and the
prívate sector. Box 1, below illustrates the case of the energy sector.

Box 1 Bulgaria: Fiscal risks associated with the energy sector
Fiscalrisksarise on account of state ownership of energy enterprises and the responsibility of the Government to
guarantee, either directly or indirectly, the obligations of state-owned enterprises (SOEs). The main sources of
existing and expectedfiscal riskassociated with Bulgaria's energy sector include: (a) non-payment/non-collection
for energy services, resulting in accounts payable; (b) inadequate tariffs to cover reasonable supply costs, resulting in
the need for direct budget support or a cross-subsidy between consumer groiips; (c) non-competitiveness of coal and
briquette enterprises, resulting in either direct budget subsidies to mitígate the social impact of mine closures or
cross-subsidizatíon by power companies; (d) implicit state guarantees for loans to energy SOEs (owing to state
ownership or control of enterprises); (e) explicit state guarantees for loans to energy SOEs; (f) explicit or implicit
state guarantees of long-term take-or-páy contracts and other obligations. Fiscal risks associated with (a)-(d) are
relatively well understood and addressed. Fiscalrisksassociated with guarantees to SOEs loans and take or pay
contracts are growing rapidly and will be the most challenging to manage and to mitígate. Investment requirements
are being driven by EU Accession requirements (nuclear safety, environmentál performance, and technical
compatibility), external and domestic demand considerations, and reúábility and efficiency improvements in existing
electricity and heat services to underpin economic growth and social protection.
. Fiscal risks associated with (e) and (f) are likely to increase rapidly if investments to modemize energy infrastructure
and meet EU Accession requirements are not carefully selected or properly structured. While the Government
rightly intends to maxinüzé the use of prívate capital to minimizefiscal risks,on-going sector restructuring and
privatization should ensure that prívate investors and operators increasingly assume all commercial risks. State
guarantees should only be considered for sound projects and where they are needed to secure prívatefinancing for
projects without which the country's economic growth or social condition wodld be jeopaidized.
With regard to (e), in 1999-2000 state guarantees have been provided or are being considered for investment loans
of about $850 million to be implemented in the 2000-[2004] period. These include investments in nuclear plant
safety, waste disposal, and plant upgrade ($380 million), electricity transmission and dispatch($150 million), district
heating ($120 million), and expansión of gas transit capacity to Turkey ($47 million). The implied level of energyrelated guarantees approximates the amount of overall investment guarantees at end 1998.
With regard to (f), an explicit state guarantee has been provided for the gas supply contract between Bulgargaz and
Gazprom, signed in May 1998, which requires Bulgaria to take or pay for pre-detennined annual volumes of gas
tmtil [2010]. The contracted volume for 1999 (4.0 billion cubic meters) is valued at $320 million. A slightly higher
volume is contracted for 2000. The contínúed decline in demand is likely to result in an obligation to pay against the
part of the contracted volume that is not taken.
Implicit or explicit state guarantees are also being considered for long-term take-or-pay power parchase agreements
(PPA) between the state-owned National Electricity Company (NEK) and privatized power producers. The first
PPA, to support a $400 million rehabilitation project (840 MW), would impose an obligation on NEK of about $180
núlUon/year .for 15 years in electricity off-take and fuel supply upon completion of the project Similarfy, the second
PPA to support a $1.0 billion new plant (600 MW) would impose an obligation on NEK of about $175 million/year
for 10-15 years. To the extent possible, these deals should be structured after reforms are more advanced and the
investors, not NEK, can take more of the market risk.

23

The government has established a simple framework for dealing with guarantees.
Particularly the government has developed a comprehensive register of guarantees (placed in the
State Audit Office as well as both the Ministry of Finance and Bulgarian National Bank) and
introduced regular publishing of the aggregate amounts of guarantees outstanding along with
government debt figures. The register covers all external and domestic guarantees, indicating the
beneficiary, creditor, project title, amount, currency, and debt repayment schedule. The
government also centralized the issuance of new guarantees and subjected each new guarantee to
executive and legislative scrutiny associated with regular budget process. Each guarantee
request, accompanied by opinion of appropriate line ministries, has to be reviewed by the
Ministry of Finance. With opinion of the Ministiy of Finance and possible input of the State
Audit Office, each guarantee needs an approval by Council of Ministers and Parliament's
ratification.
In terms of nominal limits, Decree 482 of 1997 set the annual limit on guarantees (face
valué) issued at 20 percent of expected budget revenues. A recent amendment to this Decree,
however, dropped the complementary ceiling of 20 percent of GDP on the total amount of
guaranteed debt outstanding (replacing it with a flexible ceiling to be set in the budget process
each year). This change significantly expands the legal room for the government to issue new
guarantees. Moreover, relatively lax limits on guarantee amounts are accompanied by flexible
reserve requirements. So far, the government has followed prudent reserve policy, though not
underpinned by any rules or clear risk assessment practices. The state budget included reserves
in the amount of 50 and 20 percent of the total scheduled guaranteed debt repayment (normally to
be paid directly by the debtors) in 1998 and 1999, respectively. Furthermore, under guarantee
contracts the government always covers all risks, without analyzing their determinants, and the
full amount of debtor's obligation. Official creditors, providing concessional resources for
development projects and balance of payment support, domínate the Iist of creditors and require
full risk coverage by the government. If extended to commercial creditors, this practice has,
negative implications for market behavior, creating moral hazard on the side of debtors and
creditors. Finally, for guarantees as for debt management, government lacks capacities to
analyze and control risk. The capacities to design private-public risk sharing mechanisms are
also to be developed at the MOF. Compared with other advanced EU accession countries (see
box 2), Bulgaria's problems in managing guarantees appear smaller. As we have noted,
however, its need for caution is much greater.

24

Box 2

Managing contingent liabilities in the Czech Republic

The managementofcontingentliabilitiesin the CzechRepublic suffersfromtwo major weaknesses. First, the
institutional framework does not encourage adequate prioritization in the use of guarantees. Second, the framework
neither encourages pricing of guarantees ñor sets sensible nominal limits on their amounts.
First, the law authorizing guarantees contains no guidance on when guarantees are the appropríate public policy
instrument or ho w they should be compared to other forms of state assistance. "This shortcoming has created an incentive
for entities and spending ministries unable to secure the desired level of state assistance through the budgetaiy process to
seek state assistance outside of the budget process through loan guarantees. The additional state assistance provided via
guarantees is perceived as cost free and is not included in the overall financial support provided to a given sector.
...guarantees are often extended as a result of emergency and political pressure rather than as a result of a strategic
decisión to support outcomes..."
Second, in the absence of systematic budgeting, one way to contain contingent liabilities is to place limits on
their issuance. In the Czech Republic, the sum of expected payments on guaranteed loans in a calendar year is required to
be lower than 8 percent of the expected state budget revenues. This limit is too high. In Hungaiy, the limit is set at 1
percent of the face valué of the guaranteed amount.
The Czech system is ahead of systems in other countries in as much as procedures are in place to document the
contingent liabilities, to categorize them according to risk categories, and to analyze the government exposure. The
payments made under guarantees called are also recorded.
Source: Brixi and Ghanem, 1999, and Brixi, Papp and Schick, 1999.

Large environment damages need to be fínanced
Bulgaria's Environment Protection Law states that the state will not be liable for past
environmental damages of enterprises privatized before February 1999. In addition, to comply
with EU environment regulations, by 2015 Bulgaria will need to finance large environmental
improvements as well as adopt new legislation and institutional arrangements. The Ministry of
Environment and Water estimates the 1999-2015 cost of containing past environmental damage
and financing the required treatment of water supplies, management and disposal of solid and
hazardous wastes, and air clean up to come up to US$8.5 billion14. Assuming equal annual
amounts, the government will need nearly US$570 million (about BGN1 billion) annually. The
2000-2006 PIP envisages annual amounts of US$400 million.
Based on the above overall estímate, the Ministry of Environment and Water and
Ministry of Finance have developed a forecast investment program for ecological improvements
for 2000-2006 15 . The government estimates the total minimum required expenditures in 2000-06
around USD2.8 billion.
Assessment for general costs to be incurred in order to reach EU standards in treatment of water supply,
management and disposal of solid and hazardous wastes and air pollution clean-up during 1999-2015 prepared
by Halgrove Associates funded by E PHARE. This estímate is the most comprehensive one currently
U
available and covers the overall costs of containment of past environmental damage and implementing
programs to attain EU standards for pollution and emission control, including the possibility of closing the oíd
nuclear plants.
This investment expenditure program covers: (a) support to the municipal water treatment/sewage systems
programs; (b) past environmental damage containment expenses for 'large polluters' in mining and metallurgy,
chemicals and petrochemicals, oil refíning and petroleum distribution that will be privatized; (c) support for the
disposal of solid waste programs for the municipalities; (d) disposal of hazardous waste programs for closed
mines; (e) rectifying the hazardous sulfur emission from thermal power plants; (f) past environmental damage

25

How large obligations will guaranteed agencies be permitted to accumulate?
Business Promotion Bank, recently established with 95 percent government ownership, is
based on a Germán model and a recipient of a credit line on concessional terms from Germany.
Its main objective is to channel long term credit to small and médium size enterprises. The
government guarantees repayment of the Germán credit line as well as other foreign borrowing
by the Bank. The government does not guarantee its deposits. The Bank is not operating yet,
leaving, for the time being, therisk-exposureof the state near zero.
Bulgarian Export Insurance Agency was established in 1998 as a joint stock company,
with the Minister of Trade and Tourism representing the Republic of Bulgaria as the major
shareholder, to provide exporters with insurance policies to cover political and medium-term
commercial risks. The Agency's capital is about DM10 million. On its own account, the agency
will ensure export credits, guarantees and financial losses against short term credit risks. On the
government account and with full government guarantee, the agency insures against noncommercial and médium and long-term commercial risk. According to the Export Insurance
Law, a limit for the export credit insurance benefiting from Government support should be
imposed in the State budget annually. For 1999, the Parliament approved a DM30 million limit.
The coverage against short-term commercial risk protects Bulgarian exporters against permanent
insolvency and protracted default of the foreign buyer for up to 85 percent of the incurred loss.
The coverage against non-commercial risk protects the Bulgarian exporter against default
payment by the foreign buyer as a result of war, revolution, host government's moratorium on
debt, expropriation, nationalization and similar. The insurance covers no more than 90 percent
of the insured loss. The coverage against médium and long term commercial risk protects the
Bulgarian exporter against permanent insolvency and protracted default of the foreign buyer, up
to 85 percent of the incurred loss. On the government account, the Agency has also started an
export credit insurance against short-term non-commercial risk from prívate and public foreign
buyer. Demand for this as well as other policies is, however, still very limited, keeping the
state's risk exposure very low.
State Fund for Agrículture, under the Ministry of Agriculture and with its obligations
explicitly guaranteed by the state, provides financial support to the agricultural sector through
short and long term loans, credit lines, interest subsidies and guarantees. Credit lines channeled
through commercial banks support long-term agricultural investment.
The number of
commercial banks participating in this program has increased from 2 in 1998 to 11 in 1999.
Through an irrevocable order of payments issued by the commercial banks in favor of the Fund,
credit risk appears fully assumed by the banks. The State Fund for Agriculture charges the
commercial bank a 2 percent interest rate, the banks charge the producers 6 percent, and the
remaining 4 percent cover the bank's margin. In the context of market interest rates, the implicit
subsidy to the producer is about 10 percent of the credit amount outstanding annually. This
program is growing rapidly from 150 projects approved for BGN 11 million in 1998, to 350
containment investments for the non-privatized ore mines (coal) and liquidated/closed entities; and, (g) some
estimates for compensating the land-owners for recultivation of the contaminated land around the industrial
sites. It does not cover the disposal of solid waste from closed uranium mines, phasing out the oíd nuclear
plant, and conversión of thermal power plant inputs from lignite to better quality coal.

26

projects and BGN 24 million credit in mid 1999. In addition to yearly allocation from the state
budget, this program is supported by a revolving fund financed from repayments by the
commercial banks. Direct support to agriculture producers is also provided through interest
subsidy and short-term credit. Support under these programs amounted to BGN 46 million in
1998 and was scaled down to BGN 35 million in 1999. These programs are being phased out
and expected to disappear by 2001. 1 6
While the short term support programs are subject to a yearly limit in the budget law, no
ceiling is imposed on the volume of long-term investment credit. In addition, a new instrument
is being introduced whereby commercial banks would extend long-term credit from their own
resources with a full guarantee of the State Agricultural Fund. It is unclear whether the Fund's
guarantees are subject to any specific limit. Finally, under the Grain Law, public warehouses will
be established and warehouse receipts could be pledged as collateral for commercial bank credit.
The State Fund for Agriculture is expected to particípate to this scheme through provisión of
guarantee or participation in an indemnity fund.

4. Contingent implicit risks
Enterprise obligations may turn out expensive
Enterprise restructuring has been slow in Bulgaria, and state-owned enterprises have been
accumulating losses and arrears. In 1999, arrears reached nearly 2 billion lev (around 8.5 percent
of GDP), concentrated primarily in about 90 largest enterprises. These include the Bulgarian
Railways, Bulgargaz (energy) and Neftochim (chemicals). Out of the total amount, about one half
are due in taxes and social security contributions and have led in the past to shortfalls in tax and
social security revenues. The other half is largely due to suppliers (mainly energy state-owned
enterprises), leaving around 2.3 percent in wages due to employees.
Future possible clean up of these obligations may negatively affect the budget as far as
the arrears to suppliers and enterprise employees go. Should the govemment be liable for tax and
social security contribution arrears, their payment from the budget would be an accounting
procedure of no consequence for the fiscal balance. In order to break the inter-enterprise chain of
arrears leading to accumulation of tax arrears, the govemment is implementing a strategy to
recapitalize enterprises and thus enable them to pay suppliers (typically the gas and energy public
monopolies), who would in turn pay tax and social contributions arrears to the budget. In order
to limit moral hazard, this operation is available as a once-off option at the time of privatization.
Without hard budget constraint, however, future problems in collecting taxes and social
contributions from enterprises would transíate in revenue shortfalls and affect govemment's
fiscal position

The Fund's revenue base consists of allocationsfrom state budget and of fees from leased state properties. A
2-percent levy recendy introduced on the turnover of agro-processing companies is subject to cancellation in
December 1999 under the govemment program supported by the IMF-EFF.

27

An implicit fiscalpressure from municipalities?
Municipal borrowing is subject to a legal limit of 10 percent of their annual revenues.
Out of this amount, as much as 10 percent of the preceding month's revenues can be in the form
of a short-term interest-rate free credit from the central budget. Municipalities are allowed to
seek the remaining credit from commercial banks and from other municipalities. For this credit,
municipalities can request state guarantee. Furthermore, subject to further market development
and to the approval by the Securities and Exchange Commission, municipalities will be able to
issue bonds. Sofía already successfully issued bonds in Luxembourg for 50 million Euro. The
existing law is not clear whether bond debt is subject to the overall indebtedness rules. As
experience in many south American countries, including Argentina, Brazil, and México has
illustrated, markets often perceive municipal bonds as implicitly (even when clearly not
explicitly) guaranteed by the state 17 . Possible increases in municipal credit in any form may thus
bring about strong pressure on the fiscal authorities in case of default. The govemment is closing
some of these loopholes in the draft 2000 budget law submitted to Parliament.
Fiscal risk for the central govemment may also emerge from municipal guarantees and
other contingent municipal obligations. These are not currently regulated under any existing law.
Thus, municipalities face no ceiling on the number and overall valué of guarantees they can
issue. Further risk of a financial pressure to spill over from municipalities to the central
govemment relate to the municipal ownership of loss-making and indebted enterprises and
financial institutions. The spilí-over mechanism is likely to be the stronger, the more
municipalities become responsible for the provision of core public services. Central govemment
may find it plausible to let a municipality default on its debt to commercial banks. Possible
subsequent failure of the municipality to deliver such core services as running water and
schooling may be, however, politically too sensitive to let the central govemment abstain from
"emergency" financial support. On the positive side, the State Audit Office is authorized and
active in screening municipal performance and financial management. The cost of nuclear safety
is not yet fully known.
Major govemment financing may be required by Bulgaria's nuclear power program. The
Kozloduy nuclear power plant site has 2 units of 1000 MW each and 4 units of 440 MW each.
International observers have declared the 2 large units in compliance with environmental damage
containment measures. The other 4, older, units are determined by the European Commission for
an early closure, on the basis of a Nuclear Safety Account Agreement.
The cost of
decommissioning of the power plant has not been taken into account in the past (for example,
depreciation costs have not been integrated into the energy prices). Bulgaria's govemment, while
implementing short-term safety program, indicated in the context of EU accession discussions,
its preparedness to implement the decommissioning in two steps, closing down 2 units by 2003
and the other 2 before 2008, possibly by 2006. 1 8
Dillinger and Webb, 1998.
Discussions on the timetable are still underway and exact costs of decommissioning , future demand and
capacities remaining after the closing of the oíd nuclear units not yet fully worked out. Similarly, subsequent
price, income and substitution effects on the household and enterprise sectors have not been analyzed.
Therefore, no firm estimates of new investment, if needed, to meet electricity demand are available.

28

Fiscal risk from the financial sector is currently low
Bulgarian bankers, regulators and depositare were badly burnt by the experience of 199697. Their response has been one of caution. This limits the fiscal risk that needs to be accounted
for. Risks from the past mainly relate to possible ongoing fiscal risk of past bank failures. In
practice, high inflation has eroded much of the continuing claims on banks in liquidation.
Recoveries from these liquidations are expected low but still uncertain. Bulgarian National Bank
indicates that government claims on past failed banks amount to 400 million lev. 19 Remaining
non-government deposit claims total about BGN 160 million. The government rules out any
further compensation of non-government claims (of which 50 percent were compensated).
Unrecoverable claims by state-owned banks, including State Savings Bank, on the liquidated
banks have already been fully provisioned. The privatization of the remaining four banks might
realize less funds than had been originally hoped. But further government injections needed are
unlikely to exceed BGN 100 million.
Risks from the current situation reflect several factors. One very short-term potential risk
is liquidity risk that lies in the decisión to transfer accounts of the central budgetary and extrabudgetary funds, which are currently held by banks, to the government single treasury account
held by the central bank. This transfer will not change the overall liquidity of the total banking
system. However, transferring the accounts with all their deposits, estimated at 400 million lev,
will require individual banks to liquídate some of their asset portfolios (that is their holdings of
T-Bills, foreign exchange placements with overseas banks, and similar). The selection of assets
to liquídate and conversión of remaining assets by banks may, however, cause temporaiy lack of
liquidity in the money markets. This problem appeared in the early 1999 when around BGN 80
million of funds collected from the proceeds of asset sales of banks under liquidation was
transferred to the account of the Ministry of Finance.
More generally, the major banking system risks are typically to be found in loan portfolio
weakness, interest rate or exchange rate risk, liquidity risk, or in self-dealing or fraud. Past loanportfolio weaknesses have been flushed out by the crisis of 1996-97, and indeed the real valué of
any non-performing loans at that time has been substantially eroded. The major state-owned
banks have since been under a privatization scrutiny, which has limited their appetite for any
risky lending, and this process also gives some increased confidence in the reliability of the
accounts and the classification of loans for most of these banks. Many of the second tier of
banks have received foreign capital accompanied by a degree of diligence. Therefore, the trae
aggregate capital position of the banks is likely to be fairly cióse to what they report. That means
that aggregate compliance with capital standards is good, and that no immediate threat to the
limited resources ofthe deposit insurance fund exists, unless external conditions change.

Government expects several payments from the liquidations. If transferred to the Government's account at the
BNB and sterilized in the normal way, as provided for in the currency board arrangements, this could have the
unintended effect of tightening liquidity. While this is a monetary policy issue, and as such outside the scope
of our paper, any special arrangements adopted to deal with it, such as special deposits retained by Government
in commercial banks, could conceivably present a longer-termfiscal risk.

29

Concerning the obligations of Deposit Insurance Fund, from January 1999, the blanket
deposit insurance, introduced during the crisis, was replaced by a new Deposit Insurance Fund.
The fund collects, relatively high by international standards, a levy of 0.5 percent on end-year
deposits of the banking system, and covers 95% of the first BGN 2,000 of any deposit, and 80%
of the remainder up to a máximum payoff of BGN 6,900. 2 0 No estimate is currently available of
what the average percentage payoff from the actual distribution of deposits would be. The figure
is likely to be as low as 40 percent or even less. 21
Despite a payoff of about BGN 10 million in January 1999 (see below) the Deposit
Insurance Fund still has assets of about BGN 30 million (inclusive of an endowment of BGN 17
million). Failing other calis on its resources, it will accumulate funds at a rate approaching new
BGN 25 million per annum at the current insurance levy rate. The Banking Department deposits
at the Issue Department 22 are explicitly defined as the resources available for the lender of last
resort facility in case of a systemic crisis in the banking sector. In 1999, these funds have covered
nearly 20 percent of total deposits net of settlement accounts and thus have provided additional
cushion to the Deposit Insurance Fund. The Fund also has the authority to increase this rate up
to 1.5 percent of end-of-year deposits. If still inadequate, subject to parliamentary approval, the
Fund can obtain credit from the govemment. In the context of a wider possible "too big to fail"
policy, this recourse to the govemment allows bank failure to present a fiscal risk.
Because of the relatively intense scrutiny of the larger banks, the likelihood of significant
self-dealing or fraud is limited. There is no reported evidence of malpractice in the smaller banks
either, and these are supervised by the central bank. The re-licensing process for insurance has
sharply reduced the potential for abuse in that sector.
Less is known about the business of finance houses, which enjoy very light prudential
regulation and supervisión. By far the largest of these has a balance sheet of the order of BGN
100 million and a very small capital base. Whatever the risks being faced by the investors
(whoever they may be) in such companies, it is unlikely that the fiscal authorities would meet any
deficiency here, so the fiscal risk remains small.
Banks' exposure to extemal risks entails three standard risks: (a) economic downturn, (b)
sharp and sustained interest rate movement, and (c) exchange rate change. Ideally, a systematic
Interestingly, this insurance cover differs from a common-sense reading of the relevant legislation, which
would have limited the payout to BGN 4300. However, the drafting contained an ambiguity, and the higher
amount was paid in the case of the Credit Bank which had to be dealt with only days after the new scheme
carne into effect (see below).
On the other hand, if closure was preceded by a lengthy period of suspicion, experience elsewhere shows that
uninsured depositors would likely withdraw their deposits before failure, leaving the bank with a higher share
of insured deposits.
The Issue Department and the Banking Department are the two principalfinancial departments within the
Bulgaria National Bank, following the model of the Bank of England. The Issue Department holds all of the
BNB's monetary liabilities and invests the BNB's foreign assests subject to restrictions in terms of quality and
liquidity explicitly stated in the law on the BNB. The Banking Department is responsible for enforcing reserve
requirements, monitoring financial markets and the payment system.

30

stress test would be carried out on each bank's portfolio to assess the impact of substantial
adverse developments on all of these fronts. Even without the data to conduct such an analysis,
preliminaiy broad estimates can give an indication of the broad magnitude of risks here.
Deriving these estimates, we treat the risks one-by-one, thought they could be correlated.
The risk of a sharp economic downturn that would affect loan recovery mainly correlates
with such events as Kosovo crisis or as deterioration of competitiveness of Bulgaria's
enterprises. Most of the Bulgarian banks have sufficient capital to absorb even a substantial
increase in loan-losses. The aggregate of capital and reserves in the banking system comes to
almost 60 percent of the net loan portfolio, and (with the exception of the SSB), each of the large
banks has sufficient capital to absorb the total loss of as much as one-third of its portfolio
without insolvency. Further calculations will refíne the picture, but it appears that, despite the
limited resources of the Deposit Insurance Fund, the banking system might not impose severe
direct fiscal costs arising from a major economic downturn.
Interest rate risk is negligible for most domestic banks because nearly all loans and
deposits are subject to repricing on a monthly basis. To the extent that banks are trying to profit
from the upward-sloping yield curve in foreign investments (as some are allegedly doing) there
could be a risk. However, these positions are typically in major currencies such as dollar or euro,
where interest rate volatility is not very great. Such positions would therefore present a risk to
bank profitability, but are most unlikely to be large enough to knock-on to the state budget.
Foreign exchange exposure may be a more significant issue. There are three aspects to
this. First the exposure of dollar against euro. This is strictly limited by regulation for banks.
But it could be a problem for Bulgarian borrowers, whose receivables are in euro, but whose
loans are denominated in US dollars. The existence of such cross-currency risk is generally
acknowledged, and banks find themselves in need for greater hedging. The second foreign
exchange risk relates to open positions in lev against euro. Overall, the banks are considered
long in Euro but a precise quantification is not available. Such positions are not restricted by
regulation, the argument being that the currency board eliminates all such risk. Of course, a
stress-test needs to take account of even highly unlikely events. Experience from other countries
shows it is exactly when a fixed peg comes under extreme pressure that large open positions
emerge. Should lev interest rates rise as a result of pressure on the currency board, local banks
(following the example of México, among other countries) might use their freedom to take a
large long positions in lev to profit from the high local interest rates. Were the currency board
subsequently to fail, this could result in widespread bank insolvency and a major fiscal risk.
Risks from the future are limited if government continúes in tightening prudential
regulations, supervisión and supervisoiy practice for banks and non-banking financial
institutions, streamlining the institutional structure for regulation of pensión funds, finance
houses and mutual funds and improving effective implementation of bankruptcy procedures.
Without continuous regulatory vigilance, a risk may arise from the excessive bank capacity
relatively to the current and prospective market opportunities in Bulgaria. As present behavior
would suggest, environment of low profitability may result in orderly exit of the least viable. But
international experience shows that it can also encourage reckless behavior as bank owners or

31

bank management "go for broke" when failure or exit is otherwise inevitable. Finally, the
currency board arrangement may generate a risk in the long term if it is not supported by
adequate fiscal, income and structural policies.

III. Developing Fiscal Risk Management Framework
1. Individual measures to address selectedfíscal risks
In this sectíon, we develop an action plan with measures to reduce the government's
exposure to físcal risks. We place emphasis on public finance institutional arrangements and
capacities rather than on structural policies.

Priorities for the short term

Continué cautious borrowing strategy, adjusting it to risk management objectives
Bulgaria's vulnerability to shocks, policy makers need to acknowledge the need to minimize
volatility in govemment financing and in debt service cost. By definition, a low- risk borrowing
strategy may be costlier than an adventurous one. As the country rebuilds its creditworthiness, it
will become increasingly able to manage short-term external shocks and to reduce the effective
cost of its debt portfolio.
A strategy of low risk would imply the need to adjust the govemment debt structure as
well as to strictly limit contingent liabilities (see below). The govemment will need to conduct a
detailed analysis of the structure of both relevant assets and liabilities and construct a benchmark
to outline an optimal debt portfolio. Through possible buy-backs and through new borrowing on
as needed basis the debt management office then can gradually adjust the sovereign debt
structure in the appropriate direction. Our brief analysis suggests that this direction is likely to
point toward the lev- and Euro-denominated, long-term, fixed-interest rate debt instruments. This
approach would result in a costlier, but low-volatility portfolio.
While considering any new borrowing operation, Bulgaria will need to take into account
the risk structure of its outstanding debt portfolio. This approach is particularly valuable for
heavily indebted borrowers, since its access to market-based tools is limited, principally because
of its credit rating. For example, of particular relevance for Bulgaria is the World Bank s new
loan product - a fixed-spread loan, that has an interest rate based on LIBOR, plus a spread that
can be fixed for the life of the loan. The product allows the borrowers to flexibly fix the interest
rate on disbursed amounts at any time during the life of the loan; to cap or collar it; to unfix or
refix the rate on disbursed amounts; and to change the currency and loan repayment terms, within
financial policy limits. This new product offers Bulgaria an attractive opportunity to increase the
flexibility of the govemment borrowing policy while keeping its low-risk profile. For example, it
may consider fixing the maturity of a new loan from the World Bank in such a way that it would
smooth the uneven debt servicing profile in the years ahead; or, balance the currency structure of
its outstanding obligations with the currency structure of its exports.
Additional opportunities to mitígate risks in the govemment asset/liability portfolio are
provided by the World Bank hedging products such as interest rate swaps, caps and collars;

32

currency swaps; and commodity swaps. Again, tapping international derívatives' markets may be
prohibitively costly for Bulgaria at its level of creditworthiness. The Bank's hedging producís
will only be available ío íransform íhe risk slruclure of Bulgaria's debls lo íhe World Bank. This
innovation is significant both because the bank's share of Bulgaria's overall portfolio is
substantial (and reducing the risk exposure on it would contribute to the broader risk containmení
slralegy of íhe governmenl); and because íhe sovereign borrower would thus gain experience
thaí may laler be applied in the international derívatives' markets.
On the other side, Bulgaria's return to the international capital markets may still be
premature. An early debut in the capital markets is unlikely to contribute lo the reduction of the
effective cost of external debt and may (due to Bulgaria's low rating) even exacerbate the
refinancing risk. On the contrary, another two years of strong fiscal performance and rigorous
debt servicing are bound to further improve the country's sovereign rating and to reduce the cost
of new financing. Meanwhile, further debt buyback operations, similar to the buybacks executed
in 1999, may shave some percentage points off Bulgaria's public debt burden. This will give it
larger windows of opportunity for more proactive debt management and borrowing policies years
ahead.

Define stríct limits for and build reserves to cover obligations guaranteed
government. Vulnerability to fiscal risks and of the EU accession objective require Bulgaria to
implement a tight limit on the amounts of newly issued and outstanding obligations under state
guarantees. Ideally, this limit would fully reflect government budgetary constraints. For
example, assuming a 30-percent default risk, the government would immediately set aside 30
percent of face valué of new guaranteed obligations to build its contingency reserve fund possibly
within he Fiscal Reserve Account. This fund would maintain reserves across fiscal years (rather
than being a simple budget allocation that disappears unless utilized in a given fiscal year).
Ceilings, reserve requirements, and an imperative to prioritize should apply to all sources of
explicitly guaranteed debt, encompassing obligations of all agencies, including Export Insurance
Agency, Business Promotion Bank and State Fund for Agriculture, that are backed by the
government.
Analyze and regularly report físcal risks. Require that analysis of fiscal risks be added to
government fiscal reports. Regular reports should show debt scenarios (ideally in net present,
mark-to-market valué as well as on annual, cash basis), discuss the potential future fiscal impact
of contingent liabilities and tax exemptions. These reports would form a basis for reassessing
government borrowing and risk management strategies.

Priorities for the médium term

Cióse loopholes, require due diligence, monitor all sources of físcal risk, and pr
contingency plans. Legal status and regulatory framework applying to public sector institutions,
municipalities, and agencies, backed by the state, need to cover all possible sources of their
financial risk to be effective in preventing possible spill-over on the central government. For
example, the budget law should cióse any remaining loophole allowing the municipalities to
issue guarantees and take on other contingent obligations, such as municipal insurance programs
or issuing municipal bonds outside the overall limits on indebtedness and guarantees.
Furthermore, all programs and agencies explicitly or implicitly associated with the state, should

33

be subjected to tight reporting requirements and supervisión. Supervisión as well as their design
and legal status need to ensure that their performance (including maintenance and investment)
and financial management will not create fiscal pressure on the govemment in the future. To
enhance its monitoring, the govemment should include all sources of possible obligations in its
registers of contingent liabilities. While implicit contingent obligations may remain confidential,
the govemment needs to understand their size and determinants and prepare adequate
contingency plans for how to face possible pressures in advance.23

In the Public Investment Program, incorpórate investments associated with EU acce
and refíect on the recurrent cost of all proposed investments. Bringing all investments into the
Investment Program and incorporating recurrent cost into investment decisions is cmcial to a
good fiscal performance in the future. Eventually, a medium-term fiscal framework should
replace the Public Investment Program and ensure that future investment and recurrent costs
correctly show in the medium-term fiscal planning.

Desígnate risk management responsibilities and build risk management capacities.
optimize govemment liability portfolio so that it is consistent with govemment risk management
objectives, to design programs of contingent support (such as guarantee contracts) in a way to
minimize moral hazard in the markets and fiiture fiscal cost, to assess adequate reserve
requirements so that they provide the needed cushion but without entailing unnecessary
opportunity cost, and to benefit from sophisticated hedging instruments and derivative markets,
the govemment needs to clearly outline risk management responsibilities of its agencies and
support their capacity building. This is not an easy task but experience of countries like Ireland,
the Netherlands and Sweden offers useful lessons. Meanwhile, newrisk-hedging products of the
World Bank allow Bulgaria to reduce risk exposure, particularly the government's significant
exposure to interest rate risk, with minimal counterpart risk, and to gain needed experience in
using more complex derivative instruments.

2. A better framework for fiscal management
In this section, we outline a possible development of the overall public finance
institutional framework in Bulgaria to better capture fiscal risks. This framework aims at
promoting govemment fiscal performance across all types of govemment fiscal activities and
with respect to all sources of fiscal risks as well as to the budget. One of the specific goals for
the new framework is to promote risk-awareness culture in govemment. Below, we summarize
the key features of the new institutional framework.24
Monitor and report. Statements of contingent liabilities, tax expenditures and other fiscal
risks should list the various sources of fiscal risk, discuss their nature and sensitivities,
implications on future fiscal position and equity, and where applicable provide their face and/or

For background on how to deal with implicit contingent liabilities see Schick, 1999, Polackova, 1998, and,
with respect to the financial sector, Honohan, 1999.
24

For a more comprehensive list of measures and questions associated with the management of fiscal risks,see
Annex 3.

34

estimated valué. 25 Good practice is offered by Australia, Cañada, Colombia, Netherlands, New
Zealand, and the United States. Budget documents of these countries stress-test the fiscal
baseline with respect to major macroeconomic, policy and demographic risks. Since risk
exposures change with time, as underlying conditions evolve, government should reassess
assumptions that underlie its risk analysis several times a year and adjust its reserves.
To prevent spill-over of risks held by various agencies, such as subnational governments
and other public entities, state-guaranteed funds, state-owned enterprises, and large financial
institutions and banks, it is critical for the government to enforce transparency in the markets. In
addition, government needs capacity to monitor and analyze its risk exposure vis-á-vis these
implicit sources of fiscal risk. As a part of monitoring its risk exposures, government more than
risk managers in the private sector needs to monitor behavioral determinants of fiscal risks.
These include performance (moral hazard) under its guarantee contracts, legality of claims
against tax duties, and performance of subnational governments or other entities that may affect
future government fiscal position.
Adjust medium-term físcal strategy to risks. Government medium-term fiscal strategy is
hardly viable if it does not reflect implications of fiscal risks. The best way to show mediumterm implications of fiscal risks for the overall fiscal position, is to develop baseline mediumterm cash projections and stress-test these projections with respect to specific determinants risks
(ideally, taking into account any existing correlations). Results of stress-testing will be critical
for government to re-estimate periodically the sensitivities of its risk exposure with respect to
changes in the underlying assumptions and to decide on reserve and hedging strategies
accordingly. For implicit risks, medium-term fiscal strategy should be accompanied by
confidential contingency plans on how to proceed when implicit and indirect risks materialize.
Similarly, equity and efficiency considerations are relevant only if they address the whole
range of fiscal activities and their future likely costs. The budget law should require that
guarantees and other promises of contingent government support be subject to the same scrutiny
as budgetary programs.
Create accountability for risk analysis and management. To ensure that policy makers
adequately reflect fiscal risks in their fiscal plans, they must become accountable for their
analysis and assumptions. Therefore, supreme audit institutions and budget reports should
comment on, and evalúate ex post, government risk analysis and risk management.
Share risk. To create proper incentives for the borrower, lender, and program manager
under programs of government contingent support, the budget law should demand a significant
amount of risk to be shared by these parties. The budget law may advise that state guarantees

Since no cash is spent from the budget when governments assumes a contingent obligation or provides a tax
exemption, cash-based accounting systems fail to detect suchfíscal risks. But a full accrual-based accounting
system is neither necessary ñor sufficient as a remedy. International accounting standards, for instance, require
only probable contingent liabilities (contingencies with relatively high probability of realization) to be included
in the balance sheet, leaving the others in a sepárate statement of contingent liabilities.

35

will not cover either the full credit/obligations or risks that are under control of the borrower and
lender, such as specific commercial risks.
Make provisions and account for fiscal risk Budgeting and provisioning for fiscal risk
has three main objectives: First, draw more attention to risk analysis, and thus build risk
awareness of policy makers and the public. In this context, the process of risk analysis is often
more important than the actual end estimate of govemment risk exposure. Second, make policymakers cash neutral, that is neutral between
Box 3 Prívate Sector Logic on Provisioning
alternative forms of providing govemment
support from the viewpoint
of déficit The prívate sector offer examples, why to provisión
measurement and sectoral budgetaiy envelopes. ..fully for the expected cost of contingent support
immediately in the year the program is issued.
Third, particularly important for governments Programs of contingent support can be often shown
with a limited access to debt markets, create a as a put option written (given) by the govemment.
In the prívate sector,financial institution charges
buffer for the event that risks materialize.
full option price when writing (selling) an option.
Price of an option reflects the present valué of the
Alien Schick, 1999 has summarized the
future possible loss, which may be incurred by the
underwriting institution. As illustrated by the
contemporary approaches to meet the above
Black-Scholes formula, the price increases with the
objectives in the following principies26:
time to expiry (e.g., maturity of the guaranteed
• Apply joint ceiling for the cost of budgetary loan) and with the volatility of the underlying asset
(e.g., share price of the enterprise the debt of which
and off-budget support for each sector in a
is under the guarantee). Financial institutions
fiscal year. Off-budget support is a form of
charge the full option price immediately at time of
subsidy, which should be considered along selling the option. The amount is then used either
other forms of govemment support for each to build reserves or to buy a hedge. See Hull,
1997.
sector.
• Calcúlate the cost of off-budget support as the
—27
present valué of the future expected fiscal cost.
• Reflect the cost of off-budget support in full in the year when it is issued (see box 3).
• Transfer an amount equal to the cost of off-budget support from the budgetary envelope of
the related sector to a central contingency reserve fund.
These principies pulí the expected fiscal cost of off-budget support into déficit accounting
in the year when the programs are launched. The amount transferred to the reserve fund is
accounted for like an expense. On the other hand, when contingent claims are paid out from the
reserve fund, fiscal déficit remains unaffected. Implementation of the Federal Credit Reform Act
in the US has confirmed that this approach is plausible even when accounting and budgeting
systems are not fully on an accrual basis.

For suggestions in a country-context see Brixi, Ghanem and Islam, 1999, Mody 1999, and Brixi, Papp and
Schick, 1999.
There may be a discount applied on a newly proposed policy (say drought insurance) if itsrisksare negatively
correlated with therisksof other policies (say, minimum agricultural price guarantee and flood insurance).
This is because in such a case the newly proposed policy would increase govemment overallriskexposure by
less than its full expected cost.

36

Utilize asset liability management strategy to neuter físcal risks. An integrated approach
government assets and liabilities (particularly to borrowing, debt management and to the
management of reserves) allows to find natural hedges to fiscal risks and thus reduces the
contingency reserve requirement and the need of derivative instruments. Particularly if facing
constraints in the choice of available risk management tools, the government needs to decide on
its new borrowing and on rebalancing the portfolio of outstanding liabilities in the context
broader challenges posed by the structure of its assets, future revenue streams, and contingent
and implicit liabilities.
Assignrisk management functions and build capacities. For risk management as well as
for conventional budget management, control structures and accountability are crucial. Important
features to improve control and reduce scope for fraud and corruption include centralization of
risk-taking authority and división of risk management responsibilities (see box 4). For
government risk management, crucial authority and responsibilities are likely to be best placed at
the Ministry of Finance, particularly in its departments managing debt and budget. Their
functions, as well as government risk management strategy should, however, reflect the level of
their financial and risk management capacities. And countiy experience has indicated that a lot
of attention (including, for instance, aggressive hiring approaches and independent, market-based
salary scales) is needed to develop and retain such capacities in the public sector.
Box 4

División of Responsibilities

Large banks, including JP Morgan and Deutsche Bank, have divided the functions of designing and authorizing new
transactions, analysis, and record keeping among three different offices. Thefront office centralizes designing of.
financial instruments and has the exclusive authority to enter into new derivative and debt transactions. Its objective is
to ensure required levels of available cash and optimize overall return-risk ratio. Middle-office provides analysis of .
future obligations and payoffs, and their sensitivities for the entire portfolio. Finally, back-office is responsible for
record keeping and maintaining comprehensive databases.
Maintaining these functions independent of each other improves transparency and control of portfolio risks,and it
prevents the front office from exceeding their predeterminedriskexposure limits.
¿Governments of Ireland and Sweden, among others, have successfully applied such división of responsibilities in their.
debt management. Many governments have successfully centralized the authority to issue debt, guarantees, tax
exemptions, and other off-budget programs. Further, they should now expand the scope of risk management and adjust
organizational structures and responsibilities accordingly. See Nars, 1997.

IV. Conclusions
Bulgaria's current fiscal position is strong. The CBA has been effective in imposing fiscal
discipline, but it leaves limited room for accommodating potential risks. Significant pressures
could arise from increases in international interest rates, from environmental liabilities and
investment requirements, and from failure to strengthen the collection capacities of the social
protection institutions and to establish adequate fiscal risk monitoring. Our analysis demonstrates
the importance of limiting future exposure of the government of Bulgaria to risks while
accommodating investment needs, crucial to growth and development. For this, Bulgaria needs to
find an optimal mix of debt management, fiscal reserves and risk mitigation strategies.

37

Specifically, there are several steps for the govemment to follow in dealing with the
existing risks and to constrain the scope for further accumulation of risks. In this context,
priorities include mitigating currency and interest-rate risks in the govemment liability structure,
implementing of the envisaged institutional and financing reforms of the pensión and health
systems, building adequate contingency reserves, introducing risk-sharing arrangements,
prioritizing and placing strict limits on the amounts of new guaranteed obligations, and
developing govemment capacities to analyze and manage risk. Fiscal risks should become fully
integrated with govemment policy considerations, in the process of fiscal management, and in an
integrated asset and liability management strategy.

38

V. Bibliography
Anand, Ritu, and Sweder van Wijnbergen. 1989. "Inflation and the Financing of Government
Expenditure: An Introductory Analysis with an Application to Turkey." World Bank
Economic Review 3(1): 17-38.
Anand, Ritu, and Sweder van Wijnbergen, Ajay Chhibber. 1990. "External Balance and Output
Growth in Turkey: Can They Be Reconciled?" In Tosun Aricanli and Dani Rodrik, eds.,

The Political Economy of Turkey: Debt, Adjustment, and Sustainability. Houndmil

United Kingdom: Macmillan Press.
Anand, Ritu, Sweder van Wijnbergen, and Roberto Rocha. 1988. "Inflation, External Debt, and
Financial Sector Reform: A Quantitative Approach to Consistent Fiscal Policy with an
Application to Turkey." NBER Working Paper 2731. National Bureau of Economic
Research, Cambridge, Mass.
Campos, Ed and Sanjay Pradhan. 1996. Budgetary Institutions and Expenditure Outcomes.
Policy Research Working Paper # 1646. The World Bank, Washington, DC.
Dillinger, William and Steven Webb. 1999. "Fiscal management in federal democracies:
Argentina and Brazil", World Bank Policy Research Working Paper #2121, Washington
DC.
Easterly, William. 1998. "Fiscal Alusión, "World Bank Policy Research Working Paper #2109.
Washington, D.C.
Euromoney, 1997. Excellence in debt management Kari Nars, Editor. London, England.
Ghanem, Hafez. 1997."Ukraine: Public Expenditure Review," the World Bank, Washington,
D.C.
Hemming, Richard, 1999. A Framework for Assessing Fiscal Vulnerability. IMF. Workin
progress.
Honohan, Patrick, 1999. "Fiscal Contingency Planning For Banking Crises," the World Bank
Policy Research Working Paper, Washington, D.C. (forthcoming).
Hull,John. 1997. Options, Futures and Other Derívatives. Prentice Hall, Upper Saddle River,
New Jersey.
Kharas, Homi, and Deepak Mishra. 1999. "Hidden Déficits and Currency Crises." EASPR,
World Bank, Washington, D.C.
Polackova, Hana. 1998. "Government Contingent Liabilities—a Hidden Risk to Fiscal Stability."
Policy Research Paper #1989. The World Bank, Washington, D.C.
Schick, Alien, 1999. Budgeting for Fiscal Risk. Unpublished.
Selowsky, Marcelo, 1998. "Fiscal Déficits and the Quality of Fiscal Adjustment" in European
Union Accesión: The Challenges for Public Liability Management, World Bank.

39

World Bank. 1999. "Fiscal Risks due to Off-budget Liabilities in Indonesia" in Public Spending
in a Time ofChange. Washington, D.C. Unpublished.

40

Annex 1
Table on Bulgaria General Government, 1994-2002
19
95

19
96

19
97

19
98

2000

19
99
Proj"

20
01
Proj"

Proj"

20
02
Proj"

(In percent of GDP)
Revenues
O/w Tax revenues
O/w VAT taxes
Social insurance contributions
Nontax revenues

36.1
29.3
6.7
7.9
5.7

32.6
26.5
6.7
6.9
4.9

31.7
26.7
8.5
6.9
4.4

36.8
31.4
8.5
7.7
5.3

38.0
31.5
8.6
8.7
6.4

38.0
31.7
8.6
8.7
6.3

38.1
31.9
8.8
8.7
6.3

38.3
32.0

Expenditures
Non-interest expenditures
O/w Compensation
Subsidies
Maintenance/operating
Social expenditures
O/w Pensión
Capital expenditures

42.4
27.8
4.6
1.1
5.5
10.8
8.0
1.1

45.2
24.9
3.5
0.8
4.9
9.1
7.0
0.7

34.2
25.8
3.7
0.7
6.3
8.5
6.3
1.0

35.8
31.4
4.7
2.0
5.7
10.9
8.3
2.8

39.5
34.6
5.5
1.2
5.4
12.7
8.1
3.4

39.4
34.7
5.5
0.8
5.4
13.5
7.9
3.5

38.4
34.5
5.6
0.5
5.3
13.5
7.6
3.6

38.9
34.4
5.6
0.4
5.2
1.
34
7.3
3.6

Primary balance
(Excluding costs of structural reform)
Interest
External
Domestic

8.3
8.3
14.6
3.3
11.3

7.7
7.7
20.3
3.4
17.0

5.9
6.2
8.4
3.0
5.5

5.4
5.2
4.4
3.2
1.2

3.4
4.3
4.9
3.6
1.3

3.2
4.4
4.7
3.4
1.3

3.6
4.4
4.4
3.1
1.4

3.8
4.4
4.4
3.1
1.3

Overall balance
(Excluding costs of structural reform) 2
/

-6.3
-6.3

-12.7
-12.7

-2.5
-2.2

0.9
0.8

-1.5
-0.5

-1.5
-0.3

-0.8
-0.1

-0.6
0.0

Financing
External(net)
Domestic (net, include IMF)
Privatization

6.3
-0.9
7.0
0.0

12.7
-2.4
14.9
0.0

2.5
-1.3
0.5
3.3

-0.9
-1.0
-1.5
16
.

0.9
-0.7
-1.0
2.6

1.5
0.4
0.0
1.3

0.8
-0.4
0.0
1.2

0.6
-1.2
0.9
1.0

Memorándum

8.7
6.3

item:

Nominal GDP (inoldlevbn)
880.3 1748.7 1 0 5 2 5 7 2 0 0 23472 2 2 6 2 2 1
75
17
21
51
77
Fiscal Reserve Account (% of GDP)
0.0
0.0
7.8
9.2
11.1
11.8
10.1
7.0
Sources: Ministry of Finance, and IMF
staff projection
Notes: 1 Including the activities of t e State f n for Reconstruction and Development (SFRD) and the Energy Resource F n
/
h
ud
ud
(ERF) which were closed with t e approval of t e 1 9 budget.
h
h 99
2/ From 2000, costs of structural reform include projected additional costs of social security reforms.

41

Annex
Table 1: State Guaranteed Credits, By Creditor, December 31,1998
million USD
Official creditors
European Investment Bank
World Bank
The Overseas Economic Cooperation Fund
EBRD
K F Arab Economic Development
USA govemment
Total official creditors
Prívate creditors
Siemens
Nichimen Co
Total prívate creditors
Total

Negotiated
amount

Disbursed
amount

Amount
Outstanding

386.7
304.5
182.7
172.0
40.1
15.0
1,100.9

202.6
154.0
2.7
127.3
0.0
14.9
501.4

192.6
144.8
2.7
103.4
0.0
14.9
458.4

52.3
6.4
58.7

0.0
6.4
6.4

0.0
4.8
4.8

1,159.6

507.7

463.2

42

Table 2: State Guaranteed Credits by Borrower, December 31,1998
Amount
Negotíated Disbursed
Million USD
amount
amount
Outstanding
Public borrowers
57.1
161.5
59.2
MRDPW-Transit roads
114.8
133.2
130.1
Bulgarian Telecommunication Co
Port-Bourgas
117.0
0.0
0.0
Sofía Airport
101.5
0.0
0.0
Bulgaria - energy
93.0
43.0
36.9
Railway rehabilitation
91.7
57.3
57.3
Air trañic control
67.4
46.9
46.9
33.9
NEC
42.4
63.8
11.6
11.6
Water companies restructuring
53.6
Ministry of Health
52.3
0.0
0.0
Bulgarian State Railways
45.0
26.6
26.6
NRA - project transit roads
43.0
43.0
33.6
1.4
22.5
1.4
Social insurance administration
14.9
14.9
15.0
Ministry of Agriculture
MA (project wholesales markets)
6.7
1.9
1.9
Bulgarian National Televisión
1.0
1.0
0.3
Total public borrowers

1,068.2

479.4

437.3

Table 3: State Guaranteed Credits, by Borrower, as of December 31,1998
(continued)
Amount
Negotíated Disbursed
millinn
/7?D
IllUUUIl UDLS
Outstanding
amount
amount
Prívate borrowers
KCM, S.A.
48.7
1.8
1.8
0.8
Eliseina, LTD
17.0
0.8
14.1
14.1
13.2
Prívate investment and Export
6.4
4.8
6.4
VMZ-SOPOT
5.2
5.2
5.2
Banks - Mediators APEX
Total prívate borrowers
Total (public & Prívate)

91.4

28.3

25.9

1,159.6

507.7

463.2

Notes to Tables 2 and 3:
1 Data is in accordance with the Bulgarian National Bank (BNB) register of government and publicly guaranteed debts based o
/
n
official MoF information on ratification of NA credit agreements and on official MoF information on credit servicing.
2 USD equivalent is based on central exchange rates of t e relevant currencies against BGN quoted by t e BNB on 5 July 1 9
/
h
h
99
3 16,370.64 activated state guarantee on APEX loan. Interest is paid by sources of MoF. Data is submitted by BNB
/
International Reiations Directorate.

43

Annex
How to Build Risk Management System for a Government?
What are the key principies for govemment to manage its fiscal risks? Experience in both
the public and prívate sector suggests that following two tasks are critical for govemments to
ensure future fiscal stability and allocative efficiency in the use of future as well as present public
finances. First, understand the govemment risk exposure and build risk awareness. Second,
enforce accountability for dealing with risks. Below we summarize the specific questions that
policy makers will need to ask to bring govemment fiscal performance fully under their control.
Further research is needed to elabórate on some of the recommendations below.

Questions to ask with respect to the already existing fiscal risks:
•
•
•
•
•
•
•
•
•
•
•
•
•

What are all sources of fiscal risks?
What are the sources of future possible increases in govemment obligations (direct and
contingent, explicit and implicit liabilities)?
What are the sources of future possible reductions in the valué of govemment assets and
revenues (explicit and implicit risks to assets and revenues)?
What are the types of these risks (applying the Fiscal Risk Matrixes)?
What are the limits of govemment responsibilities for implicit liabilities and for conditions of
sudden revenue and asset shortfalls?
How sensitive are the risks to their underlying variables?
What is their expected fiscal cost and valué at risk (e.g., applying stress-testing)?
To which risks is the govemment over-exposed?
To reduce over-exposure, is it possible to redesign govemment obligations, and asset and
revenue policies, and to reorient its role?
How to make the Fiscal Risk Matrixes public and benefit from transparency? Which implicit
liabilities, if any, should remain undisclosed and why?
Should the govemment build reserve funds or use hedging instruments to deal with possible
future obligation increases and revenue and asset shortfalls?
How to hedge the government's overall risk exposure that is emerging from its entire
portfolio of fiscal risks? Less optimally, how to hedge specific risks?
If hedging is not an option or not sufficient to offset govemment risk exposure, how large
reserves should be with respect to future possible? Should the govemment purchase
insurance or re-insurance instead?

Steps to build a new risk management system:
•
•

Seek to ensure that govemment has the cash available to meet its obligations and deliver.
Establish analytical framework for govemment to regularly analyze its fiscal risks, including
the future possible increases of govemment obligations and shortfalls in govemment
revenues and assets. Regularly update the Fiscal Risk Matrixes in the budget process.

44

Seek prívate sector solution to any request for govemment support to avoid risk exposure.
Analyze expected and máximum likely fiscal cost of obligations and expected reductions in
future revenue and asset valúes prior to making any policy decisión.
Design govemment support, revenue and asset policies so to achieve the policy objectives
with minimum govemment risk exposure and without inducing moral hazard in the markets.
Outline a strict policy for hedging govemment fiscal risks (future obligation increases and
asset-revenue shortfalls) according to the govemment risk management capacities. Require
hedging decisions done in the context of the entire portfolio of all fiscal risks. Regúlate the
possible use ofrisk-offsettingbonds and financial derivatives. Discourage dynamic hedging
strategies.
Outline reserve policy with respect to fiscal risks on both liability and asset-revenue fronts
according to the govemment ability to absorb shocks. Determine govemment ability to
absorb realization of fiscal risks according to its market access to borrowing (the depth of
govemment securities market). In reserve policy, specify requirements on the reserve
amounts (proportionately to govemment risk exposure), management (investment of
reserves), and on the use of reserves (only to cover the realization of risks that have been
identified ex ante). Consider new possible institutional approaches of contract-based or
market-value based reserve management.
Outline principies for govemment to purchase possible insurance and re-insurance under
competitive terms.
Establish a medium-term fiscal framework to optimize govemment management of
expenditures, obligations, revenues and assets.
In fiscal reports, require a statement of contingent liabilities, tax expenditures and other fiscal
risks.
Reflect on future possible realization of existing fiscal risks in the medium-term fiscal cash
projection scenarios and medium-term fiscal strategy.
Before designing new govemment support, revenue or asset policy, consider what is and how
to limit govemment risk exposure. For support (budget, contingent liabilities, and tax
expenditures), evalúate the fit with govemment priorities. Minimize and announce clearly
the limits of govemment responsibilities to reduce moral hazard. Require public-private risk
sharing.
Make provisions for fiscal risks already in the year when the govemment takes on the risk by
transferring the expected físcal cost from the budget allocation of the related sector to a
reserve fund (or to cover the cost of hedging and/or re-insurance).
In execution, stick to the present limits of govemment support. Before fulfilling an implicit
obligation, assess the impact on moral hazard in the markets.
Monitor fiscal risks, their sensitivities, and expected and máximum likely fiscal costs.
Monitor reserve adequacy.
Monitor hedging strategies.
Ensure sufficient transparency and controls in govemment risk management.
Involve the state audit in assessing the quality of govemment risk analysis, reserve adequacy,
risk management, and execution of risks that have realized.
Explain publicly any departures from originally envisaged realization of fiscal risks as well as
from the originally envisaged budget.

45

•

Ensure adequate monitoring of fiscal risks that may arise from both the public and private
sectors.

World Bank User
C:\Bulgaria\BlgFISCPAP2.doc
12/21/99 12:57 PM

46

RISK MANAGEMENT OF CONTINGENT LIABILITIES
WITHIN A SOVEREIGN ASSET-LIABILITY FRAMEWORK

by
Elizabeth Currie
Antonio Velandia

The opinions expressed in this work are those of the authors and should not be attributed to the World
Bank, affiliated organizations, or members of its Board of Directors or the countries they represent. We
are deeply indebted to Fred Jensen for his guidance and suggestions throughout the development of the
paper. Thanks also go to Eriko Togo and Sam Pal for their helpful comments. Naturally, we are solely
responsible of errors remaining.

Sesión 3 - 307

Table of Contente
I.

Introduction

309

II. The ALMframework

311

A. ALM approach for a private

firm...

311

B. An ALM approach for sovereigns...............................................................
III.

311

The conceptual government balance sheet

313

IV. Contingent Liabilities and Sovereign Balance Sheet Risks

317

V. Risk Measurement-Valuing Contingent Liabilities
VI.
VIL
VIII.

320

Risk Management

324

Organizational implications

327

Examples ofCL risk management with elements of sovereign ALM.
A. New Zealand

328

B. United States

329

C.

329

The Netherlands

E. Colombia

330

IX. References

333

Sesión 3 - 30

3

Abstract
Important research has already been made on the potentially destabilizing fiscal impact of
contingent liabilities. Additional insights can be gained by working within a broader
conceptual framework based on the govemment's balance sheet. This approach would place
contingent liabilities alongside other sovereign exposures arisingfrom both assets and
liabilities. A number of sovereigns using asset-liability frameworks have assigned risk
monitoring and policy formulation functions for managing risk originating from CL to debt
offices. It seems to be a logical step extending the infrastructure available for risk
management of sovereign debt to CL, since techniques and skills required are in both cases
the same. Given the risks involved for most emerging economies from recent crises, it is
expected that this pattern will be followed by more countries in the future.

I.

Introduction

The focus of this paper is the insight gained into managing sovereign contingent liabilities (CL) by
considering theirriskstogether with those of other govemment assets and liabilities. Important research
has already been made on the potentially destabilizingfiscal impact of CL, which can mask the true fiscal
situation of a sovereign1. Additional insights can be gained by working within a broader conceptual
framework based on the govemment's balance sheet. This approach would place contingent liabilities
alongside other sovereign exposures arising from both assets and liabilities, in order to systematically
identify and then quantify both the interaction between the various risk types and total sovereign risk.
Although CL have characteristics which imply a specific approach to their valuation, there are more
similarities than differences with other balance sheet items when it comes to identifying and quantifying
sovereign risk.
The balance-sheet approach also has implications for riskmanagement of CL. In general, risk can
be managed at a decentralized level, as for example, by sub-national entities and SOE. However, for many
developing countries it may be convenient to have central govemment monitoring of significant balancesheetrisks,including those originating in CL, so as to ensure that actions of third parties do not increase
its exposure as guarantor of last resort. In some cases, the central govemment may wish to go beyond this
and to manage residual public-sector risk,riskthat remains unhedged at a sub-national or SOE level. This
allows the central govemment to address riskfrom a portfolio perspective, taking into account conrelations
between different types of risks.
Also, the central govemment may deem it convenient to establish policy guidelines for
decentralized CLrisk-managementthat promote good governance and reflect the sovereign's risk
tolerance. Such guidelines may include principies such as establishing caps on CLrisk,risk-sharing,
neutralizing budgetary incentives in favor of CL compared to direct loans or subsidies, as well as
disclosure and accountability. These principies may be applied even more strictly in countries where
particular economic andfinancial vulnerabilities and/or a weakrisk-managementcapacity produce a low
risk tolerance.
1

See Polackova, Hana (1998), Shick, Allen(1999), and Kharas, Homi (1999)
Sesión 3 - 309

The central govemment may take on prívate-sector risk if it implies systemic risk, such as in the
case of a financial sector crisis. In this case, risk can be partly controlled a priori with effective regulatory
structure and tight supervisión. Additionalriskcoverage by the central govemment, beyond systemic risk,
is a matter of political economy policy. In any case, those policies should take into account their complete
fiscal implications and be expressed in a set of clear mies that govem the sovereign's commitments.
A sovereign asset and liability management (ALM) framework also has organizational implications
for the govemment. Depending on the country's context, economic vulnerability andrisk-tolerance,some
kind of centralized arrangement that systematically addresses risk-issues in a transparent and explicit
manner may be advisable. This could imply assigning, either to a formal asset-liability management unit
or to the government's debt office, the responsibility for reviewing the risk of the entire balance sheet,
establishing policies for sub-national govemment units, and monitoring of risk at various levels of the
balance sheet. In some countries, the debt management office has a significant role to play in monitoring
and managing an important part of sovereign risk, because of the fact that it directly manages part of the
liabilities and is frequently involved in granting guarantees to other state entities. Whether or not the debt
office ends up performing this function, the ALM approach strongly suggest the need for strengthening
the typical middle-office function of risk analysis at different levels of the public sector.
The degree of risk-managementcentralization, both in policy and organizational terms, may depend
on the macroeconomic context, on risk tolerance and risk management capacity, and on the type of
institutional and political arrangements, such as the degree of political decentralization, etc. The more
vulnerable the sovereign is to shocks and the weaker itsrisk-managementcapacity, as well as that of the
rest of the public sector and indeed the prívate sector, the more stringent central govemment guidelines
and monitoring should be. This is particularly important with regards to CL, which are often hidden and
unaccounted for. The central govemment may end up not only monitoring and managing balance-sheet
risks, but also promoting arisk-managementculture in the rest of the public sector.
The paper is divided into eight sections including this introduction. Section 2 describes the ALM
practice used in the prívate sector as an appropriate framework for the analysis and management of
sovereign risk. Although govemments have different objectives than those of corporates, and therefore a
different conceptualization of risk,the same portfolio management principies can be applied to sovereign
risk identification and strategies for controllingrisk.Section 3 discusses the govemment balance sheet as
the conceptual framework for applying ALM. It differs from the traditional accounting balance sheet in
that it includes only those items affecting sovereign risk, as defined in the previous section. Contingent
liabilities are treated as on-balance sheet items, as they are major contributors to sovereignrisk.Section 4
discusses risk identification in an ALM context, with particular focus on its application to CL. In this
framework it can be seen thatrisksaffecting CL are no different from those affecting assets and direct
liabilities, and that the techniques used to quantify CLriskare similar to the ones used for estimating risk
of other balance sheet items. The analysis of risk in the govemment balance sheet also helps to better
understand the typical vulnerabilities of emerging economies, particularly those relating to CL. Section 5
focuses on CL risk quantification byfirst giving a description of different valuation techniques and then
emphasizing on simulation analysis, which is the same type of technique used for risk analysis in a
sovereign ALM context. Section 6 coversriskmanagement of CL, suggesting some general principies as
an introduction to best practice. Section 7 deals with organizational issues that stem from the risk
management discussion. Finally, section 8 documents some practical applications of the two former
sections with country examples.

Sesión 3 - 3 1 0

II.
A.

The ALM framework
ALM approach for a prívate firm

The asset liability management (ALM) framework for risk management has been in use for a few
decades by banks and otherfinancial institutions2, and has subsequently been adopted by other prívate
sector agents. Traditionally, it starts with the characterization of afirm as a profit or net worth maximizer
under conditions of uncertainty. Faced with budgetary restrictions, the firm seeks to reduceriskstemming
from the unpredictability of business conditions until that point when the corresponding cost becomes
higher than theriskreduction benefit. Risk within this context can thus be measured as the volatility of net
worth or profits; more precisely,riskis perceived as potentially diminishing profits, or worse still, zero or
negative net worth i.e. bankruptcy.
The ALM approach gives insights on the way afirm can manage both assets and liabilities in order
to reduce the volatility of net worth by applying some basic principies of portfolio management, like
diversification and hedging. Indeed, a central argument of the ALM framework is to contain risk by
matching thefinancial features (e.g. interest rate characteristics) of the assets and liabilities, as then one
side of the balance sheet will be hedged -or immunized- by the other side.
Matching, however, does not need to be complete. In fact, normally banks borrowfrom the public
at short maturities and Iend at longer maturities and their retum is partly a function of managing the risk
embedded in the maturity transformation. This draws attention to liquidity risk as one of the critical
exposures faced by thesefirms: when liabilities cannot be rolled over or assets cannot be liquidated, a
bank may face liquidity difficulties, which, if prolonged, may lead to insolvency.
Forriskmanagement purposes, the ALMframework suggests that risks that cannot be hedged can
nevertheless be controlled through diversification. Banks apply this principie spreading their loan
portfolios by industry or geographically. Also, in the case of assets that run the risk of not being
recovered, the ALM approach would propose the creation of provisions equivalent to the expected loss,
and the maintenance of reserve capital in an amount equivalent to the unexpected loss.
In sum, the ALMframework helpsfirms analyze situations that can threaten their main objective of
profit maximization, by processing the information contained in the balance sheet, derivingfrom there the
potential risks, quantifying them, suggesting ways to reduce risk and providing guidelines to manage
residual risk.

B.

An ALM approach for sovereigns

Governments are expected to deliver certain services to society, for which society pays taxes. Like
corporates, sovereigns' future net worth is subject to uncertainty,
depending on a number of
unpredictable factors; e.g. business cycles, political events, developments in foreign markets, etc. Unlike
corporates, however, governments are not profit or net worth maximizers, and have the capacity to raise
taxes, if necessary. These two distinctive features have important implications for sovereign risk
definition and for the application of the ALM framework.
For a sovereign, therefore, risk could loosely be defined as the possibility that services delivered
have to be severely curtailed for a given level of taxes, or that taxes will have to be increased to maintain a
given level of government services, both of which could result from different situations, including
2

See Bitner, J h (1992) and Fabozzi, Frank (1990)
on
Sesión 3 - 3

sovereign illiquidity and insolvency. More formally, the government's objective, and by association, its
risk definition, could be deríved from the two streams of economic literature that deal with the optimal
combination of taxes and debt3. As will be shown, the tax-smoothing and the default risk literature
provide a well defined objective for the government: to manage assets and liabilities in such a way as to
minimize the need to raise taxes or the need to curtail services when faced with unexpected shocks
causing a permanent fall infiscal revenues or arisein the valué of liabilities.
In thefirst place, the tax-smoothing literature proposes that higher long-term growth is more likely
to be achieved if tax variability is minimal, since it reduces uncertainty. Taxes, other than lump sum,
create inefficiencies as they distort economic decisión- making and volatile tax rates create significant
losses inasmuch as they complícate long-term investment decisions, depress consumption and possibly
channel excess savings into short-termfinancial instruments. However, sufñciently volatile debt service,
or liquidity risk, may oblige the government to raise taxes and can actually threaten default as there is a
limit to the government's ability to raise taxes.
In the second place, the default risk literature emphasizes the costs of default in debt-servicing,
including: i) output losses from economic recession ii) financial institutions' bankruptcies, possibly
leading to systemic risk, either directly because the entities hold government debt in their portfolios, or
indirectly, as a consequence of the economic recession iii) income redistributionfrom debt-holders to non
debt-holders iv) loss of reputation and higher costs of future borrowing, and v) higher funding costs for
private sector borrowing as a result of the government's reputational loss.
The two streams of literature therefore suggest that the government should avoid raising taxes, not
only because this decreases welfare, but specially because this erodes the cushion that protects it from
incurring significant default costs.. That cushion -namely, their ability to raise taxes- is the very especial
asset of governments, and at the same time it constitutes governments' net worth. Consequently, balance
sheet and off-balance sheet items, including CL, should be managed in a manner consistent with the
objective of maintaining the level of services the government is expected to deliver without raising more
taxes.
By avoiding situations in which shocks cause liabilities to permanently exceed assets, the
government avoids the need to raise taxes, protects its net worth and keeps it for emergency circumstances
for which there is no possible hedge. For the sovereign's balance sheet management, this entails that asset
and liabilities should fall or rise together when responding to movements in interest rates or exchanges
rates, or, in other words, that liabilities should be immunized with assets, at least against shocks on the
financial variables.
Thus, the same ALMframework used by privatefirms could be applied to a sovereign balance
sheet. Mismatches between thefinancial characteristics of assets and liabilities risk having to raise taxes
and erode government's net worth. A mismatch of currency of denomination, tenor, interest-rate refixing
period and liquidity of the instruments, could result in a gap between the valué of the assets and that of
liabilities, as a response to market movements or due to an illiquid market context which makes liability
refinancing more difficult. The ALMframework would thus require a thorough analysis of the sensitivity
of both sovereign assets and liabilities, including CL, to thefinancial variables that determine their
performance over time.
Nonetheless, it should be noted that both assets and liabilities are also sensitive to macro-economic
variables and that immunization in this case can be far more complex. Useful insights can be gained from
3

For analysis on government objectives from the tax smoothing perspective see Missale, Alessandro (1997) a d Barro (1995).
n
For t e default riskperspective see Alesina (1992)
h
Sesión 3 - 3 1

analyzing the assets' sensitivities to these variables, and whether they add to the sovereign's risk, for
instance, in terms of excessive dependence on a narrow economic base. In a sense, sovereigns could take
the assets as a given, as fiscal flows may be dependent on structural macroeconomic variables of the
economy, and may be less flexible to changes than afinancial entity's assetflows. The ALM framework
would then suggest that risk-management focus on the liabilities, as will be discussed in sectíon 6.
In theory, thisriskidentification should analyze asset and liability portfolios, in order to evalúate
the interrelation among the different types of exposures, so as to take advantage of diversification benefits,
and to manage aggregateriskdimensions. A more systematic and holistic approach toriskmanagement
can thus be implemented, as opposed to dealing with individual balance-sheet items. Another advantage of
the all-inclusive, portfolio approach is that it implies a higher likelihood of identifying and perhaps better
managing, systemic risk scenarios, when the various sovereign exposures of both asset and liabilities are
triggered at the same time. This is the worst-case event that is rarely modeled, but that has the most
important implications for macro-economic stability and growth.
Risk, or the mismatch between assets and liabilities, can be measured either in present valué or in
cash flow terms, whichever is more convenient or useful. However, the complexity of the govemment
cash flows and the importance of analyzing liquidity risk through time make cash flows a preferred
vehicle for analyzing sovereign risk.
Finally, the same principies of diversification and provisioning used by the prívate sector in risk
management also hold for sovereigns. Nonetheless, governments may have less flexibility in diversifying
their portfolios, and a limited ability to build reserves against unexpected losses. Sectíon 6 discusses in
detail a number of tools available to governments for risk-management.

III.

The conceptual govemment balance sheet

Governments, unlike corporates, do not normally draw up balance sheets. Instead, their main
financial reports are the budget and the estimate of public sector borrowing required tofinance it. The
budget typically consists of a report of expected cash inflows and outflows for the year, and as such, is not
intended to reflect the fundamental, long-termfinancial condition of the sovereign.
Budgets regularly use cash accounting, which is one of the causes of the substantial growth of
contingent liabilities in some countries inasmuch as they appear to be a cheaper source of financing when
compared with subsidies or direct loans, as the latter do register an immediate expense using up the
govemment's spending capacity4. Accrual accounting, as will be discussed later, is more neutral when the
govemment decides on the vehicle for resource allocation and it is more effective for controlling
spending.
Nevertheless, for the evaluation of the sovereign'sriskexposures, the budget is not enough; instead,
a more adequate framework for the analysis of the govemmentfinancial condition is provided by the
sovereign balance sheet5. Since this departs from the traditional type of budgetary analysis, two points
merit further comment.

The contingent liability will only be accounted for as a expense until default or underperformance occurs. See Shick (1999)
n
Pg-H
5 Traditional measures of fiscal déficit and adjustment programs judged on those traditional measures also lack t e more
h
complete picture given by t e entire balance sheet. See Easterly (1998)
h

4

Sesión 3 - 3 1

First, although balance sheet items are normally treated as stocks, both assets and liabilities can be
measured and expressed as present valúes or as cash flows. In fact, information for some items like tax
revenues is primarily available as a flow, and the stock is computed by discounting the cash flow. For
other items, like foreign reserves, the stock is known and some assumptions on interest rates are necessary
to convert them into flows. As will be shown in section 4, for complexriskanalysis it is more convenient
to work with cash flows, even though fundamentally they are merely expressing the valué of assets and
liabilities in another form.
A second and more important consideration relates to the items to be included for running a
solvency test. The criteria for inclusión has to do with the focus of the analysis, namely, sovereign risk.
Thisriskhas been defined as a situation in which external shocks forcé a sovereign to raise taxes in order
to maintain the expected level of services, or worse, when the capacity to raise taxes has been exhausted,
its inability to service itsfinancial liabilities in general. Consequently, items that may cause greater
indebtedness, or that may help the sovereign fulfill its financial commitments, should be included. This
leads to the exclusión of assets that do not generatefinancial flows and the inclusión of contingent
liabilities.
Indeed, some items that would normally be part of the accounting balance sheet can be irrelevant
for this analysis, namely, illiquid assets such as national parks, military equipment, ofñcial buildings, etc.
Some SOE may also fall into this category, although a strong case can be made for including them in
balance sheet if they are part of a privatization program.
On the liability side a critical role is played by CL, that according to the traditional accounting
principies, are off balance sheet items6. In some cases CL pose such an enormous threat to tax stability,
that they represent the major source of sovereignrisk,and as such should be included in the balance sheet.
In sum, the conceptual balance sheet referred to in this paper is an economic one7. It can be seen as
the accounting balance sheet modified to include only those items representing potential financial
obligations, or that can effectively be used to meet those obligations. The difference is economic net
worth, and as explained, the government's objective is to avoid eroding net worth by raising taxes.
Based on this conceptual balance sheet, the ALM approach provides useful insights on the
sovereign's exposure, as the following examples illustrate. If, for instance, assets tend to be denominated
in local currency, then liabilities should also be denominated in local currency; issuing debt or contingent
liabilities in foreign currency exposes the sovereign to theriskof a local currency devaluation. Similarly,
if the sovereign has few real assets that generate financial flows, holding commitments indexed to
inflation exposes the sovereign to theriskof an increase in inflation.
This type of assessment can be done without actually computing the so-called economic balance
sheet8, allowing sovereigns to identify major exposures and adopt strategies that greatly reduce its
vulnerability. In fact, a number of OECD countries, most notably Belgium and Denmark, built their debt
management strategies based on an ALM approach without actually computing the current balance sheet.
It should therefore be emphasized that the conceptual balance sheet is more an organizing framework that
helps to identify the main sources of risksand the potential strategies that can help reduce it, rather than a
set of numbers from which actualriskcan be directly quantified.
According to t e Accounting Standards, a liability can only be recognized in t e balance sheet when it is likely to generate a
h
h
n
outflow of resources that can be reliably measured. See Shick (1999), pgs 3-4.
7 It is worth noting that m d m accounting theoiy is also moving in t e s m direction.
oe
h a e
8 Few countries publish even an accounting sovereign balance sheet. One such country is New Zealand, whose balance sheet can
be consulted at the web page of the Treasury www.treasury.govt.nz, and another more recent example is Colombia, see
Echeverry (1999).
6

Sesión 3 - 3 1

In practice, the main government assets are future fiscal surpluses. These are usually denominated
in local currency, and tend to be largely insensitive to the movement of short-term interest rates and to
inflation. Henee, when liabilities are denominated in local currency, have long duration and are not
indexed to inflation Índices, it is unlikely that the government will be forced to raise taxes to respond to a
shock in thesefinancial variables. Risk therefore can be minimized based on the previous analysis,
without actually computing the balance sheet.
Nonetheless, quantification of the economic balance sheet plays a critical role in the consolidation
ofthe government'sfiscal position. As a number of analysts have pointed out, the international experience
shows that significant fiscal riskarisesfrom CL that sovereigns keep hiddenfrom the budgetary system
In this sense, transparency and disclosure of CL is the first step towards proper risk management.
Quantification of the conceptual balance sheet may also be required for a more rigorous analysis that
attempts to determine the asset and liability portfolios the sovereign may wish to hold as a function of the
cost/risk trade-off and the sovereign'srisktolerance.
The following is a summarized illustration of a government's conceptual balance sheet.
Table 1
Conceptual Government Balance Sheet
Liabilities

Assets
PV of fiscal revenues
Foreign exchange reserves
Marketable securities
Onlending (e.g., IBRD loans)
Investments in SOEs
Investment in infrastructure (e.g. roads)

Direct Liabilities
PV of fiscal expenditures
Net market valué of sovereign debt 9
Contingent Liabilities
Explicit contingent liabilities
Implicit contingent liabilities
Equity
Net worth of government estáte

Following the above rationale, the assets included are mostly financial. The difference with the
traditional accounting balance sheet lies in the inclusión of CL in the liabilities, and the exclusiónfrom the
assets of government investments in national parks, militaiy equipment, and the property portfolio..
However, there might be cases when even some of these assets should be included, as in the case of
sovereigns that include their property portfolio in their privatization programs. As mentioned, the
particular content of the so-called conceptual balance sheet depends on the particular country being
analyzed.
The following items are the CL that may play a role in a sovereign economic balance sheet.
Contingent Liabilities are obligations that materialize if a particular event occurs. They can
be explicit, if the sovereign contractually acknowledges its responsibility to cover the beneficiary
under specific circumstances, or implicit, when the government is expected to do so because it
has a "moral" obligation, in most cases related to a high opportunity cost of not intervening. In

9

Includes t e monetaiy base and is net of Central Bank holdings of government debt.
h
Sesión 3 - 3 1

the conceptual balance sheet, CL are recorded as the present valué of expected costs for the
government, and not as the face valué of those liabilities10.

Explicit contingent liabilities
i).

Credit guarantees: the government accepts to fulfill the obligations of the guaranteed entity if it fails
to repay a loan. This mechanism is frequently used with subnationals, SOE's and private firms
participating in a privatization process. Insurance against losses on bank deposits are also a type of
credit guarantee.

ii). State insurance schemes: this is a mechanism by which governments insure against losses on crop
failures, natural disasters or guarantees a minimum return on private pensión funds.
iii). Legal proceedings and disputes: legal claims against the state may arise from property damage by
public projects, malpractice in public hospitals, claims against army or pólice forces, tax in dispute,
claims against sub-nationals and SOE's, and others. Once the court settles against the government, it
becomes a direct liability.

Implicit contingent liabilities
i).

Financial system bail-out: thefinancial sector may generate a huge economic cost if lack of
government intervention allows a systemic crisis to develop. As illustrated by the East Asian
experience, this item could be the most important CL on a sovereign's balance sheet. Even
governments of industrialized countries such as Sweden and the USA have in the past intervened
beyond their contractual obligation. Standard & Poor's has developed a methodology to measure the
size of this CL and its estimates are readily available.

ii). Fixed exchange-rate regime: the adoption of a currency peg as a foreign exchange policy implies that
the central bank stands ready to defend it by buying or selling foreign reserves until excess supply or
demand are cleared. Thus the regime entails a contingent liability for excess demand for foreign
exchange that could deplete the country's international reserves. This contingency might be triggered
either as a consequence of poor macro policies, or as a result of exogenous capital flows over which
the government has less control.
iii). Corporate sector bail-out: the government may bail out corporate entities, because of extemalities. In
Korea, banks' dependency on chaebols implied that not helping the latter would have resulted in the
bankruptcy of the financial system. Although diffícult to estímate, this item can be calculated from
information of previous government interventions, ownership concentration in the corporate sector
and linkages between government banks and corporates.
ív). Uncovered losses or default on non-guaranteed debt issued by subnationals, SOE's or central bank:
because of reputation and potential credit risk reappraisal from rating agencies, governments may
choose to honor those obligations, even when they have not backed them explicitly. The less explicit
the commitment, the harder it is to quantify this CL. Complexity increases with the number of
variables that should be estimated, such as entities that might be bailed out, size of the potential
default and timing. Past experience can be an indicator 1
1
In some cases however a sovereign m y decide to record the face valué of the CL as an expenditure to discourage its use as did
a
The Netherlands when guarantees werefirst incorporated into t e budget, see Shick (1999) pg. 29-30.
h

10

Sesión 3 - 3 1

v). Cleanup of liabilities of prívatized entities: governments may have to take over obligations of entities
to be prívatized, in order to attract investorsn
vi). Disaster relief: natural disasters, the dismantling of nuclear plants and environmental recovery are
forcé majeure events that compel the government to assume the costs involved and which are difficult
to predicti3.
vii). Critical social and welfare functions: failure of prívate pensión or social security funds could become
government obligations in some countries because it is believed that social services ultimately are an
obligation of the govemmentu.

IV.

Contingent Liabilities and Sovereign Balance Sheet Risks

The risk of a potential shock forcing the government to raise taxes can be analyzed in terms of a
potential fall in assets, an increase in liabilities or a combination of the two. However, for the purposes of
the present analysis four different sources are identified: i) decreased net fiscal flows ii) reduced asset
valúes iii) increased direct liabilities and iv) increased CL.
i) A severe contraction in tax revenues or a significant expansión of government expenditures
caused by external shocks may cause a fiscal shortfall, that in turn may induce a tax hike and
increase the likelihood of sovereign default. As is later argued, non-G7 countries are more exposed
in this sense because of specific features of their economies.
ii) A second source of risk is decreased asset valué, such as a fall in international reserves, or
financial losses registered by SOE's and by sub-national entities. Underlying factors such as the
defense of afixed exchange rate, the collapse of the price of main commodity exports, and credit or
liquidity problems of public entities could cause the fall in asset valué, with significant riskto the
government.
iii) A third source of riskto the government is an increase in the valué of direct liabilities relative to
asset valué. As government expenditures have already been mentioned, this alternative refers
mainly to debt and tends to reflect debt-servicing volatility caused by fluctuations of exchange and
interest rates.
iv) Last but not least, if significant CL unexpectedly materialize, there is the risk that taxes may
have to be raised. Various explicit and implicit CL could trigger this situation, including credit
guarantees, bailouts of financial and/or public enterprises, insurance for major natural disasters, etc.
A brief mention has already been made of different events that could negatively impact the stability
of taxes. This leads to the question of whether the underlying factors threatening tax stability are different
for traditional balance-sheet items, and for CL. It will be argued that potential tax volatility may result
from eitherrisksin the macroeconomic fundamentáis orfrom risksinfinancial variables, and that these
For an illustration see t e case of Hungary in Polackova (1999a) pg. 20-24
h
In t e case of Colombia, Echeverryfinds out t a during t e last decades, t e country has experienced a severe earthquake once
h
ht
h
h
every four years and a flood of major consequences every two years. He estimates a corresponding contingent liability of about
1.1% of 1 9 GDP. See Echeverry (1999)
97
14 For a illustration see t e case of t e pensión reform in Hungary in Polackova (1999a)
n
h
h
12

13

Sesión 3 - 3 1

two generic underlying factors affect CL, as well as regular assets and liabilities. Although
macroeconomic and financial risksare interrelated and can be modeled together, they are separated for the
purpose of the following analysis.
i) Macroeconomic risk: demand or supply shocks, such as a deterioration in terms of trade, may
trígger an economic contraction and lead to increased govemment expenditures, reduced tax
reserves and possibly inflation.
ii) Financial risk: a second type of riskinherent tofinancial variables. Financial risk includes: a)
market, b) liquidity, and c) credit risk.
a) Marketriskis understood as the uncertainty in the valué of sovereign balance sheet items, due
to movements infinancial variables such as exchange, interest rates and commodity prices.
Exchange rate risk tends to be significant. As evidence shows, local currencies in emerging
economies display a pattem of infrequent and unusually large jumps. A number of these
currencies have from time to time shown fundamental overvaluation, providing an opportunity
for speculative attacks and massive devaluations. Since assets are largely denominated in local
currency, whereas a substantial part of the liabilities is denominated in foreign currencies,
devaluations have depleted sovereign assets, and made direct and contingent liabilities more
expensive, thereby forcing the govemment to increase taxes and in some cases threatening to
provoke a default since the capacity to raise taxes has been taken to the limit.
Similarly, SOE and sub-nationals with foreign currency debt can face serious difficulties in the
timely servicing of foreign debt, for their revenues are largely in domestic currency. Liquidity
problems may trigger explicit and implicit CL, particularly when reputational risk has significant
extemalities. A massive devaluation could also trigger exchange-rate guarantees issued to
infrastructure projects. It could also cause a liquidity crisis in a banking system having a
significant short-term foreign currency debt, (which in turn may require a govemment bailout in
order to avoid systemic crisis), or bring about the insolvency of public sector enterprises with
foreign currency debt.
Domestic and foreign interest rate shocks may have similar effects as those described above for
exchange-rate shocks. If liabilities have a duration significantly shorter than that of assets, the
cost of servicing domestic debt may greatly increase when there is a significant and prolonged
tightening in monetary policy, thus endangering the tax stability objective. This can also be the
case of a prolonged defense of afixed exchange rate. However, if interest rates display a meanreverting pattem, the negative impact on the sovereign's net worth may be inferior to that of an
exchange rate shock. In any case, a marked increase in interest rates can make debt servicing
more expensive, erode thefinancial situation of SOE and sub-nationals, induce a contraction of
aggregate demand and affect the health of thefinancial system, all of which tend to induce an
increase of taxes and subsequently increase theriskof default.
b) The secondfinancial riskis liquidityrisk.Assets and liabilities may register mismatches in
maturities and liquidity, with assets having longer maturities, or less liquidity, than liabilities.
Most commonly, this situation arises as a result of an accumulation of short-term obligations that
cannot be rolled over and where governments lack the funds to service the obligations. CL can
also generate liquidityriskinsofar as they behave like American options, which can be called at
any time; they can thus trigger a liquidity crisis, especially with implicit CL that tend to be large
by nature, the most striking example being afinancial sector bailout. Another special feature of
CL is that normally there are no assets readily available to match those liabilities, given the lack

Sesión 3 - 3 1

of budgetary procedures that account for the expected cost and the deficiency of reserves against
unexpected costs.
c) Creditriskrefers to a debtor's inability to fulfill itsfinancial obligations. Creditriskfor the
central govemment can origínate from both sides of its balance sheet, namely, from its assets
and from its CL. In particular instances the two may be connected, such as when SOE or subnationals have central govemment equity and at the same time expect the govemment to
intervene in case of financial difficulties caused by excessive leverage, for example.
Different aspects of creditriskaffect the govemment cash flow in varying degrees. On the asset
side, a deterioration in thefinancial situation of SOE, sub-nationals and even the Central Bank,
may reduce the cash flows the govemment expects to receive from those assets. If financial
difficulties increase, explicit guarantees given by the central govemment may be called, and if
they deteriórate further, these entities could face a liquidity crunch or a solvency problem that
could trígger a more significant cash outflow from the govemment, in terms of a bail-out.
The degree of creditriskto the govemment originating in CL depends largely on the degree of
central govemment intervention in the rest of the public sector and the economy. At one extreme
there is small sized govemment with limited assets in vital sectors and enterprises, which
promotes a strong autonomy of public entities, very restricted explicit CL and virtually nonexistent implicit CL. Moral hazard is not an issue. At the other extreme, there is ampie
intervention of the govemment in the economy reflected in a wide variety of investments in the
real sector, management intervention by the central govemment in entities it invests in,
widespread issuance of explicit CL and the de facto acceptance of implicit CL. The two
typologies result in different CL creditriskimplications for the central government's cash flows.
From the previous analysis it can be concluded that CL are exposed to the same types of risks,both
macroeconomic and financial, that affect assets and other liabilities. The ALMframework shows that
situations that may forcé the govemment to induce tax instability can arisefrom different sources; henee
the importance of looking at the entire balance sheet.
This kind of analysis is particularly important for non-G7 economies, for the above-mentioned risks
are more significant than in their G7 counterparts, and therefore require better monitoring and
management. Emerging economies tend to be more prone tofinancial and economic instability than their
counterparts in industrialized countries, as the following examples show:
i) They have more volatile business cycles since their economies tend to be less diversified: GDP
and govemment revenues sometimes depend excessively on one or two economic sectors or
commodities. Price swings and negative terms of trade can have excessive effects on these
economies' asset flows, without offsetting flows on the liability side.
ii) In some countries, there is an inherent trend for pro-cyclicalfiscal policies that mismatch fiscal
revenues and expenditures in the downturn of the economic cycle. In effect, because of pent up
demand that builds up during recession, govemments can not run a restrictive fiscal policy in
periods of economic expansión, whereas in recessions, govemment expenditures are highly inertial
and difficult to cut.
iii) Emerging market countries generally lack developed deep and liquid capital markets, which
forces them to have short term domestic currency debt with increasing refinancing riskand/or incur
foreign currency debt with consequent currency risk. Since assets are mainly long term and
denominated in local currency, this represents a major cause of vulnerability.

Sesión 3 - 3 1

This vulnerability seems to be aggravated by the role played by CL. The following examples
highlight only a few important CL: guarantee issuance, the implicit backing of thefinancial system, a
fixed exchange-rate regime and the potential bail-out of SOE and sub-nationals.
i) Many non-industrialized countries are in the process of privatizing their infrastructure, searching
for benefits including new infrastructure development, reduction of budgetary subsidies and
improvements in operation of existing infrastructure15. Sovereigns may extend guarantees in order
to implement this process, and thus promote private sector participation in desirable economic
areas, or obtain funding for socially desirable projects. These are important CL that may eventually
bring problems to the sovereign's ALM.
ii) Financial sectors of emerging market countries often lack adequate regulation and supervisión
which makes it easier for the banks to exploit the implicit backing of the government by adopting
riskier positions. This situation is aggravated when the country opens its capital account, for banks
already in a vulnerable position are exposed to more risk, especially to the volatility of capital
flows.
iii) Some emerging economies have used afixed exchange-rate regime as an anchor to help reduce
inflation. This may sometimes lead to maintaining an overvalued exchange rate, that is an invitation
for speculators to make profits when the central bank starts losing significant reserves.
iv) If the public sector is relatively large within the economy, sovereign balance sheet statistics will
result in a complex layering of explicit risks. If there is an additional dimensión of moral hazard
arising from implicit CL, more privatized economies do not necessarily imply less risk for the
central government, as was learned from the East Asian crisis. CL in particular are generally not
clearly identified and documented.

V.

Risk Measurement-Valuing Contingent Liabilities

Thefirst step in quantifying riskis to valué CL, in terms of the present valué of future net costs that
the government expects to incur by accepting such liabilities. The CL valué can be estimated in various
ways, including actuarial techniques, econometric andfinancial models, and contingent claim analysis.
Actuarial techniques are widely used in the insurance industry. They estímate the future pattern of
losses based on historical tendencies. For results to be statistically significant, both a large database and a
highly stable distribution of losses to be insured are required. This method can be explained with an
example of an explicit guarantee extended by the central government to cover a portion of social security
or health benefits. The actuarial method computes the CL's expected cost as the nominal or face valué of
the guarantee, times the probability of the guarantee being used, where the probability is based on the
historical frequency of such guarantees being called. If a large part of the population receives social
security or health benefits guarantees for a sufficiently long period and maintains the same pattern of
using them, on average the estimation will beright.However, this technique has some limitations that
make it difficult to extend its application beyond these cases. Firstly, there is insufficient data on specific
events that triggered the use of CL in the past. Secondly, even when such data is available, the micro and
macro economic context may have changed in such a way that the assumption of a stable distribution of

This is the case of transitional countries t a are shifting t e role of the government from directlyfinancing a d providing
ht
h
n
services, to guaranteeing that the private sector will accomplish particular outcomes

15

Sesión 3 - 320

expected costs does not hold. Thirdly, the CL must involve large samples, so that diversification can
substantially reduce the variance on the aggregate.
Econometric andfinancial models recognize the potential instability of the expected cost of CL,
since those costs are triggered by economic andfinancial variables whose behavior can change over time.
Instead, the government's expected CL cost -or the mean of the distribution- is calculated with models
that forecast the underlying economic andfinancial variables that cause the CL to be called, but no
attempt is made to measure the variance. These models are also used to assess the impact of different
assumptions regarding the economic variables on the government's expected cost, in order to obtain a
sense of the individual risk profile of a particular CL. This technique isfrequently applied to CL
incorporating credit risk, whose behavior is a function of the general state of the economy. However,
econometric andfinancial models also require substantial data inputs, do not provide an estimation of the
variance of CL costs, and do not offer the mínimum variance resulting from large samples used in the
actuarial technique.
Since both actuarial and econometric models focus on the mean of the CL costs, these techniques
offer limited information to computerisk.An alternative approach based on option valuation techniques
presents a richer description of the behavior of CL and provides an estímate of the variance of the
distribution, which is the key component to valuérisk.Two of the techniques used to valué options that
have been applied to CL are contingent claim analysis and simulation analysis.
Contingent claim analysis is based on the concept that a CL is an option, in the sense that it can be
exercised if a certain event occurs. The distribution of the pay-out for the government issuing a CL
therefore looks like the truncated distribution of an option pay-out. Consider, for instance, a credit
guarantee granted by the government on a loan to a SOE. Theriskto the government is that the SOE may
experiencefinancial diffículties to the point where it is forced to discontinué its debt servicing, thus
forcing the lender to cali the guarantee and transfer the loan obligation to the government.
This example can be illustrated in the following graphs. In graph 1, the profits of the SOE are
depicted as the risk factor that may trigger the guarantee. It is assumed that profits are normally
distributed about a mean, assumed to be 1 in this case. At a point X, the losses experienced by the SOE
will absorb thefirm's equity, forcing the SOE to default on its debt and causing the lender to exercise the
option. Graph 2 displays the distribution of the government cash flow. If the option is not exercised, the
cash outflow will be zero; if it is exercised, the cash outflow is equal to the valué of the guaranteed loan,
less any recoveryfrom the SOE. For simplicity's sake, it is assumed that the government's net exposure
under the guarantee is equal to any losses realized by the SOE in excess of the equity. The distribution of
the government's potential exposure is therefore the same as the distribution of the SOE's income
truncated at point X.

Sesión 3 - 3 1

GRAPH1
S O E P r o » (Loss) Distribution

GRAPH 2
Government's Payout Distribution

-«

-3

-3

-1

gQ
Payout x

1

2

3

4

0

«

Thus, as graph 2 illustrates, contingent claim analysis not only gives the expected valué of the
govemment losses as the area under the curve, but also the dispersión i.e. the risk of facing unexpected
losses. Before referring to the decomposition of a CL valué into the expected valué and ariskpremium, it
is important to understand how a CL can be valued.
A useful perspective for this purpose is that of the lender. From his point of view, if the SOE does
not pay, the govemment does -guarantees are options that protect lenders against default. Thus, a loan to
the SOE backed by a guarantee is like having a risk free loan. In this sense, the govemment guarantee can
be valued as the difference between the price of ariskyand the price of ariskfree loan. Altematively, if
those loans are traded in the market, the put option can be valued based on the valué of the loan, its price
variance, the term of the option and theriskfree interest rate This is how contingent claim analysis can be
used to valué a credit guarantee or in general any CL.
The advantage of this type of model is that the valué of the option can be calculated directíy from
the valué and the variance of an underlying asset, in this case the loan being guaranteed, the term of the
option and therisk-free interest rate. This has enabled the application of the technique to valué a variety of
CL, such as loan guarantees, guarantees supporting corporate failure, mortgage guarantees and insurance
products. In the USA for instance, the Office of Management and Budget applies option technology to

Sesión 3 - 32

estímate future expenses derived from such CL as deposit insurance, pensión insurance and mortgage
guarantees.
A more important advantage of these models is that they allow to decompose the valué of the CL
into the expected valué and a risk premium. The difference between the pricing that results from the
option model and the expected cost is precisely the market valué of risk.In a simulation model this can be
thought of as the market valué of the capital allocated for risk, while the expected cost is similar to
provisioning expenses.
To better illustrate the application of estimating both the expected valué and the variance of the cost
of the CL, the same example of a credit guarantee given to a lender in favor of an SOE might be
considered. In effect, the above-mentioned credit guarantee can be treated as if the govemment was
borrowing the funds and on-lending them to a SOE. This unbundling highlights the role of the govemment
as a financial intermediary. According to the ALMframework, a bank in a similar situation should deal
with creditriskin order to price the loan to the SOE: moreover, it should first provisión against expected
losses, then allocate capital against unexpected losses and, with that information, compute the risk
premium on the operation. Replicating the same principie, the govemment could build budget provisions,
set aside reserves andfinally calcúlate the trae cost of the CL16
Options therefore can be seen as simulation models that allow aricherdescription of the behavior
of CL. Actuarial techniques, as well as econometric and financial models, valué CL on the basis of the
expected valué of the losses that the implicit or explicit commitment entails for the govemment. Options
not only provide the expected valué of losses, but also the variance of the distribution of govemment
losses, which is a more appropriate measure of risk.Henee, actuarial techniques as well as econometric
andfinancial models can be seen as special cases of option simulation.
Nonetheless, it should be recognized that data may not be available for most CL. In fact, if there
were a market for the majority of the underlying assets in which the CL could be unbundled, there would
probably be little need for the govemment to accept those CL and the corresponding risk would be
handled by the prívate sectorfrom the start. Data availability therefore is the main limitation to derive the
price of CL directlyfrom option pricing models.
Still, if options cannot be applied directly, they can be recreated from a simulation of the risk
factors that trigger CL. In the illustration given above, this is like simulating the distribution of the SOE's
profits depicted in graph 1, and from there deriving the CL pay-out shown in graph 2, rather than
assuming that the market will provide it.
This simulation technique could then be used for a number of CL, replicating the distribution of
crop results,financial system performance, occurrence of natural disasters, and so on. Nonetheless, in
order to simúlate them, information on the parameters of such distributions is required. It can be found in
historícal statistics collected by the govemment, or from evidence of other countries that have dealt with
those CL for longer periods. In the absence of statistics from which the mean and the variance can be
derived, it will still be possible to recur to a Bayesian type of simulation in which the distribution of a
occurrence of a natural disaster, for instance, collapses into the probabilities of binary events: either the
disaster happens, or it does not.

Section 6 addresses t e implications t a contingent claim analysis has for risk m n g m n purposes, namely: (i) if t e
h
ht
a ae e t
h
g v m e t should charge t e beneficiaries t e "fair valué" of t e CL, (ii) if t e govemment should provisión expenses in t e
oe mn
h
h
h
h
h
budget for a a o n equal to t e expected cost of the CL, a d (iii) if t e govemment should set aside reserves againstrisk,or if
n mu t
h
n
h
there a alternative a d less expensive way to deal with this issue.
n
n

16

Sesión 3 - 323

One important aspect is that the simulation techniques described here are inherently part of the
same type of techniques used to measureriskin the broader balance sheet context. When modeling risk
for debt service relative to the government budget, for instance, it is often difficult to have a direct
estimation of the debt service volatility vis-a-vis government revenues. Instead, debt service has to be
simulated from a given distribution of exchange rates and interest rates, for which a common approach is
to derive the distribution parameters as a combination of historical information and judgmental input.
On the other hand, the simulation of any CL, like SOE profits in the above mentioned example, is a
function of economic andfinancial variables that are the same ones that generate risk for other balance
sheet items i.e. interest and exchange rates and macro-economic shocks. In theory, it would therefore be
possible to find the elasticity of the valué of a CL to changes in the independent variables and therefore
quantify the portfolio effects of CL. With this information, it would be feasible to design a portfolio
composition that uses diversification to minimize overall risk.
These considerations also have organizational implications as the teams in charge of risk
measurement for debt are already familiar with simulation techniques and with the factors that generate
risk. In this sense, and as will be suggested in section 7, there might be some advantages in having a
centralized ALM team expanding the use of riskanalysis techniques to the entire government's balance
sheet and eventually to advice on macro hedging for the residual risk that results from the interaction of
the different balance sheet items.
To summarize, contingent claim analysis is a technique that valúes CL either directly, or by
simulating the spectrum of developments that may cause the CL to be triggered. It allows to estímate not
only the mean, but also the variance of the potential CL costs. On the one hand, this feature explains why
contingent claim analysis tends to produce higher estimates for the CL valué compared to the more
traditional techniques whose valuation is centered exclusively on the mean. On the other hand, the
analysis provides an estímate of risk as the variance of the distribution of the government expenditure,
which corresponds to the difference between the CL valué minus the expected cost of the CL. Since very
few of those CL have underlying assets that actually trade in the market, direct computation of the CL
price and its risk from option modeling might be the exception. More generally, techniques that allow to
simúlate theriskfactors that trigger CL are required. Not only are these simulation techniques part of the
same technology used to quantify risk for regular assets and liabilities, but also the risk factors macroeconomic shocks andfinancial variables- are the same ones that endanger the government objective
of tax stability and can ultimately produce a default on the government's financial obligations. This
commonality suggests the possibility of finding the elasticity of CL to thoseriskfactors, in order to better
understand their portfolio effects and to allow the construction of a portfolio that minimizes theriskof
asset and liability mismatches.

VI.

Risk Management

Thefirst step in defining a CLrisk-managementpolicy is to determine the risk exposure that the
government is willing to take. It is a political economy decisión on the part of the central government as to
whether it will manage only its own narrow balance-sheetrisk,or whether it will also cover some of the
risks generated in other parts of the public sector, and indeed in the private sector. Systemic risk
originating outside of the central government balance sheet will most likely be covered, even if it
originates in the private sector. However, there is an ongoing discussion with regards to the appropriate
limits of the central government's granting further riskinsurance and guarantees, e.g. in order to promote
growth and development. This topic goes beyond this paper, which limits itself to underlining that there

Sesión 3 - 32

should be a strategy for dealing with CL risk, and that risks taken on by the central and sub-national
govemment should be addressed explicitly and systematically and be incorporated into a broader balance
sheet management of sovereign risk.
These strategies are precisely what most governments of non-industrialized economies lack. In a
sense, CL in emerging economies have not been the result of deliberate political economy decisions, but
rather have been accumulated "passively" as a result of the lack of explicit policies addressing sovereign
risk management, combined with budgetary systems that favor CL over other forms of govemment
financial support. In these cases, governments need to measure the CL risk inherited in that manner,
evalúate their ownrisk-managementcapacity, and decide to what extent they will continué using CL in
the future. If the analysis shows that CL are inefficient vehicles of public finance vis a vis other
instruments and excessivelyriskyfor the govemment, it may be highly convenient for the govemment to
implement an explicit policy which severely restricts the issuance and acceptance of CL. Countries like
New Zealand and Argentina claim that not evenfinancial system problems represent a CL for the
govemment. These are examples of explicit political economy decisions to limit moral hazard by not
accepting certain CL, as are the cases of debt defaults by the City of New York in the 70's and, more
recentíy, by Gitic in China, where the governments did not intervene in order to avoid default.
For those governments that either accept to cover some of therisksgenerated in other parts of the
public sector, and in the prívate sector, or that inherit a stock of CL from previous administrations, there
are two broad approaches to risk management: decentralized or centralized. The difference between the
two refers to the existence of a body in the central govemment that oversees risk for the entire balance
sheet, and helps designrisk-managementpolicies both for the central govemment and for other entities
whose risk is being covered. As will be shown in sectíon 8, the United States is an example of
decentralized management with a number of regulatoiy bodies that controlriskfor different CL without a
central coordinating entity.
Centralization, however, seems to be the path followed by countries that did experience severe
difficulties associated with debt management. Countries like New Zealand, Belgium, and Ireland, that
experienced severe fiscal crises and high indebtedness which put macro stability at risk in the 70's and
80's, decided to create a debt office that was assigned the responsibility of managing what was then
considered to be the biggestriskfor the govemment: sovereign debt. Partly as a result of their success, and
partly because of economies of scale, some of these same offices subsequently expanded their scope of
influence, in order to deal with theriskof the entire govemment balance sheet, including CL. This trend,
and the advantages in terms of positive extemalities, have encouraged other countries to adopt a similar
centralized strategy, most notably Colombia and South Africa.
It is worth noting that economies of scale may be significant. In effect, and as discussed in sectíon
5, the simulation techniques used for valuing and quantifying CL risk are the same type of techniques
used to measureriskin a broader balance sheet context and in particular for sovereign debt. Therefore, the
same skills and technology are required in both cases. Investment in human capital and systems for an
adequate functioning of a debt office may take various years, and once it is available it makes sense to
extend it into an ALM office.
Thus, the trend for centralized risk management can be related to the size of the risks being
managed and probably to the economies of scale that result from extending to CL the infrastructure
available for risk management of sovereign debt. Both conditions seem to take place in most emerging
economies. The more vulnerable the sovereign is to shocks and the weaker itsrisk-managementcapacity,
as well as that of the rest of the public sector and indeed the prívate sector, the more stringent central
govemment CL guidelines and monitoring may have to be. The central govemment may end up not only
monitoring, and even managing residual balance-sheet risks, but also promoting a risk-management

Sesión 3 - 325

culture in the rest of the public sector. In this sense, it may want to bring out guidelines for CL risk
identification, quantification and management. This may be especially applicable in the case of developing
countries. In some cases, the central government may wish to go beyond this and to manage residual
public-sector risk, risk that remains unhedged at the sub-national or SOE level. This allows the central
government to address risk from a portfolio perspective, taking into account correlations between the
different types of risks.
No matter whether risk management is centralized or decentralized, moral hazard must be
contained. The policies in place should therefore ensure that beneficiaries of government backing will
behave as if the support had not existed, and that they will not undertake excessive risk increasing the
exposure of the guarantor of last resort. Two types of policies could be considered depending upon the
degree of influence the central government has on CL beneficiaries: policies designed for the public
sector, and policies directed to thefinancial system.
The following principies could be useful for containingriskof CL granted to the public sector:
i).
ii).

iii).

iv).
v).

Promote disclosure and accountability with regards to all CL
Contémplaterisk-sharingmechanisms and, whenever possible, refrain from coveringrisksthat are
under the control of the beneficiary, for this creates moral hazard. One way of risksharing is pricing
the CL and charging the corresponding valué to the beneficiaries. Another way is to establish
collateral requirements from the beneficiaries. In general, blanket risk guarantees should be
unbundled into their different dimensions, in order to determine the best way to implement risk
sharing among the different parties. It is convenient to determine what party can best manage
specific types of risk. Important progress in this area has been made in the domain of project
finance, and valuable insights can be incorporated into public sectorriskmanagement.
Promote good governance of sub-national entities and SOE's, by delegatingriskand promoting the
ALM balance sheet approach described above at a sub-portfolio level, making sub-national entities
accountable for managing their own risk and CL guidelines for state and municipal governments,
whereby theirfinancial autonomy is lost once a pre-established and quantifiable level of CLriskis
reached.
Ensure that CL compete on equal footing (e.g. in budgetary terms) with other forms of financial
support, such as direct subsidies, tax exemptions, loans, etc, so that CL choices does not lessen
publicfinance efficiency
Provisión expected costs in the budget to make CL costs visible and in some cases build reserves
against unexpected losses, possibly using the risk quantification methodologies mentioned above.
Provisioning of expected losses should always be made, but the level of reserves will depend on the
sovereign'sriskaversión and its ability to withstand unexpected losses.

For the decentralized government sector, the experience of industrialized countries shows that proper
accounting, using a medium-term horizon, with controls on direct and contingent indebtedness, disclosure
of the financial information, performance measurement and accountability, have significantly reduced if
not eliminated theriskof these CL.
In those cases where the government cannot promote the former principies because the CL is implicit
and there is limited control, as is the case of thefinancial sector, the central government should encourage
good governance and risk management with an effective regulatory structure and tight supervisión. This
includes specific limits onrisktaking, minimum capital requirements, and periodic reports to an oversight
institution for monitoring purposes. A sound financial system reduces the credit risk of the sector,
diminishes the Iikelihood of a government intervention and the size of the CL. In tum, however, the good
health of thefinancial sector depends on prevailing macro-economic policies and on risk management
practices in the prívate sector.

Sesión 3 - 326

The central government should also take the necessary actions to avoid the extensión of systemic risk
beyond thefinancial sector; again in this case, risk can be partly controlled a priori with effective
regulatory structure and supervisión on the involvement of thefinancial system with the corporate sector.
The Glass Steagal act in the United States can be understood as an example of restriction of implicit CL to
the banking system, whereas the involvement of the Korean government in the rescue of the Chaebols
offers an illustration of a systemic risk originated in the private sector that surpasses the limits of the
financial system. As mentioned, additionalriskcoverage by the central government, beyond systemic risk,
is a matter of political economy policy. In any case, those policies should take into account their complete
fiscal implications and be expressed in a set of clear rules that govern the sovereign's commitments.

VII. Organizational implications
A comprehensive sovereign ALM framework has important organizational implications for the
government. For some countries, it shows the convenience of having some kind of centralized
arrangement that systematically addresses CLrisk-issuesin a transparent and explicit manner, such as:
i) a formal asset-liability management unit (e.g. as in New Zealand or South Africa) that analyses
the sovereign's balance sheetrisk,and
ii) independent centralized operational units, that nonetheless feed information into a centralized
sovereign risk evaluation, possibly implemented by the debt management office, where the skills
and techniques are the same for debt management and for sovereignriskmanagement
A centralized organizational scheme of this sort may be particularly helpfiil for developing
countries, with the particular kind of vulnerabilities mentioned above, and with low tolerance for risk.
However, otherframeworks can be equally useful in countries with other characteristics. For example, the
USA does not have a centralized ALM monitoring of CL arising from thefinancial sector, but rather a
decentralized scheme monitored and supervised by different agencies. However, even the USA tries to
centralize an evaluation of guarantees andrisk-coveringinstruments in the Department of Treasury, in
order to standardize budgeting procedures and better control their budgetary impact. Although the best
organizational arrangement for any given country may depend on the macroeconomic context, on risk
definition, on technical capacity for risk-management,and on the type of institutional and political
arrangements, such as the degree of political decentralization, recent country experiences show a trend
towards centralized arrangements. As explained, this trend is related to the size of therisksbeen managed
and economies of scale present in the process.
Whatever the organizational arrangement decided on, the most important element is that it should
contribute to the institutional exercise of ALM and include CL within therisk-managementexercise ( e.g.
promote clear objectives, such as avoiding budgetary shocks and protecting net worth; establishment of
clear guidelines for managing CL risk,
assure mechanisms for quantifying risk-retumtrade-offs of CL;risk-identification using a conceptual
balance sheets, periodic reporting to higher authorities within a clear system of delegation, a
comprehensive information system on CL, etc.)
Typically, this will imply the need to strengthen the strategic analytical units, or the "middle
offices", responsible for risk analysis and policy. In the case of the debt management offices it is
important that the middle office group not be responsible for operational matters, such as funding,

Sesión 3 - 32

extending guarantee contracts or on-lending to SOE. CL should be one of the most important elements to
be analyzed by the middle office staff.
One organizational mechanism which helps promote CL risk management within an ALM
framework is to create a Sovereign Risk Management Committee, that would establish basic policies and
monitor sovereignriskon a periodic basis, from an overall sovereign perspective. Again, its characteristics
would be different depending on how risk is being processed, monitored and managed within the
govemment. A committee of this type might be more important in those cases whererisk-managementis
more delegated and/or where a large public sector means that the layering of different kinds of risksis
very complex, and/or where specific sub-national interests are very strong.
In some countries, the debt management office has a significant role to play in monitoring and
managing an important part of sovereignrisk.This is precisely because of the fact that it directly manages
part of the liabilities, is frequently involved in granting guarantees to sub-national govemment, and has
staff with the necessary skills for managing risk. However, it must closely coordínate with the fiscal
authorities and the central bank in monitoring the sovereign balance sheet as a whole.
There would probably be a need to coordínate between various govemment units in order to assure
the existence of a comprehensive statistical base that would enable the necessary monitoring on the part of
govemment agencies and supervisory entities. The importance of this base tends to be underestimated, in
view of its significance for sovereign risk management. As mentioned before, this may involve statistics
from both the public and the prívate sector.
Needless to say, it is essential for top govemment officials to give full support to CL risk
management within an ALM framework and its policy and organizational implementation. Without
Ministerial backing, the whole exercise will be unsuccessful, because it requires significant effort and
financial resources for aspects such as information systems improvement, reporting, creation of middle
offices and training of their staff in risk-management.

VIII. Examples of CLriskmanagement with elements of sovereign ALM
A. New Zealand
The context is one of fiscal control, and govemment balance-sheet risk management, and is
interesting inasmuch as it forces political accountability and transparency. The Fiscal Responsibility Act
of 1994 requires the govemment to follow a set of principies of responsible fiscal management, under
conditions of transparency and accountability including: i) management of the risks facing the
govemment, and ii) achieving and maintaining govemment net worth, albeit defined in strictly accounting
terms.
These balance-sheetriskmanagement objectives were reflected in the organizational reforms of the
public sector. The Treasury's operational responsibilities now include, among others, i) liabilities
management, ii) monitoring Crown assets, iii) the management and production of a complete balance
sheet of the public sector. The Asset and Liability Management Branch, under direct responsibility to the
Secretary of the Treasury, advises on govemmentfinancial policy as it relates to govemment ownership
and balance-sheet interests. It manages the Crown's sovereign debt portfolio, provides advice on the
performance and management of ownershiprisksin SOE and advises on the govemance of public sector
entities. It is responsible for producing the Financial Statements of the Government of New Zealand.
These inform Parliament how public resources are being used, and report the government's assets and

Sesión 3 - 328

liabilities, together with its revenue, expenses and cash flows, and contingent liabilities and commitments.
Both quantifiable and non-quantifiable CL are closely monitored.
Within this framework, the government takes a series of measures to contain moral hazard and limit
CLriskto the central government. For example, SOE are required to opérate on the basis of principies
and procedures contained in the State-Owned Enterprises Act 1986, that promotes efficiency, commercial
criteria and autonomy from the central government. The SOE borrow in their own ñames and on their own
credit, in most cases without a guarantee or other form of credit support from the Government.
Government policy requires that SOE disclaim in loan documentation the existence of such guarantees or
credit supports.
This organizational and policy ALM framework allows CL to be disclosed and analyzed, and
generally contained at a sub-national level, as central government intervention is strictly limited.
However, the emphasis tends to be on disclosure in accounting terms, and no evaluation is made of
expected or unexpected costs, ñor of provisioning from the budget and reserve creation. CL are given an
accounting treatment and left as an off-balance sheet item, until they are triggered and only then are they
incorporated within the balance sheet. Thus, the cost to policy makers is couched more in terms of
political scrutiny, and less in terms of financial, budgetary cost.

B. United States
The objectives of the Federal Credit Reform Act of 1990 are those of revealing long-term program
costs, to compare the cost of altemative forms of government assistance, to compare the benefit to the
borrower vs. the cost to the government, and to make the Department of Treasury ultimately responsible.
Thus, there is disclosure and transparency with new loans and guarantees and CL in general. In addition,
the Treasury must budget for the estimated present valué cost of new obligations, prior to the approval of
new CL such as loan guarantees, as well as new direct loans. Thus, policy-makers feel a direct financial,
budgetary impact of CL, and these are placed on the same footing as other forms of central government
financial assistance.
One of the consequences is that programs are better structured to increase cost-sharing with
borrowers, andrisk-sharingwith creditors. A medium-term horizon is used, and accrual accounting helps
direct the attention to the effects of CL in terms of medium-term cash flow risk(expressed in discounted
terms).A cash-based system of budgeting creates biases, in terms of just accounting for expected cash
outlays associated.
One disadvantage is that the problem of variance around expected costs, the coverage of unexpected
losses is not included as part of the cost of a program.

C. The Netherlands
The Netherlands has set a single budget ceiling within each sector for expenditures and guarantees. The
sector decides on the form of assistance within the joint ceiling. The objective is to make sector ministries
cash-neutral in choosing policy instruments e.g. between direct loans and guarantees, and to minimize the
government'sriskexposure.
Guarantees and their risk analysis need to be approved by the MoF. The government guarantees are
disclosed in terms of face valué (in order to disincentive their use), and the full face amount of the
government obligation is transferred from the sectoral budget allocation to a central reserve fund.
Sesión 3 - 32

D. Cañada17
CL by federal departments and Crown corporations in Cañada amounted to nearly $70 billion in
1997-98, representing about 8% of GDP. About two-thirds of the CL relate to loan and loan guarantees,
mostly for borrowing undertaken by Crown corporations. The government's current loan and loan
guarantee policy was established in 1986, as a response to significant losses incurred in the early 1980's
mostly due to ad hoc loan guarantee programs which resulted in large cash payouts. The govemment
began to implement accmal accounting with the explicit recognition that allowances need to be
established for those cases where expected losses with respect to contingent liabilities could be reasonably
estimated.
The policy is based on the following principies:
Budgetary recognition of subsidies and concessions: with respect to loans, any interest-rate subsidy or
other concessions are treated as part of budgetary expenditures
• Demonstrated need: the sponsoring department must demónstrate that the project could not be
financed on reasonable terms and conditions without a govemment loan or guarantee
• Satisfactory rate of return: an economic analysis is made to demónstrate that the project's expected
cash flow is adequate to cover repayment of the guaranteed debt, interest and operating cost and yield
a satisfactory rate of retum
• Real prívate sector equity at risk: no project is considered for a govemment guarantee where prívate
equity sponsors are not supplying a substantial portion of funds required from their own resources.
• Lender risk-sharing: bankers should be prepared to lend funds under conditions where they would
bear at least 15% of the net loss associated with any default
• Upside benefit: where the govemment is requested to bear significant downsiderisks,considerations
should be given quid pro quo on the upside should the project prove to be successful.
• Cost recovery: fees are set for loan guarantees to recover the estimated cost of future losses and help
cover administrative costs
• Up-front provisioning for all loans and loan guarantees: the latter will be provisioned up-front at the
time that they are made. The provisión will be determined based on theriskassessment. Sponsoring
departments are required to fund these provisionsfrom either the fees they charge orfrom their annual
appropriations.
• Approval ofMoF and Parliament: all new loan and loan guarantee programs and changes to existing
programs must be approved by the Minister of Finance. Parliamentary authority is required for all new
loan and loan guarantee programs. Such programs are subject to a limit that can only be changed by
Parliament.
• Ongoing review and audit: Departments and Crown corporations are required to submit regular
reports on their CL. These are published on an annual basis as notes to the government's financial
statements. The estimates of the CL and losses are audited annually by the Auditor General of Cañada,
who is independent of the govemment and reports directly to Parliament.
•

E. Colombia
The Constitution of 1991 allowed prívate capital in public services and infrastructure development,
and in order to encourage prívate sector participation, public entities guaranteed the servicing of projects'
financial liabilities, and then proceeded to grant even wider guarantees. Guaranteedrisksinclude demand
17

Resolving Institutional Problems of Contingent Government Liabilities, Presentation to PREM Seminar Washington D.C., July

14,1999

Sesión 3 - 330

or traffic, prices, revenue, cost overruns, inflation, exchange rates, tariff scheme and regulatory
frameworks. The disadvantages of this scheme included:
•
•

lack of an established policy over whichrisksshould be guaranteed
in thefirst generation of guarantees, private sector parties enjoyed retums higher than corresponded
with therisksundertaken
• crude methodologies were used to quantify CL and their possible effects for the nation's bud
• the budgetary framework favored the substitution of explicit liabilities by implicit ones.
budgetary period is of only one year, so that thefinancial impact of guarantees is only reflect
partially in the year in which they are given. This is an incentive to use instruments that d
require an immediate budgetary register, and allows the evasión of the real costs of CL
By 1997, CL in these sectors summed 1.6% of GDP, equivalent to 20% of that year's total investment
in infrastructure. In the last three years, actual guarantee payments have exceeded the budgeted amount by

118%.

A law was passed in 1998 (Law 448) in order to increase the budgetary discipline of these CL, to
recognize and valué CL, to create the Contingent Fund of Public Entities and to assure that transfers to this
Fund had the same priority as debt servicing. Thus, a formal and standardized procedure to valué and
budget these liabilities is being designed. The General Directorate of Public Credit must approve the
methodology for valuing the CL. Public entities must do yearly transfers of resources to the Contingency
Fund, according to a specific payment plan, in order to cover expected losses of the CL. The liability
valué, payment plan and deposits are adjusted to reflect the remaining life of the contract. If the CL were
to be called, the entities have access to the accumulated deposits. This scheme thus improves the liquidity
of guarantees.
In the interim, the central government must include guarantees within the credit ceiling limits
approved by Congress. At the sub-national level, guarantees have to be included in the indebtedness
Colombia's efforts to improve on the management of CL have also included the draw up of an
economic balance sheet. Based on Bill Easterly paper18 which emphasizes on the need to measure the
budget déficit looking at the public sector net worth instead of focussing on the change in public sector
liabilities, Juan Carlos Echeverry, from the national Planning Department, tests the sustainability of the
Colombian fiscal situation by estimating a complete balance sheet for the government, including
contingent liabilities19.
Data of the general balance sheet elaborated by the Nation's General Accountant for 1997 are used as
a baseline, and then adjusted for coverage and errors. Some items are recalculated, while others -mainly
contingent liabilities- are included for the first time. The public net worth is then computed as the
difference between assets and liabilities, desegregating short term assets and liabilities tofind a liquidity
acid test.
Thefindings show that a more complete analysis of the balance sheet would be less optimistic than
the official estimates: a positive net worth of 62% of GDP according to the official figures, would be
substituted by a negative net worth representing 69% of GDP, mainly as a result of the recalculation of
pensions. This clearly shows that thefiscal situation in the médium and long term is unsustainable and that
drastic structural reforms are needed. However in the short term the situation does not seem to be so
critical since short term assets are greater than short term liabilities by 1% of GDP.
18
19

See Easterly (1998)
See Echeverry (1999)
Sesión 3 - 3 3

Some of the reforms suggested along the lines of this study are the reform of the pensión system,
particularly the system covering education and health sectors, the creation of the Fund for Territorial
Liabilities so that sub-nationals will be obliged to fund and make provisions for their obligations,
downsizing the public sector by merging and eliminating entities, and the increase in taxes in order to deal
withfinancial crisis and with natural disasters.

Sesión 3 - 33

IX.

References

Alessina Alberto et al 1992 Default Risk, Economic Policy, October
Bitner, John 1992 Successful Bank Asset Liability Management: a Guide to the future GAP.. John Wiley
and Sons
Barro 1995. "Optimal Debt Management" NBER Working Paper No5327, October
Easterly, William. 1998. "When Is Fiscal Adjustment an Illusion?" World Bank, Washington, D.C.
Economic Policy, April.
Echeverry, Juan 1999. "El Balance del Sector Publico y la Sostenibilidad Fiscal en Colombia" Paper
presented in cooperation with the World Bank by the Colombian National Planning Department,
June.
Fabozzi Frank and Konishi Atsuo 1990Asset Liability Management. Probus Publishing Co, Chicago II.
Kharas, Homi, and Deepak Mishra. 1999. "Hidden Déficits and Currency Crises." EASPR, World Bank,
Washington, D.C. Forthcoming
Missale Alessandro 1997 Managing the Public Debt: the Optimal Taxation Approach. Journal of
Economic Surveys Vol 11 No3
Mody, Ashoka and Cristopher Lewis 1997 " The Management of Contingent Liabilities: a Risk
Management Framework for National Govemments" In Irwin, Timothy, and others, eds. 1997.
Dealing with Public Risk in Prívate Infrastructure. World Bank Latin American and Caribbean
Studies,. Washington, D.C.
Mody, Ashoka and Dilip K. Patro. 1996. "Valuing and Accounting for Loan Guarantees." The World
Bank Research Observer 11(1)( February).
Polackova, Hana. 1998. "Contingent Government Liabilities: a Hidden Risk for Fiscal Stability." Policy
Research Working Paper 1989, World Bank, Washington, D.C., October 1998
Polackova, Hana, Papp A., and Schick Alien. 1999a. "Hungary: Fiscal Risks and the Quality of Fiscal
Adjustment." Policy Research Working Paper 2176. World Bank, Washington, D.C., September
1999
Polackova, Hana. 1999b. "Fiscal Adjustment and Contingent Government Liabilities: Case Studies of the
Czech Republic and Macedonia." Policy Research Working Paper 2177, World Bank,
Washington, D.C., October 1999
Schick, Alien 1999.. Budgeting for Risk. mimeo Washington, D.C.: World Bank.
Talvi, Ernesto and Vegh Carlos 1998. Tax base variability and Pro-cyclical Policy. Mimeo. Office of the
Chief Economist, Inter-American development Bank

Sesión 3 - 333

Sesión 3 - 334

SESION 4
PAPEL, ALCANCE Y COSTOS FISCALES
DE LAS REDES DE PROTECCION SOCIAL

Sesión 4 - 335

Sesión 4 - 336

SOCIAL ISSUES IN IMF - SUPPORTED PROGRAMS

by
Sanjeev Gupta, Lotiis Dicks-Mireaux, Ritha Khemani,
Calvin McDonald, and Marijn Verhoeven *

* This paper is being published in the IMF's Occasional Paper series. At the CEPAL seminar, it is being
presented by Adrienne Cheasty. A number of IMF staff guided and contributed to the paper. In particular,
the authors would like to thank Jack Boorman, Vito Tanzi, Peter Heller, Thomas Leddy, Ke-young Chu,
and Russell Kincaid for substantial comments on earlier drafts; Hamid Davoodi, Frank Engels, Tetsuya
Konuki, Henry Ma, and Gustavo Yamada for contributions to various sections of the paper; and
Randa Sab and Erwin Tiongson for providing statistical assistance. The authors are also grateful to staff of
departments within the IMF and the World Bank for feedback on the earlier versions. Cecilia Pineda,
Larry Hartwig, and Cecilia Lon performed the administrative tasks related to the production of the paper
and Jeanette Morrison of the IMF's External Relations Department edited the Occasional Paper and
coordinated its production

Sesión

- 33

Table of Contente
I.

Introductíon

342

II. Evolution ofthe IMF's Policy Advice

343

A. Two Social Sector Issues......
B. Key Steps in the Evolution—

344
344

III. Social Safety Nets

347

IV. Public Spending on Education and Health Care

A.
B.
C.
D.

352

Aggregate Spending on Education and Health Care
356
Composition of Spending
358
Impact on Education and Health Indicators and Implications for Poverty........................360
Experience with Program Targets, Conditionality, and Monitoring in ESAF Countries .362

V. Collaboration with the World Bank and Other International Agencies

369

A. Postscript

371

References

372

IMF Executive Board Discussion

376

Glossary

Boxes

378

2.1.
2.2.
3.1.
3.2.
4.1.

Evolution of the IMF's Social Policy AdviceS
International Social Development Goals and Performance Indicators
Strengthening Social Safety Nets in ESAF-Supported Programs, 1994-98
Social Safety Nets: Issues in Transition Economies
Quality of Social Spending and Indicators Data

4.2.

Targets for Public Spending on Education and Health Care in ESAF-Supported Programs 94-98

Tables
4.1.
4.2.
4.3.
4.4.
4.5.

Improvement in Social Indicators in Member Countries, 1985-97
Social Development Indicators, Selected ESAF Countries
Monitoring Arrangements, Selected ESAF Countries
Public Education and Health Care Spending Targets, Selected ESAF Countries
Social Policy Conditionality (Prior Actions, Performance Criteria, and Benchmarks), Selected
ESAF Countries

Figures

4.1.

4.2.
4.3.
4.4.

Changes in Education and Health Care Spending in Selected Countries with IMF-Supported
Programs, 1985-97
Spending Levels on Education and Health Care in Countries with IMF-Supported Programs, 1997
Allocation of Education and Health Care Spending in Countries with IMF-Supported Programs,
1994
Benefit Incidence of Public Spending on Education and Health Care in Countries with IMFSupported Programs, Early 1990s

Sesión

- 33

Preface
Ai part of its mandate, the IMF seeks to create the conditions necessary for
sustained high-quality growth, which encompasses a broad range of elements. These include
sound macroeconomic policies, growth-enhancing structural reforms, good govemance, and
social policies, such as cost-effective social safety nets and targeted social expenditures.
Over the years, the IMF and the international community have increasingly recognized that
macroeconomic and structural policies have important social implications, which in turn
have ramifications for the domestic ownership of economic and reform agendas and
promoting sustainable growth. In particular, social safety nets that are aimed at cushioning
the adverse social impact of adjustment programs on vulnerable groups have become
increasingly important. Also, there has been greater focus on the prioritization of public
spending on areas of social policies that promote growth through poverty reduction,
specifically with regarás to increased access of the poor to education, health, and economic
opportunity. The IMF has therefore taken a progressively more active stance on social
policies to ensure that they are well integrated into IMF-supported programs and IMF
policy advice. A key element of the IMF's involvement in social policies has been
cóllaboration with other relevant international agencies, especially the World Bank, in order
to draw upon their social policy expertise.
This paper—the most recent of a series of reviews of the IMF's involvement in social
issues undertaken since the late 1970s—reviews the IMF's policy advice in two key areas of
social policy: social safety nets and public spending on education and health care. It was
initiated as part of the work by the World Bank and IMF to strengthen the poverty focus of
adjustment programs in low-income countries, in particular within the framework of the
Initiative for Heavily Indebted Poor Countries (HIPCs). This review looks at such social
policies in IMF-supported programs in two broad groups of countries. The first group
comprises a sample of up to 65 countries that implemented IMF-supported programs
between 1985-97. The second group comprises a sample of up to 12 countries that
implemented programs under the IMF's Enhanced Structural Adjustment Facility (ESAF),
including under the HIPC Initiative during the latter part of the 1990s. The paper pulís
together the main lessons from this experience and sets out proposals for strengthening the
social content of IMF-supported programs including through mechanisms to strengthen
World Bank—IMF cóllaboration in the social policy area. An earlier versión of the paper was
discussed by the IMF Executive Board on September 13, 1999; a summary of this Board
discussion by the Chairman appears following the review.
Soon after these and other discussions (at the IMF and World Bank) on how to
strengthen the poverty focus of adjustment programs in low-income countries, it was agreed
at the IMF—World Bank 1999 Annual Meeting to put poverty reduction at the centre of
adjustment programs supported by the international community in the poorest countries. To
implement this significant shijt in policies, poverty reduction was made a central objective of
the ESAF which was transformed in December 1999 into the Poverty Reduction and Growth
Facility (PRGF); this will also be a key element ofthe enhanced initiative for debt reduction.
A central feature ofthe PRGF is the Poverty Reduction Strategy Paper (PRSP)—one ofthe
policy recommendations of this review—which articúlate a country's comprehensive strategy
for poverty reduction and integróte macroeconomic, structural, and social policies. The
PRSP would be nationally-owned and prepared through a broad participatory approach.

Sesión 4 - 339

Overview
This paper reviews the IMF's policy advice in two key areas of social policy: social safety nets
and public spending on education and health care. While the IMF has been helping countries promote
sustainable economic growth and, thereby reduce poverty through macroeconomic policy advice, it has
also been strengthening its dialogue with member countries on the social implications of its advice. This
paper offers preliminary conclusions on how to improve the integration of IMF policy advice on social
safety nets and public social spending into program design1 within a sustainable macroeconomic
framework.
In the family of international organizations, the social components of country programs are
primarily the responsibility of the World Bank and other organizations, not the IMF. The World Bank has
primary mandates, responsibilities, and expertise on social issues. Whenever feasible, the IMF has drawn,
and will continué to draw, upon the work of the World Bank and other organizations. Henee, enhanced
inputs from and closer collaboration with these organizations are essential. Another important element is
more dialogue with civil society groups, in particular labor unions and nongovemmental organizations
(NGO)s.

Social safety nets
The design of social safety nets and the timing of their establishment in countries have been
influenced by both social protection needs and constraints. The needs reflected the specific adverse social
effects of reform measures and the characteristics of affected groups. The constraints reflected the
availability of social policy instruments such as oíd age pensions and unemployment insurance, and
administrative andfinancing capacity. Whenever social policy instruments were available, the foremost
challenges have been to ensure their targeting and to increase their financing.
This review identifies three key requirements for strengthening social safety nets in
IMF-supported programs:
•

more comprehensive ex ante analysis of the likely social impact of key macroeconomic and
structural reform measures; such analysis needs to be undertaken before or at the time of program
formulation;
•
adequate follow-up of performance and monitoring of safety nets during program implementation
to ensure that intended poor groups receive adequate support; and
•
introducing appropriate social policy instruments before the onset of crises and economic reforms.
IMF staff needs to rely on the expertise of the World Bank and other organizations in conducting
the ex ante analysis. IMF staff reports should discuss such analysis and also the performance of social
safety nets. When the World Bank or other relevant international institutions are unable to provide needed
advice within a suitable time frame, IMF staff should attempt to fill the gap. These situations, however,
should be infrequent.

Public Spending on Education and Health Care
On average, in the past decade, education and health care spending has increased—in real per
capita terms, as well as in relation to GDP—in countries with IMF-supported programs. For many
countries, these increases have been accompanied by improvements in a broad range of social indicators.
For program descriptions see the Glossary at the e d of this volume.
n

Sesión 4 - 3

Still, countries differ considerably on spending relative to GDP for both education and health care and on
the speed of improvement of social indicators, reflecting in part differences in the efficiency of public
spending.
There is scope for improving the efñciency and targeting of existing
spending on education and
health care as a means of improving social indicators. This improvement could be achieved through,
among other things, strengthening budget formulation and implementation capacity, increasing resources
spent on primary education and basic health care, and reducing excessive out-of-pocket expenses borne by
the poor in the form of user charges for primary education and basic health care. To consolidate the
progress already made, this review identifies some steps that should be taken in programs supported by the
IMF for:
•

establishing quantitative targets for education and health care spending more systematically,
particularly for primary education and for basic health care; these targets should be reported in
IMF staff papers for the Executive Board, and efforts should be made to strengthen the monitoring
of such spending;
•
occasionally setting performance criteria on minimum spending thresholds; and
•
in some circumstances, monitoring budgetary allocations for selected key inputs, such as books
and medicine (although an excessive level of detail in IMF-supported programs would be neither
feasible ñor appropriate).
These steps should be taken, in collaboration with the World Bank, by building on the progress
that has already been achieved. IMF staff should continué to assess budgetary allocations for social
sectors, relying on available World Bank input, in particular timely Public Expenditure Reviews (PERs).
To help promote social reform, IMF-supported programs could use as reference points the targets
established by the authorities for selected intermedíate social indicators (e.g., primary and secondary
school enrollment rates and immunization rates). Especially where social indicators are failing to improve,
despite increases in public spending, IMF staff should report to the Executive Board on discussions with
the country authorities, World Bank, NGOs, and other institutions.

World Bank-IMF collaboration
World Bank-IMF collaboration could be significantly improved by better integrating
macroeconomic and social objectives, policy measures, and related work agendas. A shared understanding
of the key social and macroeconomic issues is essential.
•

Such collaboration could take place through the formulating of a poverty reduction strategy
together with the country authorities in a participatoiy process. The main elements of the strategy
would be set forth in a Poverty Reduction Strategy Paper (PRSP), which would be endorsed by
the govemment, the World Bank, and the IMF. The PRSP would set out medium-term
macroeconomic, structural, and social policies consistent with the government's poverty reduction
objectives. An IMF- or Bank-supported program should be consistent with the policy framework
contained in the country-strategy paper.
•
When timely World Bank input is either not available or insufficient, program design should
allow for the fuller integration of relevant social policies at a later stage (e.g., at the time of
program reviews), as additional analysis becomes available.
The PRSP should include several components that would facilitate World Bank-IMF
collaboration. These components, which would reflect the two institutions' respective operational
responsibilities in a country, should contain policy advice,financing needs, and work programs, in

Sesión 4 - 341

particular in the context of IMF-supported programs and World Bank lending operations.2 Through this
process, an iterative dialogue between the staffs of the IMF and the World Bank would be intensified,
assuring the consistency between a macroeconomic framework and a cost-effective strategy for
sustainable growth with poverty reduction.
The social policy components of the countries' Comprehensive Development Frameworks (CDFs)
could also be integrated into their macroeconomic programs. In this regard, drawing upon the advice of
the World Bank, the authorities should formúlate, at an early stage of their macroeconomic programs,
comprehensive social strategies that include specific action plans that provide a much-needed road map
from objectives to policies.
High-level poverty monitoring units in governments could help strengthen coordination at the
local, national, and international levels and collect data for monitoring social progress.

Data and Institutional Capacity
IMF staff should make an effort to identify and highlight data weaknesses in the area of social
spending indicators and social protection arrangements. This would help draw the authorities' attention to
the urgent need to redress the data weaknesses in collaboration with the World Bank and other
international agencies. In this regard, IMF staff should also assess the scope for technical assistance.
Greater attention could also be given to inputs prepared by civic groups, NGOs, and donors.

I.

Introductíon

The IMF's mandate is, among other things, "to facilítate the expansión and balanced growth of
international trade, and to contribute thereby to the promotion and maintenance of high levels of
employment and real income . . . of all members as primary objectives of economic policy."3 To this end,
the IMF promotes sound macroeconomic policies, growth-enhancing structural reforms, and good social
policies—conditions for high-quality growth. The IMF has paid increasing attention to these
considerations in its policy advice.
This paper explores how attention to social issues can be accentuated. It is one of several papers
that respond to the request of G-7/G-8 Finance Ministers at the June 1999 Cologne Summit to the World
Bank and IMF to work together to strengthen social policies in the design of adjustment programs to
protect the most vulnerable and to develop an enhancedframework for poverty reduction.4 It looks at two
social policy issues that are important for economic reform and growth—social safety nets and public
social sector spending. It also identifies ways to better integrate sound social policies into the IMF's policy
advice and program design within a sustainable macroeconomic framework.5
2

3
4

5

A m r detailed discussion of the proposed Poverty Reduction Strategy Paper is provided in the World Bank-IMF staff paper,
oe
"HIPC Initiative—Strengthening t e Link Between Debt Relief and Poverty Reduction," which is available o t e Internet:
h
n h
http:/www.imf.org/extemal/np/hipc/hipc.htm. See also the Glossary at the end of this volume.
Article I (ii) of t e Articles of Agreement of t e International Monetary Fund.
h
h
See also t e World Bank paper, "Building Poverty Reduction Strategies in Developing Countries" a d t e joint World Bankh
n h
IMF paper, "HIPC Initiative—Strengthening the Link Between Debt Relief and Poverty Reduction." Both are available on the
Internet: http://www.imf.org/extemal/hipc/hipc.htm.
The scope of social policies, and the channels through which macroeconomic policies c n have a social impact, cover a broad
a
area The selective focus of this paper on social s f t nets and public spending on education and health care allows in-depth
ae y
consideration of a limited number of important social policies that are closely linked to t e design of economic programs. For
h

(continued...)
Sesión 4 - 342

This paper examines the experiences of two sets of sample country groups. The first set comprises
large samples of up to 65 countries with programs supported by the IMF, including the countries that
implemented Stand-By Arrangements and programs supported by the Enhanced Structural Adjustment
Facility (ESAF) from 1985 to 1997; the analysis aims to identify broad pattems and reach general
conclusions. The second set comprises samples of 11-12 countries that implemented ESAF programs,
including under the Initiative for Heavily Indebted Poor Countries (HIPC), during the latter half of
thel990s; the analysis focuses on the use and monitoring of program targets. (See the Glossary for
descriptions of these facilities). The availability of data influenced the choice of countries, which may
introduce some bias.
To a large extent, the analysis and policy advice on social issues lie outside the areas of IMF
expertise. IMF staff relies on inputs from other international agencies, notably the World Bank. Thus, this
paper also discusses collaboration with the World Bank and other agencies in the social policy sphere,
including on internationally accepted principies, goals, practices, and indicators, such as those developed
collaboratively by the Organization for Economic Cooperation and Development (OECD), United Nations
(UN), and World Bank.
The paper has five sections. Section II describes the evolution of the IMF's social policy advice.
Issues related to social safety nets as a mechanism to mitígate the immediate adverse effects of economic
crises and reform programs on poor groups are presented in Section ni. In Section IV we set out policy
issues concerning the adequate provisión of education and health care services—crucial for achieving
countries' social development goals. We discuss IMF collaboration with the World Bank and other
international organizations in Section V.

II. Evolution of the IMF's Policy Advice
Sound economic policies favor both growth and the poor. The contribution of macroeconomic and
structural reforms to long-run economic growth and poverty reduction is now well established. Research
has demonstrated that low fiscal déficits and price stability promote economic growth,6 and economic
growth is the most significant single element that contributes to poverty reduction.7 Macroeconomic
adjustment generally benefits the poor.8 Dismantling product and factor marketrigiditieshelps reduce
poverty by increasing not only the supply of essential goods, but also the poor's access to them.9 In

6

7
8
9

this reason, t e focus is on the expenditure side of t e budget, which has offered better opportunities t a the tax side for
h
h
hn
poverty reduction (see Harberger, 1998). Likewise, t e paper does not discuss t e social impact of t e IMF's macroeconomic
h
h
h
policy advice, per se, except to t e extent it bears directly on t e design of social policies; ñor does it at m t a broader review
h
h
te p
of IMF policy recommendations in t e social sphere in bilateral and multilateral surveillance.
h
As regardsfiscal déficits, see Fischer (1991); Levine and Zervos (1993); Easterly and Rebelo (1993); Bredenkamp and
Schadler (1999); and Goldsbrough a d others (1996); as regards inflation, see Barro (1995); Bruno and Easterly (1995); Sarel
n
(1996); and Ghosh and Phillips (1998). Macroeconomic stability—lower (and stable) inflation—has also been shown to be
conducive to higher long-run growth (World Bank, 1996).
World Bank (1996).
Demery and Squire's (1996) review of six African countries has shown t a macroeconomic adjustment has generally
ht
benefited t e poor.
h
Sahn, Dorosh, a d Younger (1997), in a comprehensive study (based o computable general equilibrium models of 1 subn
n
0
Saharan countries) concluded t a under structural adjustment programs supported by t e IMF and t e World Bank, m s of
ht
h
h
ot
t e poor experienced small net gains. They also show that structural reforms hurt those reaping rents from distortionary
h
policies who t n to be nonpoor.
ed

Sesión 4 - 34

addition, based on cross-country studies, there is increasing evidence that lower inflation also enhances
income equality (Milanovic, 1994; Bulír and Gulde, 1995; Sarel, 1997; Bulír, 1998; and Guitián, 1998).

A.

Two Social Sector Issues

Nevertheless, in the short run, measures needed for macroeconomic stability can adversely affect
some poor groups, while helping other such groups. For example, currency devaluations may hurt the
urban poor who consume imported grains, while helping low-income smallholders producing export crops
in rural areas. Mitigating the adverse effects of reform programs on poor groups should be an important
aspect of the IMF's policy advice and program design.
The size and quality of public social spending can affect long-run growth and poverty reduction.
The relationship between public social spending, growth, and poverty reduction, however, is complex and
dynamic, involving many other factors, including prívate spending on education and health care. Forging a
consensus on a proper balance between macroeconomic stabilization and sound public spending through a
participatory dialogue among the govemment, civil society, and the international community can be
facilitated by establishing social and poverty reduction programs that are integrated into a sustainable
medium-term budgetary framework.

B.

Key Steps in the Evolution

Over the years, the IMF has taken a progressively more active stance on social policies to ensure
that they are well integrated into IMF-supported programs and IMF policy advice. The IMF has
strengthened the integration of social policies into its operations by establishing the Structural Adjustment
Facility (SAF) in 1986 and its successor, the Enhanced Structural Adjustment Facility (ESAF) in 1987,
and the Initiative for Heavily Indebted Poor Countries (HIPCs) in 1996 (Box 2.1).10
A key element of these new instruments has been the collaborative role of the World Bank. The
IMF and the World Bank have collaborated through the Policy Framework Paper (PFP) in the ESAF and
through the joint HIPC Initiative. More broadly, the IMF has intensified collaboration with other
international agencies that have social policy expertise. IMF staff has participated in international social
policy forums relevant to the IMF's economic policy advice. For example, the staff contributed to the
discussion at the 1995 World Summit for Social Development in Copenhagen and recently to
UN/OECD/World Bank discussions on a core set of international development goals and indicators (Box
2.2). The IMF also has organized conferences on Income Distribution and Sustainable Growth (1995) and
on Economic Policy and Equity (1998) (see Tanzi and Chu, 1998; and Tanzi, Chu, and Gupta, 1999). In
addition, IMF management and staff have engaged the representatives of civil society groups, including
labor unions, NGOs, and religious groups, in a dialogue on social concems and IMF policy advice. For
example, meetings with such groups are now commonplace during staff missions or at headquarters.11

10
11

See, for example, IMF (1995) and Gupta and others (1998).
Among t e frequent contacts in civil society groups are Oxfam, Friends of t e Earth, World Vision, the Swiss Council of
h
h
Development Organizations, Witness for Peace, Christian Aid, Results International, and Caritas Intemationalis. The IMF has
organized numerous seminars for academics, labor unions, environmental groups, religious organizations and development
NGOs. In 1996, the Managing Director addressed a World Congress of the International Confederation of Free Trade Unions
(ICFTU) and in 1 9 the World Conference of Labor. In addition, t e IMF greatly increased t e dissemination of information
97
h
h
through press releases, public information notes, a d t e publication of s m staff reports and other studies.
n h
o e

Sesión 4 - 344

Box 2.1
Evolution of the IMF's Social Policy Advice
Recent Evolution
Specific operational guidance has been provided to the staff of the IMF through
Executive Board discussions and guidelines.
•
In the mid-1980s, Board discussions were held on poverty, fiscal policy, and income
distribution in IMF-supported programs. 12
•
In 1993, the Executive Board considered issues concerning the design of social
safety nets and their integration in adjustment programs, and in the mid-1990s the
composition of public expenditures.
•
In September 1996, the Interim Committee (now the International Monetary and
Financial Committee) stressed the need for an enhanced
approach
to social
policies in a declaration entitled Partnership for Sustainable Growth. It states that
"because the sustainability of economic growth depends on development of human
resources, it is essential to improve education and training; to reform public pensión
and health systems to ensure their long-term viability and enable the provisión of
effective health care; and to alleviate poverty and provide well-targeted and
affordable social safety nets." 13
•
In June 1997, guidelines for improving the monitoring of social expenditures and
social indicators were issued to IMF staff. The social indicators to be monitored
included the core set of international development goals and indicators laid out in the
March 1995 "Copenhagen Declaration on Social Development and Program of
Action" of the World Summit for Social Development. Also, several of the
recommendations of a recent external evaluation relating to the social aspects of the
ESAF are being incorporated in ESAF-supported programs,14 including through a
pilot program for enhanced World Bank-IMF collaboration launched in 1998. Under
this pilot program, World Bank and IMF staff work with six countries to make deeper
assessments of the social impact of adjustment policies and to address these effects
in the design of the countries' IMF-supported programs.
•
More recently, especially in the aftermath of the Asian financial crisis, the Managing
Director, Michel Camdessus, has stressed the need for a social pillar in the

architecture

of the international financia)

system.

Current Developments
The World Bank was requested by the Development Committee in October 1998 to
develop general principies of good practice in social policies, and a paper was discussed at
the Spring 1999 Meeting of the Development Committee. The President of the World Bank,
James Wolfensohn, has also proposed a Comprehensive Development Framework. A social
pillar would need to be founded on a strong statement of social objectives by countries in
their policy frameworks, supported when needed by external financial and technical
assistance.

12
13
14

See, for example, IMF (1996) and various Development Committee documents.
See IMF (1996).
See IMF (1997) a d Abed a d others (1998) for the intemal review of ESAF.
n
n

Sesión 4 - 34

sector

Box 2.2
International Social Development Goals and Performance Indicators
Goals of Social Development
Since the early 1990s, various global UN conferences have established goals for social policies, as well
as for the environment, human settlements, human rights, drug control, and crime prevention. In particular, the
Copenhagen Declaration on Social Development (March 1995) laid out a program of action which, among other
things, included the goals of eradicating poverty, promoting social integration, and achieving universal and
equitable access to education and primary health care.
Key goals of social development are, by the year 2015 to:
•
reduce the proportion of people living in poverty by at least one-half relative to 1993;
•
achieve universal primary education in all countries;
•
make progress toward gender equality by eliminating gender disparity in primary and secondary
education (to be achieved by 2005);
•
reduce maternal mortality rates by three-fourths and reduce infant and child mortality rates by two-thirds
relative to 1990; and
•
provide access to reproductive health services to all individuáis of appropriate ages.
Indicators for Measuring Progress
Several sets of social indicators have been identified in various forums to assess social development
and monitor key social development goals. Examples of such sets of social indicators are:
•
the OECD/UN/World Bank core set of working indicators of international development goals; 15
•
the Common Country Assessment (CCA) indicators of the UN Development Assistance Framework
(UNDAF);
•
the UN/CCA Task Forcé on Basic Social Services for All (BSSA) indicators; and
•
the UN Statistical Commission's Mínimum National Social Data Set (MNSDS).16
The IMF has been supportive of these efforts.
The OECD, World Bank, and UN, in cooperation with developing countries and bilateral donors, have
established the following working set of core indicators on social development:
•
Poverty: share of the population living below $1 a day in purchasing power parity terms; the poverty
gap (the resources needed to lift all those below the poverty line out of poverty); prevalence of
underweight children under 5 years of age; and the share of the poorest 20 percent in national
consumption.
•
Education: net enrollment rates in primary education; completion rate of fourth grade of primary
education; and literacy rate of those between 15 and 24 years of age.
•
Gender equality: ratio of girls to boys in primary and secondary education; ratio of literate females to
males (ages 15to 24).
•
Health: infant mortality rate; under-5 mortality rate; maternal mortality rate; percentage of births
attended by skilled personnel; contraceptive prevalence rate; and HIV prevalence in pregnant women
aged 15 to 24 (for lack of data, currently the overall HIV prevalence rate is used).
In addition, the OECD/UN/World Bank core list includes 6 environment indicators, as well as ten
background indicators of development, such as adult literacy rate, total fertility rate, and life expectancy.
The CCA/UNDAF list includes all indicators in the OECD/UN/Worid Bank core set, but for some
development goals, the list has a more extensive scope, including, for example, more indicators on gender
equality and women's empowerment, child welfare, and food security. In addition, the CCA/UNDAF list has
indicators relating to employment, housing, drug control, and crime prevention.
Compared with the CCA/BSSA and MNSDS sets, the OECD/UN/World Bank list is more extensive and
includes a wider range of social development indicators. However, the CCA/BSSA and MNSDS sets also include
some indicators not found in the OECD/UN/World Bank list, such as average years of schooling (MNSDS) and
access to primary health care services (BSSA).
All these sets are intended to be used flexibly, and need to be adapted to the specific circumstances of the
country to which they are applied. Indicators may be added if they capture an aspect of social development not
included in the sets, while lack of data may require that some indicators omitted.
15
16

Available via the Internet: http://www.oecd.org/dac/indicators/htm/tables.htm.
These indicators and their definitions are available in t e World Bank's World Development Indicators database and in t e
h
h

UN Development Program's (UNDP)

Human

http://www.undp.0rg/hdr0/indicat0rs.html#devel0ping.

Development Reports,

Sesión 4 - 346

which are available via the Internet:

III. Social Safety Nets
In countries where the authorities could foresee that reform measures would have a sizable
adverse social impact, the policy mix and sequencing have aimed to take this impact into account within a
sustainable macroeconomic framework. For instance, IMF-supported programs have aimed to phase out
subsidies for food and other items gradually, rather than at once (e.g., Indonesia, 1998; and Senegal,
1994—95). The adverse impact, however, cannot be totally eliminated even with an appropriate policy mix
and sequencing. For instance, a change in relative prices that hurts the poor—such as a devaluation, that
could adversely impact the urban poor through increasing prices of imported products—may be at the
heart of a reform program. A tensión may emerge, therefore, between stabilization and social protection
objectives.
Social safety nets are a means of easing this tensión. IMF staff—drawing on the work of other
institutions—has increasingly sought to incorpórate social safety nets into adjustment programs. The
IMF's Executive Board, following its discussion of social safety nets in 1993, endorsed this approach.17
More recently, ESAF-supported programs have sought larger budget allocations for social safety nets.
There has been an increase in the use of structural benchmarks and performance criteria aimed at securing
social protection objectives (Box 3.1).
This review indicates that most IMF-supported programs have incorporated social safety nets,
although there is scope for further improvement in their quality and implementation.
•

•
•

The social safety nets in IMF-supported programs have included new temporary arrangements—
such as temporary subsidies and public works programs as well as existing social protection
instruments adapted to the needs of target groups—such as pensions and other permanent social
security programs (see Chu and Gupta, 1998). Measures designed to foster financial stability, such
as the adoption of deposit guarantees (limited or general) and otherfinancial restructuring
measures to maximize asset recoveries, to redistribute losses, and to sustain credit to the small and
medium-industry segments, also have protected small depositors and vulnerable groups.
Over time, permanent social protection arrangements (e.g., pensions, unemployment insurance)
also have been established in the context of reform programs.
Family-based safety nets have cushioned income losses during adjustment periods in many
countries. For example, in Indonesia, in the wake of the recent crisis, about a fourth of families
received informal transfers (Frankenberg, Thomas, and Beegle, 1999). Such informal
arrangements have been generally well targeted.18 Thus the design of public social safety nets has
sought not to duplícate the system of voluntary, prívate transfers.

The timely implementation of social safety nets has been hampered—all too frequently—by a lack
of existing social policy instruments. Often these instruments can be speedily adapted to the needs of the
new target groups. But sometimes the adaptation is difficult. Countries often have not had the will to
reform costly existing social protection mechanisms or to shift social protection priorities. Those segments
of the population suffering from the adverse effects of reforms may prove to be different from those
protected by the permanent arrangements. In addition, the lack of data and administrative and financial
constraints have hampered implementation and monitoring.
17

18

T e two conferences on income distribution and economic policy organized by t e IMF in recent years have also emphasized
h
h
t e need for cost-effective social safety nets during reform periods. See Tanzi and Chu (1998) a d Tanzi, Chu, a d Gupta
h
n
n
(1999).
See Cox, Okrasa, and Jm n z (1997) and Cox, Eser, and Jm n z (1997) for such networks in Poland and t e Russian
i ee
i ee
h
Federation, respectively.

Sesión 4 - 34

Box 3.1
Strengthening Social Safety Nets in ESAF-Supported Programs, 1994-98
Policy Framework Papers (PFPs) and Memoranda of Economic Policies (MEPs) for
44 countries that had ESAF-supported programs during 1994-98 were reviewed to ascertain
what kind of measures had been incorporated to mitigate the adverse social effects of
structural adjustment policies. These measures were classified into two groups:
•
Budgetary allocations (either unspecified general allocations or quantitative
commitments); and
•
Measures to strengthen social protection through more specific targeting, better
monitoring of the affected population groups, widened coverage of safety net
measures, or related reforms.
A significant group of countries have incorporated allocations for social safety nets in
their ESAF-supported programs, and the use of structural benchmarks and performance
criteria for achieving social protection goals has grown.
•
Half the countries had commitments in their PFPs for allocations for financing social
safety nets during the program period (on average 3-32 years), with a third of the
44 countries setting these targets in quantitative terms. During the review period,
countries also included measures for strengthening the design and coverage of
social safety nets; some established targets for the number of vulnerable people to
be shielded by social safety nets (Armenia, Guyana, the Kyrgyz Republic,
Mozambique, and Nicaragua).
•
About three-fourths of countries announced spending on social safety nets in their
MEPs during the program period, with 60 percent of the 44 countries specifying
quantitative allocations. In recent years, however, the use of quantitative targets with
respect to social safety nets has declined. A similar trend is noticeable in the listing
of structural measures for strengthening social safety nets. Although it is difficult to
pinpoint the reasons for this trend, there are several possibilities. Priorities in
programs could have shifted; the monitoring of social safety nets could have become
more difficult because of data and other constraints; and in a number of countries,
structural reforms with the greatest adverse social impact could have been felt earlier
in the program periods. On a country-by-country basis, significantly fewer countries
(about one-half) that sought allocations of expenditures on social safety nets in PFPs
made the same commitments in their MEPs. Nine countries specified targets for the
number of vulnerable to be shielded by social safety nets in MEPs (Albania,
Armenia, the Republic of Congo, Haiti, the Kyrgyz Republic, the former Yugoslav
Republic of Macedonia, Mozambique, Nicaragua, and Yemen).
Greater use of structural benchmarks and performance criteria for social safety nets
in ESAF countries is a recent phenomenon. Structural benchmarks have included, for
example, improving the transparency of energy subsidies, replacing generalized subsidies
with targeted subsidies, and establishing labor retrenchment funds. Programs for six
countries incorporated benchmarks (Azerbaijan in 1996, Cameroon in 1998, the Kyrgyz
Republic in 1994,1995, and 1998, the former Yugoslav Republic of Macedonia in 1997 and
1998, Pakistan in 1995-98, and Yemen in 1997-98). Of these, two countries have included
performance criteria for achieving social protection objectives, for example, by strengthening
the revenue position and by reforming the benefit structure of the employment and the
pensión funds (the Kyrgyz Republic in 1994,1997, and 1998, and the former Yugoslav
Republic of Macedonia in 1998).

Sesión 4 - 348

In most cases, IMF staff has relied largely on the World Bank, and regional development banks to some
extent, to take the lead in the design of social safety nets for IMF-supported programs.
•

•

For example, in Indonesia, a targeted rice subsidy and community-based public works programs,
designed by the World Bank, were incorporated in the 1998 IMF-supported adjustment program.
This was also the case for the public works programs financed by the Asian Development Bank
and the World Bank in Thailand in 1998. In Brazil in 1999, the Inter-American Development
Bank and the World Bank cofinanced a special adjustment loan for a social protection project that
was integrated into the IMF-supported program.
In some cases, the IMF Fiscal Affairs Department has provided limited technical assistance on
social safety nets (e.g., Ecuador and Belarus in 1999). And in 1999, the African Department
(AFR) recruited two social policy specialists with backgrounds in sociology to assist the
department in incorporating appropriate social safety nets in countries' adjustment programs. To
the extent possible, IMF missions have built on the work of the World Bank, regional
development banks, other UN agencies, and NGOs.19 In many cases, program-related missions
have also focused on social issues.

A.

Design Issues

The design of social safety nets has been influenced, among other things, by the availability of
existing social policy instruments. For example, transition economies had a broad range of social
instruments that were poorly targeted (e.g., Moldova and Ukraine). A wide range of benefits covered the
bulk of the population, including the nonpoor. Thus the principal aim of IMF-supported programs in these
countries has been to make spending—for instance, food subsidies—better targeted, rather than to create
new instruments. In contrast, low-income developing countries had limited social policy instruments, and
the effort there has been to create arrangements that could reach affected population groups (such as
transitory subsidies for the urban poor in the CFA franc zone countries in the aftermath of the 1994
devaluation). In general, establishing cost-effective social safety nets would have been easier had welltargeted social policy instruments been already in place before the onset of crises and economic reforms.20
The specific adverse effects on and characteristics of target groups have determined the types of
social safety net instruments.
•
A sharp fall in output, reinforced by a large increase in prices of important staples, can result in a
significant real income loss for those poor households who are net consumers of food. This
occurred in Indonesia in 1998 and in transition economies in the initial stages of transformation.
In these circumstances, income transfers or targeted food subsidies became critical.
•
When the prices of essential goods rose in countries where the elderly constituted a high
proportion of the population, low-income pensioners needed to be helped through an adequate
minimum allowance (e.g., the Russian Federation and Ukraine).
•
When there were regional pockets of unemployment, special programs to supplement incomes
have been implemented (e.g., community-based public works programs in Indonesia and
Senegal).

19

20

This approach is consistent with that set forth in the 1 9 Development Committee paper prepared jointly by t e staffs of t e
93
h
h
IMF a d t e World Bank. Executive Directors in t e 1 9 discussion o safety nets "agreed that t e IMF's policy advice
n h
h 93
n
h
through technical assistance o social safety n t should be continued to t e extent that staff resources were available."
n
es
h
Ferreira, Prennushi, a d Ravallion (1999).
n

Sesión 4 - 34

Nevertheless, the selection of target groups has raised fairness issues. To ensure political support
for reform, social safety nets have been extended to politically vocal middle-income groups: for example,
a subsidy for premium gasoline (in Indonesia, initially in 1998), severance payments for departing civil
servants (Ghana and Lao People's Democratic Republic) and for public enterprise employees (Argentina
and Bolivia), and food subsidies (Jordán). Unemployment benefits—largely for formal sector
beneficiaries—were strengthened in the presence of a large informal sector, where the majority of the poor
may be residing (Brazil in 1998).
To identify target groups, IMF staff has generally relied on the national authorities and the World
Bank which have provided a measure of the poverty line and household expenditure survey data on
household characteristics.21 The latter, however, are often conducted infrequently with results that are not
always timely or comparable across time.22 Because of these constraints, poverty lines and poverty
profiles have typically been unavailable for the year when reform programs were put in place (e.g., Brazil
and Thailand).
The weak administrative capacity in many countries has hampered the targeting of benefits,
particularly on the basis of incomes. This has meant a greater reliance on programs that have self-targeting
features, such as
•
•
•

public works with below-market wages (e.g., Indonesia, Malawi, Thailand, and Senegal);
subsidies on commodities consumed by the poor (e.g., inferior ricein Indonesia); and
shielding of groups that are easily identified as poor (pensioners, the unemployed, single mothers,
and children).
Means testing based on wage income has been used in some countries—for the housing subsidy
program in Ukraine, for example—but this carnes the risk of mistargeting, especially in countries with a
large informal sector, and could create disincentives for the supply of labor.
Financial constraints have limited the scope of social safety nets. The need to redress
macroeconomic imbalances has typically precluded increasing total public spending; thus, a reallocation
within the existing budgetary envelope to better-targeted programs has been necessary (e.g., Brazil).
•
•
•
•

In some cases such as Venezuela , eliminating subsidies has freed resources for more targeted
social protection programs.
Significant budgetary savings have often been achieved by reforming social safety nets, such as
by replacing generalized subsidies with targeted ones (e.g., rice subsidies in Indonesia).
External donors, including the World Bank, have played a role in some cases, in particular by
funding severance payments for departing civil servants and by providing food aid (e.g., Senegal).
In transition countries, the large decline in output has increased the demand for social benefits
while reducing the availability of financing (Box 3.2).

When the adverse impact exceeded expectations, program targets have been changed to
accommodate larger budgetary outlays for social safety nets. Indonesia, Korea, and Thailand raised
spending on social protection programs to 5.2 percent of GDP in 1998/99, 2 percent of GDP in 1999, and
21

22

The Poverty Assessments prepared by the World Bank have been particularly useful. In s m cases, the World Bank and
o e
national authorities use different poverty lines (e.g., Belarus). In t e design of social safety nets, the usual practice has been to
h
use country-specific poverty lines, rather t a international poverty lines defined in t r s of U.S. dollars in purchasing power
hn
em
parity t r s for daily consumption of a individual.
em
n
Less t a two-thirds of countries with IMF-supported programs have conducted at least one household or demographic
hn
survey, of which t e last one in more t a half of t e countries was conducted before 1996.
h
hn
h

Sesión 4 - 350

2 percent of GDP in 1998/99, respectively, from between Vi percent and 1 percent of GDP in each country
before the crisis.
In general, and beyond the context of immediate program requirements, it would be desirable to
identify the need for social policy instruments, and advise government authorities to seek necessary
assistance from the World Bank and others, in the course of IMF surveillance. Such efforts could, among
other things, speed the establishment of cost-effective safety nets if difficulties arise and reform measures
need to be undertaken.
Box 3.2
Social Safety Nets: Issues in Transition Economies
Key Issues
Transition economies have been trying to reform expansive, but increasingly
ineffective, social protection arrangements, including subsidies, pensions, unemployment
benefits, and family allowances. Declining taxes and social contributions have severely
weakened the ability of many transition economies to provide the needed benefits.
•
The declining social contributions have transformed the earnings-related pensions
and other social benefits into virtually fíat minimum benefits.
•
Offsetting tax obligations of enterprises against obligations of the government has
limited the ability of many governments to pay cash benefits. Moreover, these
obligations of the government do not necessarily represent spending of high social
valué.
In many transition economies, the benefits have yet to be fully reformed.
•
In Moldova, the Russian Federation, and Ukraine, pensión systems have allowed
workers to collect benefits at a relatively young age, and workers in certain
occupations have been eligible for pensions even earlier. Pensions have been based
not only on the number of years of contributions, but also on years spent studying or
taking care of a young child.
•
A large part of social contributions collected for assisting the unemployed continué to
be wasted on low-priority programs and benefits administration; only a small share of
the unemployed actually receive assistance in any form.
•
Traditional extensive and generous privileges for politically influential groups (judges,
parliamentarians, internal security personnel) prevent the targeting of limited
resources to the genuinely needy.
IMF Advice
Thus, the emphasis of IMF staff advice has been on improving compliance with tax
laws, simplifying the rate structure, and stopping the collection of taxes and social
contributions in kind (e.g., Azerbaijan, Moldova, and Ukraine). Some progress in
simplificating the social contribution rate structure has been made, but significant change in
other areas has been elusive.
In these and other similar cases, IMF staff have called for raising the pensión age,
eliminating privileged and early pensions and untargeted benefits, targeting social assistance
and subsidies, increasing the size and coverage of unemployment benefits, and making
social benefits more transparent. Pensión reforms are under consideration in many transition
countries, but the progress in implementing far-reaching reforms remains slow. For example,
despite the emphasis given to pensión reform, including the use of performance criteria, in
IMF-supported programs in the Kyrgyz Republic, relatively little progress was achieved; a
conceded and more comprehensive and resource-intensive World Bank adjustment
operation appears to have had more success in pensión reform since 1998.

Sesión 4 - 3

B.

Labor Market Implications

Laboi market incentives have been a key concern in the design of unemployment benefits. The
challenge has been to strike an appropriate balance between social protection and disincentive effects.
Though certainly, this balance may differ among countries, depending on social preferences, norms, and
other factors. In 1998, because of their concerns about possible disincentive effects, the Korean authorities
initially hesitated to broaden the coverage of unemployment benefits. A relatively high wage for
participants of public works programs can undermine their effectiveness as a safety net by attracting
already employed workers.23
Establishing social safety nets, however, can help promote a fundamental labor market reform.
Before thefinancial crisis in 1997, lifelong employment in large enterprises had been an important aspect
of social protection in Korea. But it constrained the ability of enterprises to restructure in the face of
changing economic conditions. Broadening the coverage of unemployment benefits to 70 percent of the
labor forcé in early 1999 from around 30 percent supported labor market reform aimed at promoting labor
market flexibility by providing income transfers to those switching jobs.

C.

Monitoring

The staff monitoring of social safety nets has been infrequent. A review of 12 countries with
ESAF arrangements during 1994—98 indicates that, although over three-fourths of Policy Framework
Papers (PFP) and Memorándum of Economic Policies (MEP) reported on the performance of social safety
nets under IMF-supported programs, such monitoring typically occurred only once or twice.24 The
infrequent monitoring of social safety nets may have reflected weak national capacity to monitor the
implementation of social policies and their impact on poverty.
The staff s assessment and reporting of the effectiveness of social safety nets have been uneven.
In only a third of the 12 countries reviewed did staff papers assess coverage and incidence of social safety
nets during thefive-year period (e.g., with respect to temporary food subsidies and to fertilizer provisión
to smallholders in Malawi). For a larger number of countries (more than three-fourths), staff papers
reported on the improvements in the benefxt structure andfinancing of social protection mechanisms. In
contrast, staff papers on IMF-supported programs for the countries affected by the recent crisis in Asia
(Indonesia, Korea, and Thailand) reported extensively on social safety net developments.25

IV. Public Spending on Education and Health Care
The relationship between public social spending, social indicators, and poverty reduction is
complex and dynamic. How much public social spending reduces poverty depends not only on the amount
allocated for education and health care, but also how efficiently these allocations are spent and how well

23

24

The wages paid to participants are the mo critical determinant of overall program cost and t e effectiveness of job creation
st
h
through public works schemes. Experiencefrom a range of countries shows that programs are m r effective when t e wage
oe
h
is maintained at a level below t e prevailing m r e wage for unskilled labor. See Subbarao and others (1997).
h
ak t
Azerbaijan, Bolivia, Georgia, Guyana, the Kyrgyz Republic, t e Lao People's Democratic Republic, the former Yugoslav
h
Republic of Macedonia, Malawi, Mongolia, Pakistan, Senegal, a d Vietnam.
n
With the support of the World Bank, a few countries have recently established a mechanism to monitor social o t o e o an
uc m s n
on-going basis (e.g., the Social Monitoring a d Early Response Unit in Indonesia).
n

Sesión 4 - 35

they are targeted to the poor.26 Education and health care indicators are affected not only by govemment
outlays on education and health care but also by prívate spending, demographic trends, and public
spending in other areas such as for sanitation and safe water. Empirical research on the link between
increased aggregate public spending on education and health care and improvements in related social
indicators has yielded conflicting evidence.27 Note, also, that today's illiteracy and infant mortality rates
are normally the result of yesterday's social policies; poverty reduction reflects past increases in spending
on primary education, primary school enrollment, and literacy. Finally, some indicators reflect
intermedíate outputs, not final outcomes. For example, widespread immunization of infants under
12 months against measles does not by itself yield a low infant mortality rate, especially if other variables,
such as access to safe water and female education attainment, are relatively poorly developed.28
This review suggests that considerable progress has been achieved in strengthening spending
policies on public education and health care, but that some areas require further efforts.
•

Countries have made considerable progress in establishing comprehensive and structured policy
frameworks for such spending; more recently, there have been discemible improvements in
targeting and monitoring public spending and positive, albeit modest, developments in related
social indicators and outcomes.
•
Efforts to raise spending on education and health care have achieved relatively more success in
the HIPC decisión point countries (Bolivia, Burkina Faso, Cote d'Ivoire, Mozambique, and
Uganda) than in other IMF-supported program countries.
Further improvements, however, are needed to address some inadequacies:
•
Lack of adequate data is commonplace. Data on the composition of education and health care
spending are often not available. Data on subnational govemment spending are scarce. Education
and health achievement indicators are either unavailable or available with a long lag (Box 4.1).
•
Policy objectives have not always been clear or articulated in terms of well-defined targets against
which progress can be measured, and the definition of targets and monitoring often has changed
over time within a single country.
The HIPC Initiative framework has yielded relatively more progress than have other programs,
across a more comprehensive range of social sector reforms. The most marked improvements in social
indicators during the period under review also have taken place in HIPC countries. No causal association,
however, can be established between increased spending and outcomes, because their link is affected by
many other factors.

26

27
28

There are wide disparities in t e cost-effectiveness of govemment spending on education and health care across countries in
h
Africa, and, in general, these countries were found to be less efficient t a those in Western Hemisphere and Asia. See Gupta,
hn
Verhoeven, a d Honjo (1997).
n
See Mingat and Tan (1998); Filmer, H m e , a d Pritchett (1998); and Gupta, Verhoeven, and Tiongson (1999).
a mr n
For example, for the latest year for which data are available, Zambia has a higher immunization coverage than Kenya, but also
a higher infant mortality rate. In contrast, for a group of 48 program countries for which recent data are available, there is a
negative and statistically significant correlation between immunization coverage and infant mortality rate.

Sesión 4 - 353

Box 4.1
Quality of Social Spending and Indicators Data
Social Spending
Many deficiencies exist in data on public spending on education and health care.
•
In general, spending by local governments is not included; this can be a major handicap
in countries that have devolved or are devolving expenditure responsibilities to lower
levels of the government, particularly those related to basic education.
•
In many cases, data coverage in fiscal accounts is limited to current outlays, in part
owing to the inability of governments to separate donor-financed capital spending by
function.
•
Jn-kind donor contributions to education and health care are not included.
•
Data typically become available with a lag, which for some countries can be as long as
two to three years.
•
Virtually no country has consistent annual series for expenditure allocations within the
education and health care sectors (e.g., separating between primary and tertiary
education, or preventive health care and curative health care), and the data available in
many cases are not consistent with aggregate fiscal data.
•
Despite the importance of books and medicine for developments in social indicators,
separate data for nonwage and wage outlays in education and health care sectors are
available for only very few countries.
•
Data on prívate sector outlays on education and health care are not collected on a
regular basis.
Social Indicators
The most serious shortcoming of data on social indicators is that they are generally
produced infrequently and with a long lag, or, in many cases, are not collected at ail. For
instance, data for 1997 are available for only 11 out of 18 indicators of well-being and social
development in the working set identified by the OECD/UN/World Bank, and only for a small
number of developing and transition countries. Current data for many important indicators are
derived from models, rather than from actual observations. For example, for 102 countries,
actual observations on infant mortality rates are not available for 1985 or later.
Furthermore, some of the key indicators become available only every five years. In
some cases, there is a trade-off between the availability of data on social indicators and their
quality. For example, net enrollment rates, which correct for grade repetition, are available for
only about half of program countries, whereas gross enrollment rates, which are available for
most countries, count all students regardless of age as part of the school-going population, and
thus overstating enrollment to the extent students are repeating grades (see Table 4.1). There
may also be inconsistencies among data sources and compilation methods, raising questions
about data comparability across countries and over time. Because indicators are constructed by
using data collected at the national level through censuses, sample surveys, and administrative
records, data quality to a large extent depends on the national statistical capacity.

Sesión 4 - 354

Table 4.1: Improvement ¡ Social Indicators i M m e Countries, 1985-97
n
n e br
(Current level and average annual percent improvement; n m e of countries i parentheses)
u br
n
Current Level 1 /
ESAF
Non
Program
Program
program
Countries Countries
Countries2
Only
Education
Illiteracy rate 3/
15.8 (22) 27.4 (49)
Female
18.6 (22) 32.8 (49)
Gross primary school enrollment rate 4/
104.2 (25) 92.3 (55)
Female
102.8 (25) 88.6 (52)
Male
105.8 (25) 96.5 (52)
Gross secondary school enrollment rate 4/ 61.8 (25) 50.5 (53)
Female
65.0 (24) 49.5 (49)
Male
62.6 (24) 51.9 (49)
Net primary school enrollment rate
84.8 (18) 74.6 (31)
Persistence to grade 5
88.5 (21) 75.3 (23)
Health care
Life expectancy (in years)
Infant mortality rate 31
Under-5 mortality rate 3/
Births attended by skilled staff
Contraceptive prevalence 5/
Access to health care
Percent under 12 months immunized
DPT vaccination
Measles vaccination
Dther basic services
Access to safe water
Access to sanitation

68.1
31.8
40.1
78.5
66.3
94.6

(33)
(34)
(29)
(25)
(4)
(16)

63.3
48.9
66.6
64.0
49.3
74.0

(45)
(48)
(29)
(37)
(14)
(13)

Average Annual Percent Improvement

Non program Program
Countries 21 Countries

ESAF
Program
Countries
Only

HIPC
Countries
Only

SubSaharan
Africa

45.5
53.4
80.9
76.6
88.0
30.6
28.4
32.5
58.9
68.0

(21)
(21)
(25)
(24)
(24)
(24)
(23)
(23)
(17)
(11)

3.4
3.6
0.1
0.3
0.1
2.8
3.2
2.7
0.2
0.5

(22)
(22)
(25)
(25)
(25)
(25)
(24)
(24)
(18)
(21)

3.5
3.5
0.8
0.8
0.7
1.1
1.9
0.6
0.7
1.7

(49)
(49)
(55)
(52)
(52)
(53)
(49)
(49)
(31)
(23)

2.2
2.1
1.0
1.1
0.9
0.8
2.3
0.2
0.6
2.0

(21)
(21)
(25)
(24)
(24)
(24)
(23)
(23)
(17)
(11)

2.2
2.0
0.9
1.2
1.4
0.9
2.8
0.5
0.5
2.9

(16)
(16)
(15)
(14)
(141
(14)
(13)
(13)
(11)
(8)

2.4
2.3
0.6
1.1
1.0
1.2
3.1
0.3
0.1
2.3

(15)
(15)
(15)
(14)
(14)
(14)
(14)
(14)
(10)
(9)

57.6
72.8
113.2
52.4
36.0
48.7

(25)
I25)
(14)
(18)
(5)
(3)

0.4
1.7
3.8
6.1
3.6
1.4

(33)
(34)
(29)
(25)
(4)
(16)

0.2
2.5
3.7
1.8
3.1
3.7

(45) 0.2
(48)
1.5
(29) 2.9
(37) 3.2
(14)
5.6
(13) 11.2

(25)
(25)
(14)
(18)
(5)
(3)

0.1
1.6
2.3
0.8
6.7
4.0

(12)
(13)
(8)
(10)
(3)
(4)

-0.1
1.0
0.8
1.7
6.9
6.6

Asia
and the Pacific

Transition
Economies

(8)
(8)
(10)
(9)
(9)
(10)
(10)
(10)
(4)
(5)

2.9
3.0
0.7
0.5
0.6
1.4
1.2
1.4
0.9
2.5

(14)
(14)
(14)
(13)
(13)
(14)
(10)
(10)
(11)
(7)

7.9
8.3
1.0
0.1
0.8
-0.8
-0.5
-0.9
0.4

(7)
(7)
(13)
(13)
(13)
(12)
(12)
(12)
(4)
(0)

(13)
(14)
(8)
(9)
(3)
(5)

0.7 (4)
2.7 (5)
(0)
4.2 (10)
2.0 (5)
6.2 (2)

0.5
3.3
3.2
-0.1
2.5
0.4

(9)
(10)
(7)
(11)
(5)
(4)

0.0
3.0
5.7
-0.3
0.0

(16)
(16)
(13)
(4)
(1)
(0)

7.7 (10)
9.1 (10)

86.8 (33) 82.6 (61)
85.0 (33) 83.0 (60)

74.6 (29)
75.3 (28)

4.5 (33)
5.4 (33)

4.8 (61)
4.4 (60)

6.9 (29)
6.1 (28)

7.3 (16)
6.8 (15)

5.7 (15)
4.3 (14)

81.7 (27) 66.0 (41)
73.9 (23) 57.6 (37)

58.1 (19)
41.6 (17)

4.1 (27)
2.6 (23)

2.9 (41)
4.4 (37)

4.2 (19)
6.7 (17)

4.0 (12)
3.2 (10)

4.2 (12)
2.7 (10)

3.1
3.1
0.5
1.4
0.0
2.2
3.5
1.6
0.9
-0.3

Latin America
and the
Caribbean

7.7 (8)
14.2 (8)

3.8 (15)
3.3 (15)

3.7 (16)
2.8 (16)

1.6 (14)
3.5 (14)

-3.5 (5)
-9.9 (3)

Sources: World Bank, World Development Indicators 1998 and 1999 database.
1/ Latest data available. Mostly refers to 1995-97. Except for life expectancy, which is in years, all the indicators refer to shares of the relevant population groups. See also Box 2.
21 Nonprogram countries, shown for comparison, included: Angola, The Bahamas, Bahrain, Belize, Bhutan, Botswana, Colombia, Cyprus, Eritrea, Fiji, Grenada, I.R. of Irán, Kuwait, Lebanon,
Malaysia, Maldives, Malta, Myanmar, Namibia, Netherlands Antilles, Omán, Paraguay, Qatar, Seychelles, Solomon Islands, South Africa, St. Kitts and Nevis, St. Lucia, St. Vincent and the
Grenadines, Swaziland, Syrian Arab Republic, Tonga, Turkmenistán, and United Arab Emirates.
3/ Tire annual percent improvement in illiteracy, infant mortality, and under-5 mortality rates refers to a decline in these rates. An annual percent improvement of 3.4 in illiteracy, for example,
means that illiteracy rates are falling by 3.4 percent per year.
4/ Gross enrollment rate is the ratio of total enrollment, regardless of age, to the population of the age group that corresponds to the level of education shown. If, for example due to grade
repetition, students that fall outside the age bracket for primary or secondary education are nevertheless enrolled, the gross enrollment rate may exceed 100 percent.
5/ Contraceptive prevalence rate is the percentage of women who are practicing, or whose sexual partners are practicing, any form of contraception and is usually measured for women aged
15-49.

A.

Aggregate Spending on Education and Health Care

IMF-supported programs have sought to promote universal access to basic social services.
Programs have increased public spending for such services in countries where this spending was low,
supported high-quality expenditure in these sectors, and protected or sought real increases in these
expenditures during adjustment periods when poor households might lack the ability to pay for basic
social services. The importance programs have attached to these objectives was reflected in therisinguse
of quantitative targets, structural benchmarks, and performance criteria aimed at raising education and
health care spending (Box 4.2).29
Box 4.2
Targets for Public Spending on Education and
Health Care in ESAF-Supported Programs, 1994-98
Policy Framework Papers (PFPs) as well as Memoranda of Economic Policies (MEPs)
for 44 countries that had ESAF-supported programs during 1994-98 were reviewed to ascertain
the extent to which they:
•
incorporated targets for budget allocations for education and health care, either in
unspecified general or specific quantitative terms;
•
called for structural improvements in the provisión of social services; and
•
monitored changes in, and established targets for, social indicators.
ESAF-supported programs have increasingly sought to raise public spending on
education and health care and to implement structural reforms in the sectors. Benchmarks and
performance criteria have also been increasingly widely used to achieve increases in such
spending.
In PFPs, about 80 percent of the 44 countries sought increases in public spending
on education and health care during 1994-98, and a slightly lower proportion (60 percent)
set quantitative targets for such increases. Targets were most commonly set once
during the period while, on average, ESAF-supported programs were in place in the
countries for 3-3Vz years during 1994-98. PFPs for around 60 percent of the 44 countries aimed
at increased budgetary allocations for primary education and basic health care during the period,
and about one-third of these set quantitative increases. All programs called for structural
measures to strengthen the provisión of social services during the period, for example, by
increasing the number of teachers and doctors and enhancing the role of the prívate sector.
About 45 percent of the 44 countries targeted improvements in social indicators in both
unspecified and quantitative terms. The most commonly used indicators were primary school
enrollment, including, separately for girls, literacy, infant mortality, and immunization rates.
The picture is broadly similar for MEPs with respect to the proportion of countries that
committed to increase budgetary expenditures on health care and education. Compared with
PFPs, however, a much smaller percentage of countries (about 45 percent) sought increases in
budgetary allocations for primary education and basic health care (either unspecified or in
specific quantitative terms) and some 16 percent established specific targets for quantitative
increases. A lower percentage of countries (30 percent) identified improvements in education
and health care indicators as a policy objective. The use of quantitative targets increased by
40 percent during the period under review.
29

Notwithstanding t e problems with comparability, IMF staff has compiled cross-country data on public education a d health
h
n
care spending. This data set covers 6 countries t a are implementing or have implemented IMF-supported r f r programs,
5
ht
eo m
of which 3 are low-income countries with ESAF-supported programs. The GDP deflator was used to convert nominal
1
expenditures into real terms. In principie, deflating by public sector wages would provide a m r accurate reading of real
oe
trends in education and health care spending. But such wage data are rarely available for low-income countries. In 1
0
countries in Africa for which data on public sector wages are available, real per capita spending on education a d health care
n
increased, on average, by 2 percent a year under IMF-supported programs—a result consistent with spending t e d derived
rn s
from the GDP deflator.

Sesión 4 - 35

On a country-by-country basis, commitments to social spending measures was less
frequent in MEPs than in PFPs. For example, only about half the countries that sought increases
in social expenditures in their PFPs mentioned such increases in MEPs.
In recent years, programs have relied on benchmarks and performance criteria to seek
increases in, and strengthen the efficiency of, social spending. To this end, the MEPs of six
countries included benchmarks (Armenia in 1996, Azerbaijan in 1997, Cameroon in 1997,
Georgia in 1997 and 1998, the Kyrgyz Republic in 1995, 1997, and 1998, and Uganda in 1997),
and two countries included performance criteria (Ghana in 1998 and the Kyrgyz Republic in
1998).

Overall, considerable progress was made in increasing social spending during 1985-97. Although
the lack of consistent data hinders the assessment of public social spending, program countries, on
average, have achieved an increase in social spending:
•

For 65 of the 107 countries with IMF-supported programs during 1985-97, govemment spending
on education and health care, on average, has increased both as percent of GDP and in real per
capita terms. The share in GDP of spending increased by 0.3 percentage point during the program
period (about eight years, on average); the spending increased by 2.4 percent a year in real per
capita terms (Figure 4.1).30
•
In a subset of 29 countries, of which 19 are ESAF countries, that have data on military spending,
such spending on average declined during 1990-97, whereas education and health care spending
together increased in relation to both GDP and total govemment spending.
Real per capita social spending has declined in some countries, and the increases have been
relatively low in some regions, notably education spending in sub-Saharan Africa.31 In transition
economies, real per capita spending on education and health care has declined considerably. A modest
decline of 0.1 of a percentage point in the share of spending in falling GDP masks a larger decline in real
per capita terms. In these countries, however, education and health care spending have been historically
high and inefficient.
Countries with ESAF-supported programs have shown relatively strong results. In the
31 countries with ESAF programs, the real per capita growth of spending on education and health care
over 1985-97 (4.0 percent and 4.9 percent, respectively) has outstripped, on average, that in other
program countries.
In 1997, public expenditure on education and health care as a share of GDP in countries with
ESAF programs approximated those in other program countries. In HtPCs, spending levels remain
below those in other program countries in part because of very low initial levels (Figure 4.2). Education
and health care spending as a share of total govemment spending—indicative of the priority assigned to
these types of spending—shows the same pattem.

30
31

See Gupta, Verhoeven, Yamada, and Tiongson (1999).
The countries where spending on education a a share of GDP declined are the Republic of Congo, Cóte d'Ivoire, Guineas
Bissau, Madagascar, Mali, Mozambique, a d Nigeria. In health care, t e Comoras, Guinea-Bissau, Nigeria, and Zambia
n
h
experienced a decline in spending as percent of GDP. The countries where real per capita spending on education declined are
t e Comoras, Cóte d'Ivoire, the Republic of Congo, Kenya, Madagascar, Mali, Nigeria, and Sierra Leone. In part, this
h
reflected cuts in salaiy from a high level (e.g., Cóte d'Ivoire). In health care, the countries where real per capita spending o
n
health care declined are the Comoras, Cóte d'Ivoire, Kenya, Madagascar, Nigeria, a d Zambia.
n

Sesión 4 - 357

Figure 4.1. Changes in Education and
Health Care Spending in Selected
Countries with IMF-Supported
Programs, 1985-97
(Average change between preprogram year
and latest year in percent of GDP; number of
countries in parentheses)
| ] Education g ] Health Care
B
T
13

1
.
1
09
i
0.7

as
Q3

ai
-ai
<13
AU program

ESAF (31)

countries (651

Sesión 4 - 358

HIPC (19)

S±>-Sataran
Africa (18)

Tramtian ( 1 9

Figure 4.2. Spending Levels on Education and Health Care in
Countries with IMF-Supported Programs, 1997
(ID percent of GDP; number of countries in parentbeses;
latest year for which data are available)1

•Education

All Program
Countries

ESAF (31)

• Health Care

HIPC (19)

Sub-Saharan Transition
Africa (18)
(16)

(65)

Sources: National authorities; and IMF staff.
'Mostly 1997.

Sesión 4 - 359

B.

Compositíonof Spending

Available data for 1985-97 suggest that budget expenditure shares shifted, on average, from
current to capital outlays in both ESAF countries and HIPCs, and more so than in other program countries.
Whether this led to an increase in such key components in the delivery of education and health care as
books and medicine, however, is unclear because available data for these outlays are reported with other
types of spending under other goods and services. In all program countries, average spending on other
goods and services fell during the period under review.
Although many programs have sought to improve the allocation of budget resources within the
education and health care sectors, more can be done. On average, program countries have devoted a
relatively large share of their education budget to tertiary education and even a larger part of health care
outlays to curative services (Figure 4.3). This suggests that low-income households would benefit from a
shift in budgetary resources toward primary education and basic health care.32

C.

Impact on Education and Health Indicators and Implications for Poverty

On average, the education and health care indicators in the OECD/UN/World Bank working set of
core indicators for measuring social development have improved for program countries.33 But there are
important exceptions. In sub-Saharan Africa, average life expectancy has declined, reflecting the toll of
HTV and conflicts (Table 4.1). Improvements in social indicators in ESAF countries and HIPCs have not
been commensurate with the spending increases. Progress in improving infant mortality and primary and
secondary enrollment has been slower in these countries than in other program countries. Transition
economies have experienced declines in enrollment rates in secondary education and immunizations;
reforms in these two areas have been slow, thus increasing theriskthat the declines in spending may lead
to a permanent setback in social indicators.34
Weak administrative capacity to formúlate and execute the budget has reduced the impact of
education and health care spending on social indicators. In particular, the capacity to spend resources
efficiently can vary at different levels of government, and is likely to be lacking at lower levels of
government, at least initially, during a period of devolution of expenditure responsibilities.35 The
allocation of budgetary resources within the social sectors (e.g., between primary and tertiary education) is
also important, as is the presence of corruption, which can distort the composition and level of social
spending.36
Although improvements in social indicators reflect a country's social development, they have not
been necessarily translated into reduced poverty, which in itself is multidimensional. For example, the
poor tend to be less educated and less healthy than the nonpoor.37 For targeted spending, to have a
considerable payoff, however, the benefits from improved basic social services have to be accompanied by
income-earning opportunities.
32

35
36

37

See World Bank (1993 and 1995); and Gupta, Verhoeven, and Tiongson (1999).
As yet, there is no consensus in international f r m on the precise social indicators to be used for assessing development a d
ou s
n
social performance. As noted in Box 2.1, there are a present four sets of indicators in use.
t
For a discussion of t e reasons for t e slow pace of reforms in education and health care spending in transition countries, see
h
h
Gupta (1998).
This has been a issue during t e decentralization of government in Ethiopia, see Ter-Minassian (1997, Chapter 20).
n
h
For example, if budget allocations in t e health sector are m d on t e basis of the n m e of beds in a hospital, hospital
h
ae
h
u br
administrators and doctors m y increase the number of beds and keep t e occupied, squeezing allocations for medicine. See
a
hm
also Tanzi (1998) and Mauro (1998).
See Anand and Ravallion (1993) and Bidani and Ravallion (1997).

Sesión 4 - 360

Figure 43. Allocation of Education and Health Care Spending in
Countries with IMF-Supported Programs, 1994

(In percent of total education and health care spending; number of countries in
parentheses; latest year for which data are available)1

Health Care (20)

Education (34)

Curative care

5%
0

Tertiary

Primary and secondary

Basic health care

Sources: IMF, Government Finance Statistics database; World Bank, Public Expenditure
Review and Poverty Assessment, various issues; and IMF staff estimates.
'Mostly 1994.

Data for 29 program countries show that the targeting of education and health care spending could
be improved, particularly in sub-Saharan Africa and in the transition economies (Figure 4.4). The poor's
access may be constrained by out-of-pocket costs for (both formal and informal) using public services,
excessive distance to the nearest school or health center, poor quality of public services, and gender bias.
For sub-Saharan Africa,
•

14 percent of total spending on public education and 12 percent of health care spending, on
average, accrue to the poorestfifth (quintile) of households compared with 30 percent for the
richest quintile for both. These gaps widen for spending on secondary and tertiary education and
hospital care;
•
spending on primary education is somewhat better targeted than that on secondary and tertiary
education, and the targeting of public spending on education and health care is improving in some
countries (e.g., Cóte d'Ivoire and Malawi).
Targeting has a geographical dimensión. For example, govemment public policy choices with a
pro-urban bias reduce the access to vital social services for the poor, most of whom live in rural areas.
Data on the geographic distribution of education spending were available for only 10 countries with
IMF-supported programs, and show that education spending, on average, has disproportionately favored
the urban population, particularly spending on secondary and tertiary education.38 A similar urban bias is
discemedfrom limited data on health care spending for six countries.

38

The rural population in Albania and Ghana, however, receives as m c as 70 percent of the benefits from public spending on
uh
primary education. Furthermore, the incidence of education spending has improved over time in favor of the rural population
in Malawi. See also Davoodi and Sachjapinan (forthcoming).

Sesión 4 - 3 6

Figure 4.4 Benefit Incidence of Public Spending on Education and Health Care in Countries with IMF-Supported

Programs, Early 1 9 s
90

(In percent of total spending; numherofprogram countries inpaientheses;
latest ye arfo rwttch data are available)

I Rural • Urban

• Bottom income quintile
• Top income quintile

All
education
(26)

D.

Primary
education
(29)

AU health
(23)

All education (8)

All health (6)

Experience with Program Targets, Conditionality, and Monitoring in ESAF Countries

A review of 11 countries with ESAF arrangements illustrates a range of approaches and success
with public spending on education and health care.39 It is difficult to establish any clear links between
meeting benchmarks, performance criteria, and prior actions, on the one hand, and improved access to
basic social services, on the other hand. Social indicators in ESAF countries, with the exception of the
transition economies (Armenia, Georgia, and the Kyrgyz Republic), were generally poor at the outset of
their programs. In nontransition countries, programs emphasized reorienting expenditures to raise
spending on education and health care; in Bolivia, Lao People' s Democratic Republic, and Uganda, déficit
and/or overall expenditure levels were also programmed to rise as a share of GDP, in part to raise
spending on education and health care. In the transition economies, where existing high and inefficient
levels of education and health care spending tottered on a collapsing revenue base, programs emphasized
quality and efficiency improvements.
In the nontransition countries the program objectives for education and health care spending were
set in the context of longer-term goals drawn from the authorities' strategic plans for poverty reduction.40

39

40

The countries were selected to cover a range of social policy strategies and progress in implementing policies, with a view to
identifying ways of strengthening t e social c n e t of IMF-supported programs. The ESAF arrangements reviewed cover the
h
o tn
lastfive years, during which operational staff guidelines on public social spending were issued. The countries include five
heavily indebted poor countries that have reached their decisión point under t e HIPC Initiative: Bolivia (with t e three-year
h
h
ESAF arrangements starting in 1 9 a d 1998); Burkina Faso (1998); Cóte d'Ivoire (1998); Mozambique (1996); a d Uganda
94 n
n
(1997). Other countries included in the review are Armenia (1996); Georgia (1996); Kyrgyz Republic (1994 and 1998);
Ghana (1995 and 1999); Lao People's Democratic Republic (1993); and Malawi (1995).
In some cases, strategic plans specifically for the education a d o health care sectors were also formulated.
n/r

Sesión 4 - 362

•

In several cases, the strategic plans were formal national govemment plans prepared in
collaboration with the World Bank and other development partners (e.g., Bolivia, Cote d'Ivoire,
Ghana, Malawi, and Uganda).
•
In only a few countries were quantified long-term goals for social outcomes included explicitly in
IMF-supported programs, for example, doubling the general literacy rate and increasing life
expectancy to 57 years (Burkina Faso).
•
For the HIPCs that reached the decisión point (Bolivia, Burkina Faso, Cóte d'Ivoire,
Mozambique, and Uganda) and Ghana (1999), the use of quantitative goals for indicators was
significantly expanded (see Tables 4.2 and 4.3). In these cases, goals established for education and
health care indicators were broadly in line with the OECD/UN/World Bank core set, with
adjustments made for the authorities' specific objectives and local conditions.
•
At the outset of the arrangements, the transition economies did not have overall poverty reduction
plans, although for Armenia and Georgia poverty assessments had been completed by the World
Bank at about the same time. In general, goals were east in qualitative terms. For example, the
long-term goals for spending on education and health care included a higher standard of living,
development of human capital, and alleviation of poverty (Bolivia); better provisión of social
services by enhancing the efficiency of social expenditures (Georgia); improving basic education
and health care and human resources (Ghana); and providing adequate fimding for human
development and improving the living standards (Lao People's Democratic Republic).
All of the programs reviewed included measures to improve fiscal govemance. In general, these
reforms addressed general expenditure management issues and covered the setting of budgetary spending
allocations and priorities, expenditure controls, and reporting, as well as transparency.41 In Mozambique,
reforms included the publication of budget plans and outcomes.
Programs often focused on improving cost-effectiveness and spending composition.
•

•

To this end, programs often included measures to reduce the cost of delivery of social services
(e.g., reducing excessive numbers of teachers) and to improve the quality of spending (e.g.,
improving the school curriculum and introducing second shifts in schools to overcome capacity
constraints on the feasible number of hours of teaching).
Program targets for public spending sought to reorient expenditure composition toward education
and health care, and, in particular, protect such spending during fiscal adjustment. Most of the
earlier ESAF-supported programs did not include quantitative budget targets for education and
health care spending at their outset, but these were introduced at a later stage (Table 4.4). This in
part reflected the introduction in 1997 of staff operational guidelines on social spending. All of the
programs of the heavily indebted poor countries that had reached their decisión points under the
HIPC Initiative have included quantitative targets. In the transition economies, part of the strategy
for increasing access to education and health care services has been to involve the prívate sector.

The scope of spending targets that could be monitored was constrained. The constraints included a
lack of data on the intrasectoral allocation of education and health care spending, incomplete coverage of
these sectors, and lags in the availability of data. Quantitative spending targets and benchmarks singled
out education and health care spending as priorities to be protected from cuts (Bolivia, Kyrgyz Republic,
Malawi, and Uganda), or aimed at redirecting public spending in favor of education and health care
(Ghana and Malawi) (see Table 4.4). Because the link between expenditure on education and health care
andfinal outcomes is complex and uncertain, programs monitored actual spending and developments in
intermedíate social indicators such as the hiring or firing of teachers and consolidation of schools
(Armenia), increasing water supply and number of classrooms created (Cóte d'Ivoire), reducing of stays in
41

In t e Kyrgyz Republic (1998) a d Burkina Faso (1996), some of t e reforms were directly targeted at social spending.
h
n
h

Sesión 4 - 36

hospitals (the Kyrgyz Republic), and increasing the share of textbooks in the budget (Malawi). Monitoring
was also undertaken with a view to assessing the impact of policies on the provisión of social services.
Program documents also have included qualitative assessments of progress.

Indicators

Education
Admission and repetition
rates
for primary
schools
Gross/net enrollment ratios
for
primary and/or
secondary
schools
Primary school completion
ratios
Gender ratios (to increase
ratio
of girls to
total pupils
enrolled in primary schools)
Book/pupii ratio
Cumulative number of
schools
benefiting from
participation in quality
improvement programs
Health
Life expectancy
Fertility rate
Infant and maternal
mortality
rate
Child and/or pregnant
woman
vaccination
rates
DPT vaccination coverage
Childhood malnutrition rate
Access to clean water
Utilization rate of health
centers
Proportion of health posts
staffed exclusively with
untrained personnel
Proportion of houses
receiving
measures against endemic
diseases (e.g., chagas,
malaria)

Table 4.2
Social Development Indicators, Selected ESAF Countries 1/
Other
ESAF
HIPO Decisión Point Countries
Country
Burkina
Bolivia
Ghana
Mozambique Uganda
Faso

X
X

X
X

X
X
X

X

X

1/ The countries included are a subset of the 11 ESAF countries for which the experience on program targets,
conditionality, and monitoring was reviewed.

Sesión 4 - 36

Table 4.3
Monitoring Arrangements, Selected ESAF Countries 1/

Ghana: A broad-based
Technical Committee on Poverty
comprising officials from various
ministries and government
departments, donors, and NGOs
was established in 1997 and
meets regularly. The committee
collects and disseminates
information on poverty reduction
programs. As part of this
Continuous Poverty Monitoring
System, a pilot survey has been
undertaken to provide a
systematic and frequent update
on key poverty indicators. Also,
donor committees meet
frequently on health and
education issues to coordínate,
exchange information, and
monitor programs.

Maíawi:
A governmental poverty
Uganda:
A poverty monitoring
and social policy monitoring unit unit including education, finance,
was established in 1995to
and health ministry officials was
monitor core social indicators to
established in 1998 and made
be presented to donors and at
responsible for collecting
Consultative Group meetings.
poverty and social sector
information at the local level.
Health districts collect quarterly
data on health-related indicators
and report them to the sector
monitoring unit. Expenditure
tracking studies were
undertaken to identify obstacles
that prevented budgeted
expenditures from reaching their
intended use.

1/ The countries included are a subset of the 11 ESAF countries for which the experience on program targets,
conditionality, and monitoring was reviewed.

A number of innovations have strengthened monitoring.
•

•

The introductíon of explicit targets has contributed to improved monitoring of developments. The
staff reports on programs of the HIPCs that had reached the decisión point (Bolivia, Burkina Faso,
Cote d'Ivoire, Mozambique, and Uganda) have systematically covered developments relating to
spending targets and social outcomes. In the other countries, the onset of improved monitoring
also has reflected the issuance of the staff guidelines on social spending. But significant
weaknesses remain in the quality of reporting in some countries. For instance, definitions of
targeted spending as set out in the initial request for the ESAF arrangement and those that were
subsequently monitored have sometimes been different. As noted earlier, information on actual
spending was typically available only with a considerable time lag, which meant that targets had
to be based on partial estimates for the preceding year(s). As a result, a clear picture of spending
developments and their impact on social indicators emerged only after the passage of several
annual programs.
In more recent ESAF-supported programs, formulating a clear framework in program documents
that integrated social spending targets with a time path of specified indicators and outcomes
improved the focus and monitoring of social policy. Such frameworks were used in the case of
HIPCs that had reached the decisión point and in Ghana's program (1999). In these cases, the
authorities explicitly noted their commitment to education and health care output targets in the
Memorándum of Economic Policies (MEP), reinforcing the emphasis given to social issues in the
Policy Framework Papers (PFP). The advantages of a well-defined framework also carried
through to more focused, comprehensive, and forward-looking assessments of education and
health care sector developments. In other countries, improved monitoring of education and health
care spending is evident in programs approved after 1997, following the introductíon of the staff

Sesión 4 - 36

guidelines on social spending, and program documents provide more specific information on
related developments and report spending on education and health care at a disaggregated level.
•
In the HIPC decisión point countries and Ghana, information to monitor social developments was
drawn from a wider variety of sources, including bilateral donors (Burkina Faso and Ghana).
These arrangements also made effective use of information from internal reviews conducted at the
local government and community levels. Also, in several countries, poverty monitoring teams and
units were set up (examples are given in Table 4.3).
In specifying public spending targets and policy measures and in monitoring, IMF staff
collaborated with the World Bank and regional development banks. The World Bank provided policy
analyses for many countries. Except for Burkina Faso and Georgia, however, comprehensive World Bank
Public Expenditure Reviews were not available at the time of the initial request for an ESAF arrangement
to guide budget policy. Thus, it was not always possible to ensure that budget allocations for education
and health care were in line with an appropriate overall composition of expenditures. Expenditure reviews
have since been undertaken or are scheduled to commence in 1999 in all 11 countries.
Conditionality42 was attached to public spending targets and to key reforms for which timely
implementation was essential to the success of the program.
•

•

42

Conditionality was used sparingly, and primarily took the form of benchmarks (Table 4.5).
Performance criteria and prior actions have rarely been used. For the most part, conditionality was
applied to minimum levels of budget spending with a view to protecting spending on education
and health care from the pressure of overall spending restraint, in parallel with World Bank
programs to improve the quality of social sector spending (Georgia and the Kyrgyz Republic), or
to ensure that additional resources were not diverted to other uses (Uganda).
Conditionality was used to encourage the timely completion of sector and national poverty
reduction strategies and action plans, which were prerequisites for establishing a clear operational
strategy for improving access to social services over the médium term. In Cote d'Ivoire a prior
action on adopting an antipoverty national plan was introduced to provide stronger evidence of the
authorities' commitment to strengthening spending on education and health care, an area in which
there had been slippage and an unmet benchmark in the previous ESAF arrangement. Establishing
performance criteria on education and health care spending in the Kyrgyz Republic and Uganda
programs was combined with measures to ensure the quality of such spending.

Policy measures that members intend to follow as a condition for the use of IMFfinancial resources.

Sesión 4 - 36

Table 4.4
Public Education and Health Care Spending Targets, Selected ESAF Countries 1/
Quantified
Expenditure
Country 2/
Monitoring
Definition
Comment
Target
3/
Transition Economies
Armenia
(1996)

Yes

Georgia
(1996)

Yes

Kyrgyz
Republic (1994)
Kyrgyz
Republic (1998)
4/

Current health and
education expenditures in
percent of GDP.
Social expenditures in
percent of GDP (1996).
Budgetary appropriations
for health and education
(from 1997).

No

Yes

Yes

Yes

Health and education
expenditures in percent of
GDP and total spending.

Yes

Bolivia (1995)

Yes

Yes

Bolivia (1999)

Yes

Burkina Faso
(1996)

Yes

Cote dDIvoire
(1998)

Yes

Mozambique
(1996)

Yes

Uganda
(1997)

Yes

Health and education
expenditures in percent of
GDP.
Health and education
expenditures in percent of
GDP.
Health and primary
education expenditures in
percent of GDP.
Health and education
expenditures in percent of
GDP.
Current expenditures on
health and education in
percent of GDP.
Health and education
expenditures in percent of
GDP.

Quantitative target specified
and monitored from the
second annual arrangement.
Quantitative targets specified
in the reports but definitions
differ across reports. Reports
include mostly quaiitative
discussion of social
expenditure developments,
but without reference to
previously specified targets.

HIPC at Decisión Point

Other ESAF Countries
Ghana
(1995)

Yes

Ghana
(1999)

Yes

Lao People's
Democratic
Republic
(1993)

No

Health and education
expenditures, excluding
foreign financed capital
expenditures, in percent of
GDP.
Health and education
expenditures, excluding
foreign financed capital
expenditures, in percent of
GDP.

Sesión 4 - 36

Yes

Quantitative target specified
and monitored from the third
annual arrangement.

Yes
Yes
Yes
Yes

Yes

Yet to be
reviewed

Quantitative target specified
and monitored from the
second annual arrangement.

Target specified and
monitored from the second
annual arrangement.

Country 71
Malawi

(1995)

Quantified
Expenditure
Target
Yes

Definition
Current expenditures on
health and education, in
percent of GDP.

Monitoring
3/
Yes

Comment
Quantitative target specified
and monitored from the third
annual arrangement.
Definition of targets differs
across reports. Qualitative
discussion of social
expenditure and outcome
developments, including with
reference to previously
specified quantitative target.

Source: IMF staff reports.
1/ A program is defined to have a quantitative target when either the staff report, MEP, and/or PFP provides a projection for the
category of public spending. This is a broader coverage than used in Box 6, which is restricted to MEPs and PFPs.
21 Year of approval in parentheses.
3/ Defined as a reference and/or a discussion of the developments with respect to the quantitative target specified in the request
for the ESAF arrangement or subsequent annual arrangements.
4/ Program has performance criteria set in nominal terms as a floor on expenditures on health and education.

Country
Armenia

Table 4.5
Social Policy Conditionality (Prior Actions, Performance Criteria, and Benchmarks)
Selected ESAF Countries 1/
Type of
ESAF
Policy Measure
Observance
Conditionality
Arrangement 21
ESAF. 3 (1998)

Cóte d'Ivoire

ESAF. 1 (1998)

Georgia

ESAF. 1 (1996)
ESAF. 2 (1997)
ESAF.3 (1998)

Ghana

ESAF.2 (1995)

Kyrgyz
Republic

ESAF.1 (1998)

Mozambique

ESAF.3 (1998)

Uganda

ESAF.1 (1997)
ESAF.2 (1998)

Develop and approve a three-year
strategic plan for the health sector by
end-September 1999.
Adopt a national plan to fight poverty.

Benchmark

Review
pending

Prior action

Met

Reduce number of budgetary positions1
primarily in education, by 10,000 by
September 1996.
Minimum amount of health expenditures of
the republican govemment cumulative
from January 1, 1997.
Minimum amount of health expenditures of
the republican govemment.
Complete medium-term expenditure
framework (MTEF) for priority sectors of
education, health, and roads.
Cumulative floor on budgetary
expenditures separately on health and
education beginning April 1998. 3/
Complete National Poverty Assessment
and Poverty Action Plan by end-1998.
Minimum amount of nonwage expenditures
on priority program areas in health and
education.
Minimum amount of nonwage expenditures
on priority program areas including
universal primary education component of
domestic development expenditures.

Benchmark

Met

Benchmark

Not met

Benchmark

Not met

Performance
criteria

Met

Performance
criteria

Waived 4/

Benchmark

Met

Benchmark

Met

Performance
criteria

Review
pending

Source: IMF staff reports.
1/ Social policy conditionalities were used in only 7 of the 11 ESAF countries.
21 Year of approval shown in parentheses; suffix indicates the program year (1 to 3).
3/ Program targets are set in terms of expenditures as a percent of GDP.
4/ Prior actions on the level of spending in the quarter following the test date were set as conditions for completing the mid-term
review, of which that on health was met. In addition, measures to strengthen overall expenditure management were to be
introduced and the authorities undertook to ensure that spending on education and health care would be kept at least constant in
real terms throughout 1999.

Sesión 4 - 36

V. Collaboration with the World Bank and Other International Agencies
The IMF is concerned with the social dimensions of its economic policy advice. The analysis and
policy advice on social issues, however, are to a large extent outside the principal areas of IMF expertise.
Between the Bretton Woods institutions, the primary responsibility for social policies lies with the World
Bank, and IMF staff rely upon the World Bank in this area. Other international institutions, such as
regional development banks and UN agencies (e.g., International Labor Organization, United Nations
Development Program, and World Health Organization), the donor community, NGOs, and civil society,
on a country-by-country basis, also can provide valuable inputs.
World Bank and IMF staffs have well-established procedures for collaboration, including on
social sector issues, in support of members' efforts to achieve sustainable growth and poverty reduction.43
On social sector issues, IMF staff looks to the World Bank for inputs on social sector policy goals,
analysis, reforms, and their budgetary cost, as well as data on social indicators. An iterative interaction
between the staffs of the IMF and the World Bank ensures the consistency of the overall macroeconomic
framework, including notably the fiscal envelope, with sustainable, cost-effective social policies and
strategies for poverty reduction.
Frequent interaction between the staffs of the World Bank and the IMF on social issues takes
place through several channels.
•

•
•
•

•
•

43

In addition to exchanges at headquarters, including through the review process, IMF missions to
program countries have usually involved parallel World Bank missions or the participation of
World Bank staff (and vice versa), as well as in-field consultations with resident representatives.
In ESAF countries, overlapping missions are the norm.
In ESAF-supported programs, the PFP includes a separate section on poverty and social sector
issues.
Reflecting the joint nature of the HIPC Initiative and its specific emphasis on achieving social
improvements and development, collaboration is intensive in HIPCs, particularly as they approach
their decisión points under the Initiative.
In the fall of 1998, a pilot program for enhanced World Bank-IMF collaboration in low-income
(ESAF/IDA) countries was launched in six countries (Cameroon, Ethiopia, Nicaragua, Tajikistan,
Vietnam, and Zimbabwe). The pilot includes a specific focus on social sector issues; in particular,
it is envisaged that the World Bank would not only identify measures to mitígate adverse effects
on the poor and vulnerable, but also assess the social impact of program design more broadly, ex
ante and ex post.
Periodically, joint institution-wide, forward-looking reviews of work plans and priorities in public
sector work are carried out. These reviews aim, among other things, to coordínate IMF and World
Bank public sector work and to ensure the timely availability of public expenditure analyses.
At the general policy level, the IMF's Fiscal Affairs Department and their World Bank
counterparts in the Poverty Reduction and Economic Management and Human Development

Collaboration between the World Bank and IMF has been periodically reviewed and staff guidance notes o collaboration in
n
certain policy areas have been issued.

Sesión 4 - 36

Networks have recently initiated regular senior-level meetings to coordínate work programs and
to help resolve collaboration issues that may arise.
The importance of social sector issues for the IMF has led to a strengthening of collaboration with
the UN system and regional development banks. For example, in the area of labor market and related
social policy reform, guidelines were issued in 1996 to IMF staff on collaboration with the International
Labor Organización (ILO) that provide for more systematic contacts between staff at the country level,
especially through resident representatives. Also, pilot countries were selected for enhanced IMF-DLO
collaboration, and interaction on general policy issues has been increased, most recently in the context of
the Asian crisis.44 In late 1998, the World Health Organization (WHO) and the MF strengthened
collaboration on health-related issues in low-income countries.
World Bank-IMF collaboration in the social area, however, has not always been seamless. In part,
this reflects a mismatch between the timetables adopted for World Bank and IMF work programs and the
operational methods of IMF- and World Bank-supported programs, which, in some respects, are not
well-attuned to each other since IMF-supported programs are typically formulated under relatively tight
deadlines, often demanded by a crisis situation.
•

The pressing country requirements can, unfortunately, cut across the grain of the longer timeframe of preparatory work, including a broad participatory approach for social policy formulation.
This highlights the importance of an ex ante dialogue and ongoing policy analysis and
recommendations in the social policy area.
•
Sufficient time is also not always available to collect information to develop well-targeted social
sector reforms.
•
On occasion, World Bank input has not been available within the time-frame required by the
countries' circumstances, and the IMF staff and the country authorities have devised policies as
best they could. In such cases, the focus has tended to be on mitigating the adverse impact on
vulnerable groups through specific social protection mechanisms, rather than in terms of program
design based on ex ante social assessments.
From an IMF perspective, more World Bank involvement, including in monitoring and following
up on social sector issues, would be desirable, particularly for ESAF countries. The World Bank has
shifted away from comprehensive Public Expenditure Reviews (PERs) toward a sequence of more indepth sector-specific expenditure reviews, which are useful in their own right, but do not provide a
comprehensive analysis of budget priorities. The World Bank staff understandably focuses their work
programs on the Country Assistance Strategies (CAS's) and their own lending operations. Thus, the PFP
is not regarded as a priority in the work agenda of World Bank staff. It is no longer discussed by the
World Bank's Executive Board and is generally not directly relevant to World Bank's operations, which
contrasts with the role of the PFP at the IMF. This difference has created gaps in needed inputs for
ESAF-supported and other programs and has, in some cases, led IMF staff to work with the authorities to
fill the gap through IMF technical assistance.
The most recent internal assessment of the World Bank's support of poverty reduction found that,
relative to past benchmarks, performance has been good (World Bank, 1999). At the same time, it raised
many of the same concems noted above.45 As a result, mechanisms for more active monitoring of
44

45

IMF sénior staff have participated in the high-level tripartite ILO meetings in Bangkok in 1997, 1 9 , a d 1999, and, in May
98 n
1998, the IMF organized a seminar with the ILO to improve t e staff s understanding of core labor standards and t e ILO's
h
h
role in setting and monitoring these standards.
The World Bank report underscored three points: (i) only somewhat more t a half of the Country Assistance Strategies
hn
prepared in FY1998 were judged fully satisfactory in their integration of poverty issues into the framing of a forward-looking
strategy; (ii) progress in completing poverty a s s m n s h s at times been slow, reflecting a greater use of participatory and
ses e t a
consultative methods, staff resource constraints, a d t e degree of domestic political commitment; and (iii) m c less
n h
uh

(continued...)
Sesión 4 - 3

developments in these areas are being implemented. Various World Bank initiatives should contribute to a
better and more systematic integration of social sector issues in World Bank lending operations and henee
in IMF-supported programs:
•

developing the Comprehensive Development Framework (CDF), which includes countries' social
sector objectives and policies;
•
strengthening the poverty focus in Country Assistance Strategies;
•
preparing social and structural policy reviews;
•
improving the delivery of Public Expenditure Reviews and their synchronization with countries'
budget processes;
•
implementing the joint ESAF/IDA pilot scheme; and
•
developing and implementing the Principies and Good Practices in Social Policy.
World Bank-IMF collaboration in the poverty reduction and social policy area could be
strengthened by the formulation, together with the authorities, of a Poverty Reduction Strategy Paper
(PRSP) through a participatory process. The PRSP would aim at ensuring the consistency between a
country's macroeconomic, structural, and social policies, and the goals of poverty reduction and social
development. It would
•

be owned by the govemment and endorsed by the Boards of the World Bank and the IMF as a
basis for the institutions' operations;
•
include a macroeconomicframework and structural and social policies consistent with the poverty
reduction and social goals;
•
set out technical assistance needs and expected providers; and
•
identify overall externalfinancing needs.
When all these elements are available in the PRSP, it could replace the Policy Framework Paper
(PFP), and provide a framework for World Bank and IMF lending operations for the countiy.46 While the
PRSP might be prepared every three years, there would be annual updates.

A.

Postscript

Following the Executive Board's consideration of this and other papers, important changes were
made to the ESAF to give the goal of poverty reduction greater prominence, and improve its effectiveness
as a policy instrument to favor growth and reduce poverty.
The key change is that the complementarity of macroeconomic, structural, and social policies will
be given greater recognition. These policies would be articulated in a Poverty Reduction Strategy Paper
(PRSP) which would integrate these policies in a mutually reinforcing manner. Many of the
recommendations for strengthening the design and implementation of social policies, particularly in
efficiency and targeting of public spending would feed into the PRSP. Importantly, the PRSP prepared in
collaboration with the IMF and World Bank, will be nationally owned: the preparation would be
govemment-led based on an open consultative process, including with civil society. The PRSP would also
serve as the vehicle for strengthened IMF-World Bank collaboration. Reflectíng these changes, the ESAF
was transformed in December 1999 into the Poverty Reduction and Growth Facility.

46

progress has been m d in evaluating t e impact of specific interventions (only 1 percent of Country Assistance Strategies
ae
h
3
include monitorable poverty benchmarks t a were time-bound a d output- or outeome-oriented).
ht
n
A m r detailed description of t e proposed process a d t e PRSP is provided in t e Bank-Fund paper, "HIPC Initiative—
oe
h
n h
h
Strengthening the Link Between Debt Relief and Poverty Reduction," which is available on the Internet at
http://www.worldbank.org/hipc/hipc-review/hipc-review.html.

Sesión 4 - 3

References
Abed, George, and others (1998), Fiscal Reforms in Low-income Countries: Experience Under IMFSupported Programs, IMF Occasional Paper No. 160, Washington: International Monetary Fund.
Anand, Sudhir, and Martin Ravallion (1993), "Human Development in Poor Countries: On the Role of
Prívate Incomes and Public Services," Journal of Economic Perspectives, Vol. 7 (Winter), No. 1,
pp. 133-50.
Aninat, Eduardo, Andreas Bauer, and Kevin Cowan (1999), "Addressing Equity Issues in Policymaking:
Lessons from the Chilean Experience," in Economic Policy and Equity, ed. by Vito Tanzi, Keyoung Chu, and Sanjeev Gupta, Washington: International Monetary Fund.
Barro, Robert (1995), "Inflation and Economic Growth," Bank of England Quarterly Bulletin, Vol. 35
(May), pp. 166-76.
Bidani, Benu, and Martin Ravallion (1997), "Decomposing Social Indicators Using Distributional Data,"
Journal of Econometrics, Vol. 77 (March), No. 1, pp. 125-93.
Bredenkamp, Hugh, and Susan Schadler (1999), Economic Adjustment and Reform in Low-income
Countries, Washington: International Monetary Fund.
Bruno, Michael, and William Easterly (1995), "Inflation Crises and Long-Run Growth," NBER Working
Paper No. 5209, Cambridge, Massachusetts: National Bureau of Economic Research.
Bulír, Ales (1998), "Income Inequality: Does Inflation Matter?" IMF Working Paper 98/7, Washington:
International Monetary Fund.
, and Anne-Marie Gulde (1995), "Inflation and Income Distribution: Further Evidence on
Empirical Links," IMF Working Paper 95/86, Washington: International Monetary Fund.
Chu, Ke-young, and S. Gupta, eds. (1998), Social Safety Nets: Issues and Recent Experiences, Washington:
International Monetary Fund.
Cox, Donald, Zekeriya Eser, and Emmanuel Jimenez (1997), "Family Safety Nets During Economic
Transition," in Poverty in Russia: Public Policy and Prívate Responses, ed. by Jení Klugman,
Washington: Economic Development Institute of the World Bank.
Cox, Donald, Wlodek Okrasa, and Emmanuel Jimenez (1997), "Family Safety Nets and Economic
Transition: A Study of Households in Poland," Review of Income and Wealth, Vol. 43 (June), No.
2, pp. 191-209.
Davoodi, Hamid R., and Sawitree Sachjapinan, forthcoming, "How Useful are Benefit Incidence
Studies?" IMF Working Paper, Washington: International Monetary Fund.
Demery, Lionel, and Lyn Squire (1996), "Macroeconomic Adjustment and Poverty in Africa: An
Emerging Picture," The World Bank Research Observer, Vol. 11 (February), No. 1, pp. 39-60.
Easterly, William, and Sergio Rebelo (1993), "Fiscal Policy and Economic Growth: An Empirical
Investigation," NBER Working Paper No. 4499, Cambridge, Massachusetts: National Bureau of
Economic Research.
Sesión 4 - 3 2

Ferreira, Francisco, Giovanna Prennushi, and Martin Ravallion (1999), "Protecting the Poor from
Macroeconomic Shocks: An Agenda for Action in a Crisis and Beyond," Policy Research
Working Paper No. 2160, Washington: World Bank.
Filmer, Deon, Jeffrey Hammer, and Lant Pritchett (1998), "Health Policy in Poor Countries: Weak Links
in the Chain," Policy Research Working Paper No. 1874, Washington: World Bank.
Fischer, Stanley (1991), "Growth, Macroeconomics, and Development," NBER Working Paper No. 3702,
Cambridge, Massachusetts: National Bureau of Economic Research.
Frankenberg, Elizabeth, Duncan Thomas, and Kathleen Beegle (1999), 'The Real Costs of Indonesia's
Economic Crisis: Preliminary Findings from the Indonesia Family Life Surveys," Labor and
Population Program Working Paper 99-04, Santa Monica: RAND.
Ghosh, Atish, and Steven Phillips (1998), "Warning: Inflation May Be Harmful to Your Growth, " Staff
Papers, International Monetary Fund, Vol. 45 (December), pp. 672-710.
Goldsbrough, David, and others (1996), Reinvigorating Growth in Developing Countries: Lessons from
Adjustment Policies in Eight Economies, IMF Occasional Paper No. 139, Washington:
International Monetary Fund.
Guitián, Manuel (1998), "Monetary Policy: Equity Issues in IMF Policy Advice," in Income Distribution
and High-Quality Growth, ed. by Vito Tanzi and Ke-young Chu, Cambridge, Massachusetts: MIT
Press.
Gupta, Sanjeev (1998), "Economic Transition and Social Protection: Issues and Agenda for Reform," IMF
Paper on Policy Analysis and Assessment 98/14, Washington: International Monetary Fund.
, Marijn Verhoeven, and Keiko Honjo (1997), 'The Efficiency of Government Expenditures:
Experiences from Africa," IMF Working Paper 97/153, Washington: International Monetary
Fund.
, and others (1998), The IMF
International Monetary Fund.

and the Poor, IMF Pamphlet Series, No. 52, Washington:

, Marijn Verhoeven, and Erwin Tiongson (1999), "Does Higher Government Spending Buy Better
Results in Education and Health Care?" IMF Working Paper 99/21, Washington: International
Monetary Fund.
, Marijn Verhoeven, Gustavo Yamada, and Erwin Tiongson (1999), "Education and Health
Spending Continúes Rise in Countries with IMF-Supported Programs," IMF Survey, Vol. 28
(March), pp. 79-80.
Harberger, Arnold C. (1998), "Monetary and Fiscal Policy for Equitable Economic Growth," in Income
Distribution and High-Quality Growth, ed. by Vito Tanzi and Ke-young Chu, Cambridge,
Massachusetts: The MIT Press.
Iglesias, Enrique V. (1999), "Equity Issues in Latin America," in Economic Policy and Equity, ed. by Vito
Tanzi, Ke-young Chu, and Sanjeev Gupta, Washington: International Monetary Fund.

Sesión 4 - 3

International Monetary Fund (1986), Fund-Supported Programs, Fiscal Policy, and Income Distribution,
IMF Occasional Paper No. 46, Washington: International Monetary Fund.
(1995), Social Dimensions ofthe IMF's Policy Dialogue, prepared by the Staff of the International
Monetary Fund, IMF Pamphlet Series, No. 47, Washington: International Monetary Fund.
(1996), "Partnership for Sustainable Global Growth," an Interim Committee Declaration, IMF
Survey, October 14, p. 327.
(1997), The ESAF at Ten Years: Economic Adjustment and Reform in Low-income Distribution,
IMF Occasional Paper No. 156, Washington: International Monetary Fund.
Levine, Ross, and Sara Zervos (1993), "Looking at the Facts: What We Know About Policy and Growth
from Cross-Country Analysis," Policy Research Working Paper No. 1115, Washington: World
Bank.
Mauro, Paulo (1998), "Corruption and the Composition of Government Expenditure," Journal of Public
Economics, Vol. 69 (August), No. 2, pp. 263-79.
Milanovic, Branko (1994), "Determinants of Cross-Country Income Inequality: An 'Augmented'
Kuznets' Hypothesis," Policy Research Working Paper No. 1246, Washington: World Bank.
Mingat, Alain, and Jee-Peng Tan (1998), "The Mechanics of Progress in Education: Evidence from CrossCountry Data," Policy Research Working Paper No. 2015, Washington: World Bank.
Sahn, David E., Paul A. Dorosh, and Stephen D. Younger (1997), Structural Adjustment Reconsidered:
Economic Policy and Poverty in Africa, New York: Cambridge University Press.
Sarel, Michael (1996), "Nonlinear Effects of Inflation on Economic Growth," Staff Papers, International
Monetary Fund, Vol. 43 (March), No. 4, pp. 199-215.
(1997), "How Macroeconomic Factors Affect Income Distribution: The Cross-Country
Evidence," IMF Working Paper 97/152, Washington: International Monetary Fund.
Sen, Amartya (1999), "Economic Policy and Equity: An Overview," in Economic Policy and Equity, ed.
by Vito Tanzi, Ke-young Chu, and Sanjeev Gupta, Washington: International Monetary Fund.
Subbarao, K., and others (1997), Safety Net Programs and Poverty Reduction: Lessons from CrossCountry Experience, Washington.
Tanzi, Vito (1998), "Corruption Around the World: Causes, Consequences, Scope, and Cures," Staff
Papers, International Monetary Fund, Vol. 45 (December), pp. 559-94.
, and Ke-young Chu, eds. (1998), Income Distribution and High-Quality Growth, Cambridge,
Massachusetts: The MTT Press.
, Ke-young Chu, and Sanjeev Gupta, eds. (1999), Economic Policy and Equity, Washington:
International Monetary Fund.
Ter-Minassian, Teresa, ed. (1997), Fiscal Federalism in Theory and Practice, Washington: International
Monetary Fund.

Sesión 4 - 3

World Bank (1993), World Development Report 1993: Investing in Health, Washington: World Bank.
(1995), Priorities and Strategies for Education: A World Bank Review, Washington: World Bank.
(1996), Social Dimensions of Adjustment: World Bank Experience, 1980-93, Washington: World
Bank.
(1999), Poverty Reduction and the World Bank: Progress in Fiscal 1998, Washington: World
Bank.

Sesión 4 - 3

IMF Executive Board Discussion
(Chairman's Summing-Up, September 19,1999)
Directors considerad that the IMF's primary role is to promote macroeconomic stability and
structural reforms necessary for achieving sustainable and rapid growth. They underscored the crucial
importance of economic growth for poverty alleviation, but also recognized that the IMF must be sensitive
to the social implications of its policy advice. In particular, Directors noted that IMF-supported programs
have sought to help members address the potential adverse impact on vulnerable groups of their
adjustment and reform efforts as well of exogenous shocks, and that such efforts in turn can make a vital
contribution toward sustaining economic reforms and protecting living standards. They also observed that
sound macroeconomic policies, coupled with effective social and infrastructure spending, foster faster
long-term growth. In light of these considerations, Directors observed that social safety nets and
appropriately targeted productive public spending, particularly in the social area, can provide critical
support for the success of members' adjustment and reform programs.
Broader requirements for improving living standards were also discussed, including promoting
faster growth and employment creation and better integrating poorer countries into the international
economic system. In addition, Directors suggested that the international community should work to
improve these countries' access to industrial country markets, as well as to halt the excessive flow of
weapons to developing countries. Directors stressed the importance of good governance, transparency, and
accountability for ensuring the effective use of public resources.
Directors discussed the role of the IMF with regard to social policies. Recognizing the need for
mutually reinforcing macroeconomic and social policies, they underscored the importance of more closely
integrating, with the help of the World Bank, social issues and poverty concerns into IMF-supported
programs. Directors agreed that greater attention to social issues was necessary in the context of lowincome countries, including HIPCs, where structural reforms are particularly critical. With regard to other
countries, a variety of views was expressed on the degree and modalities of IMF involvement in social
issues. Some speakers attached importance to these issues in countries other than low-income countries,
pointing to the recent experience in the Asian crisis countries and elsewhere. However, a number of
Directors, while concurring that IMF-supported programs should encompass members' social policies and
poverty reduction efforts, cautioned nevertheless that the IMF should not allow its primary mandate to be
diluted. They viewed the World Bank as taking the leading role in developing adequate social safety nets
and effective social policies, and the IMF as contributing to poverty reduction mainly through its support
of economic policies that provide a conducive environment for sustained growth.
All Directors emphasized that, as regards social issues, the World Bank and other relevant
international organizations have the primary mandate and expertise. In this connection, they noted, for
example, that the relationship between public social spending, growth, and poverty is complex, and that it
is therefore critical to ensure that spending is used productively. Several Directors cautioned that the IMF
did not have the panoply of expertise needed to assess the quality of social spending and related issues.
Directors, therefore, underscored that the social components of countries' IMF-supported programs should
draw, to the fullest extent possible, on the work of the World Bank and other relevant institutions. Some
Directors were of the view that when timely inputs from these institutions on essential social components
of IMF-supported programs were not available, IMF staff would necessarily have to provide policy advice
to the extent feasible. Many other Directors were skeptical about such IMF involvement, particularly if it
were to require additional staff expertise. All Directors suggested that, to facilítate such cooperation, these
institutions should be encouraged to provide more timely input. Overall, Directors agreed that more
intensive cooperation between the IMF and the Bank is essential, proceeding along the lines of these
institutions' respective responsibilities and comparative advantages, and thus avoiding duplication of

Sesión 4 - 3 6

efforts. They welcomed therefore the recent enhanced IMF-Bank cooperation, pointing to the Enhanced
Structural Adjustment Facility/International Development Association (ESAF/IDA) pilot program and the
preparation of the reports for this discussion.
Turning to specific issues in the area of social policies, Directors noted the scope for further
improving the quality and implementation of social safety nets, through comprehensive ex ante analyses
and monitoring, relying on the expertise of the World Bank and other organizations. Several Directors
recommended that the staff should assess, in the course of surveillance, the adequacy of social policy
instruments, the performance of social safety nets, and the potential social ramifications of
macroeconomic andfinancial policies. Many others, however, cautioned that this should not detract from
the appropriate focus of Article IV surveillance. We will have to come back to this matter on the occasion
of the biennial review of surveillance.
Spending on education and health care has increased in real per capita terms and in relation to
GDP in most countries with IMF-supported programs during the past decade. While this has been
accompanied by an improvement in a broad range of social indicators, Directors noted the diversity in
outcomes caused by the differences in the effectiveness of social spending. Although some cautioned that
too much emphasis on the absolute amount of social spending could send the wrong message, Directors
more generally stressed the importance of efficient and well-targeted spending for ensuring that gains in
social indicators are commensurate with spending increases. Further improvements in these areas could be
achieved, inter alia, by strengthening a country's budget formulation and implementation capacity.
The establishment of national quantitative targets for poverty reduction—consistent with the
International Development Goals for 2015 to which countries have subscribed—could also prove
beneficial, especially if the higher targeted spending is used productively. Directors suggested that, in
setting targets, spending needs in priority areas for poverty reduction other than health, education, and
social safety nets—such as basic sanitation, rural roads, and access to clean water—should also be taken
into account; such priority spending may contribute as much or more to poverty reduction. Some Directors
thought that core labor standards had a valuable contribution to make to the achievement of these targets,
while other Directors were concerned that this issue is outside the IMF's main areas of responsibility. We
will come back to this issue after further consultations with the World Bank and the ILO.
Directors considered that in countries where social spending is so low as to be a critical area of
weakness, structural benchmarks could continué to be used selectively to protect social spending and to
promote key institutional reforms. While many Directors thought that such benchmarks should only be
used in ESAF-supported programs, some other Directors saw the valué in applying performance indicators
(performance criteria and structural benchmarks) to a broader range of IMF-supported programs. In
establishing such structural benchmarks, IMF staff will rely on inputsfrom the World Bank and others to
ensure, inter alia, the targeting and quality of spending.
Directors noted with concern the widespread poor quality of data on social spending, social
indicators, and social protection arrangements, which inhibited the design and implementation of effective
social programs. They saw an urgent need for country authorities to identify weaknesses in data and data
collection, and to make data improvements in collaboration with the World Bank, other international
agencies, and civil society.

Sesión 4 - 3

Glossary
Benchmarks. In the context of IMF-supported programs, a point of reference against which program
implementation is monitored. Benchmarks are not necessarily quantitative and frequently relate to
structural variables and policies.

Comprehensive Development Framework (CDF). A holistic approach to development developed by the

World Bank and implemented on a pilot basis since early 1999. The CDF highlights the interdependence
of all elements of development—social, structural, human, govemance, environmental, economic, and
financial—and stresses ownership by the country and partnership with government, civil society,
assistance agencies, and the private sector.
Conditionality. Economic policies that members intend to follow as a condition for the use of IMF
resources. These are often expressed as performance criteria (e.g., monetary and budgetary targets) or
benchmarks, and are intended to ensure that the use of IMF credit is temporary and consistent with the
objectives of adjustment program designed to correct a member's macroeconomic and structural
imbalances, and promote stronger growth and extemal payments viability.

Country Assistance Strategy (CAS). The central vehicle for World Bank Executive Board review of its

assistance strategy for International Development Association (IDA) and its borrowers. The CAS
document describes the strategy based on an assessment of priorities in the country and indicates the level
and composition of assistance to be provided based on the strategy and the country's portfolio
performance.

Enhanced Structural Adjustment Facility (ESAF). This facility, established by the Executive Board in

1987 and extended and enlarged in February 1994, is the principal means by which the IMF provides
financial support, in the form of highly concessional loans, to low-income member countries facing
protracted balance of payments problems and loans and grants under the HIPC Initiative. In late 1999, the
ESAF was succeeded by the Poverty Reduction and Growth Facility (PRGF).

ESAF/IDA pilot project. Program for enhanced World Bank-IMF collaboration in low-income

(ESAF/IDA) countries, including a specific focus on social sector issues in six countries—Cameroon,
Ethiopia, Nicaragua, Tajikistan, Vietnam, and Zimbabwe.

Heavily Indebted Poor Countries (HIPC) Initiative. Initiative designed to provide exceptional
assistance to heavily indebted poor countries following sound economic policies to help them reduce their
extemal debt burden to sustainable levels. In the second half of 1999, the Initiative was strengthened to
provide deeper and faster debt relief through the Enhanced HIPC Initiative.
Memorándum of Economic Policies (MEPs). Document prepared by a member country describing the
policies that it intends to implement in the context of its request for financial support from the IMF.

Performance Criteria. Macroeconomic indicators such as monetary and budgetary targets that must be
met for the member to qualify for phased purchases under Stand-By Arrangements in the upper credit
tranches, Extended Fund Facility (EFF), and Enhanced Structural Adjustment Facility Arrangements
(ESAF).
Policy Framework Paper. Document prepared by a member country in collaboration with the staffs of

the IMF and the World Bank and updated annually, which describes the authorities' economic objectives,
macroeconomic and structural policies for a three-year adjustment program supported by resources under

Sesión 4 - 3

the Enhanced Structural Adjustment Facility (ESAF), as well as associated externalfinancing needs and
major sources of financing.

Poverty Reduction and Growth Facility (PRGF). The IMF's successor arrangement to the Enhanced

Structural Adjustment Facility (ESAF). In the PRGF, poverty reduction is a key element of a renewed
growth-oriented economic strategy. The cornerstone of the new approach is a nationally owned,
comprehensive, Poverty Reduction Strategy Paper (PRSP).

Poverty Reduction Strategy Paper (PRSP). The central document in the new Poverty Reduction and

Growth Facility (PRGF) and will identify priorities for public action to achieve the greatest impact on
poverty reduction. It will also address the critical, and often complex issues related to enhancing good
governance and supporting transparency in policy making. The PRSP will be country driven, being
prepared by the authorities with assistance from the World Bank and the IMF, and reflect the outcome of
an open, participatory process involving civil society, relevant international institutions, and donors.

Public Expenditure Review (PER). The major vehicle for the World Bank to analyze public sector

issues, and general public expenditure issues in particular, as part of its economic and sector work with its
member countries. PERs help countries establish effective and transparent mechanisms to allocate and use
available public resources in a manner that promotes economic growth and helps in reducing poverty.

Stand-By Arrangements. Stand-By Arrangements give member countries the right to draw up to a

specified amount of IMF resources during a prescribed period (usually 12-18 months, although they can
extend up to three years). The release of drawings is conditional upon meeting performance criteria and
the completion of periodic reviews.

Sesión 4 - 3

Sesión 4 - 380

ECONOMIC SHOCKS, SAFETY NETS AND
FISCAL CONSTRAINTS: SOCIAL PROTECTION
FOR THE POOR IN LATIN AMERICA

by
Norman Hicks and Quentin Wodon *
World Bank

* This paper was prepared for the XII Seminario Regional de Política Fiscal in Chile, in January 2000.
The paper relies in part on the last chapter of a regional study on poverty and policy (Wodon, 2000)
completed with funding from the Office of the Chief Economist (Guillermo Perry) for the Latin America
and Caribbean región at The World Bank. The authors are grateful to Ana-Maria Amagada, Judy Baker,
Charles Griffin, Margaret Grosh, Kathy Lindert, and William Maloney for useful discussions. The views
presented in this paper are those of the authors, and need not represent those of the World Bank, its
Executive Directors, or the countries they represent

Sesión 4 - 3

Abstract
Confronted with recurrent macroeconomic shocks, governments in Latin America and the
Caribbean have increasingly been concerned about establishing or strengthening systems of
social protection and safety net programs. The goal of these programs is to help mitígate the
impact on the poor of shocks before they occur, and to help the poor cope with the shocks
once they have occurred. In this paper, we focus on publicly funded or mandated safety nets
functioning as risk-coping mechanisms. The paper reviews the characteristics of a good
safety net against the main types of safety nets currently in place. In general, itfinds that no
single program meets all of the criteria in terms of efficiency and effectiveness, although
some are better than others. Finally, what has been the actual record in terms of protecting
the poor through targeted public spending during crises? The paper finds that because of
fiscal constraints during a crisis, social spending is often pro-cyclical when ideally it should
be counter-cyclical.

I.

The context: Persistent poverty and vulnerability to shocks

The Latin America and Caribbean Región (LAC) has been plagued by macroeconomic shocks over
the last twenty years, with serious consequences for the poor (e.g. Glewwe and Hall, 1998; Lustig, 1995,
1999; Ganuza et al., 1999.) Macroeconomic shocks are so-called "covariant" shocks (Holzmann and
Jorgensen, 1999) which affect the real incomes of a large share of the population, typically through a
reduction in both real hourly wages (via inflation) and the number of hours worked (via unemployment or
underemployment.)
Beyond these income effects which may vanish once growth resumes,
macroeconomic shocks may also have longer term consequences. For example, the reduction in the
quantity and quality of public health care due to budgetary cuts during a crisis may induce irreparable
damage. Also, when coping with a crisis, parents may send their children to work in order to compénsate
for their own loss of income1. If there is substitution between child labor and schooling, and if the children
do not return to school at a later stage, they will incur long term wage losses due to their lower endowment
in human capital (Wodon and Siaens, 2000.) All this may help explain why De Janvry and Sadoulet
(1999)find evidence that the increase in poverty that follows a macroeconomic shock may be larger than
the subsequent reduction in poverty with growth.
There are signs that the reforms enacted in many countries of the región in the 1990s have been
bearing at least some fruits. Wodon et al. (2000) estimates that in 1996, slightly more than one third of the
LAC population (36.7 percent) was poor (i.e., not able to afford basic food and non food needs), and one
out of every six person (16.1 percent) was extremely poor (i.e., not able to afford basic food needs.) This
1

The impact of adult wages on child labor has been discussed a o g others by Basu and Van (1998.) In testing their theory, Ray
mn
(1998)finds t a higher wages for adults reduce t e probability that children will be working in Perú, but not in Pakistan.
ht
h
Psacharopoulos (1997)finds that the impact of household income on schooling and child labor is significant in Venezuela, but
less so in Bolivia. In Peni, Patrinos and Psacharopoulos (1997)find a impact of family income on a measure of age-grade
n
distortion for children, but no impact on child labor. Jacoby a d Skoufias (1997)find t a child labor helps s o t rural
n
ht
m oh
incomes in India. For Perú, Ilahi (1999)finds a gender bias. Although t e educational a t i m n of boys a d girls are t e s m
h
t an e t
n
h a e
on average, the schooling of girls is m r sensitive to shocks s c as changes in adult female employment or t e sickness of
oe
uh
h
household members.

Sesión 4 - 3 2

represents a progress versus 1992 when the incidences of poverty and extreme poverty were both higher.
However, the absolute number of the poor has not been reduced as much due to population growth.
Moreover, if the comparison is made with 1986 instead of 1992, the numbers of the poor and extreme
poor in 1996 have risen considerably (see Lustig and Arias, 2000, for a survey of poverty estimates.)
Projections of further poverty reduction between 1996 and 1998 using elasticities of poverty reduction to
growth and actual levels of growth observed in LAC suggest only limited gains in percentage terms, with
the numbers of the poor and extreme poor remaining constató2.
Table 1
Population and number of poor and extreme poor ¡n LAC, 1986-1998

Year

Population
(millions)

Share (%) of
population
poor

Number of
poor
(millions)

1986
1989
1992
1995
1996
1998*

407.38
430.98
454.65
478.21
486.06
501.87

33.75
38.26
39.65
36.92
36.74
35.83

137.49
164.89
180.27
176.56
178.58
179.84

Share (%) of
population
extremely
poor
13.32
17.59
18.65
15.94
16.10
15.55

Number of
extreme poor
(millions)

Source: Wodon et al. (2000.) Poverty numbers for 1998 are based on projections, not surveys.

54.26
75.81
84.79
76.23
78.26
78.05

Apart from having suffered from macroeconomic instability and high levels of poverty, Latin
American households have also been affected by the forces of globalization. The move towards greater
openness has generated substantial benefits in many countries of the región, but it has also facilitated the
transmission of international shocks and, as importantly, it has accelerated the adjustment of the región's
economies to the requirements of international competition. Openness may have resulted in a reduction in
employment security (Rodrick, 1999) and in widening wage differentials between better and less skilled
workers3. While the current changes in labor markets should ultimately bring net positive aggregate gains,
they may also induce idiosyncratic (i.e., household specific) shocks for the individuáis Ioosíng their jobs at
a time when the región has not yet fully developed systems of social protection.
It should not be surprising under these circumstances that there is a malaise in the región. While
average levels of real per capita incomes haverisenin a majority of LAC countries in the 1990s, this may
not yet have been translated into improvements in subjective perceptions of welfare. Table 2 presents the
results of a 1999 opinion survey conducted by the Wall Street Journal in fourteen LAC countries. Almost
two thirds of the respondents believe that their parents had a better life than themselves. Less than half
believe that their children will have a better life than themselves. This pessimism probably reflects both a
feeling of economic insecurity and the relative lack of progress in reducing poverty observed over the last
two decades and documented above.

2
3

The LAC región has performed better in terms of non-monetary indicators of well-being, with improvements for adult illiteracy,
infant mortality, life expectancy, gross secondary school enrollment, and access to safe water.
A convincing body of empirical evidence is lacking on many of these issues. Lustig and Arias (2000) argüe for example that
there have been widening retums to skills, but Gilí (1999) suggests that the empirical evidence is still very much mixed. See for
example Fajnzylber and Maloney (1999.)

Sesión 4 - 3

Table 2
Subjective perceptions on changes in living standards in LAC
Taking everything into consíderation, would you say that
your parents lived better, the same, or worse than how
you live today?
And regarding your children, do you believe that they will
live better, the same, or worse than how you live today?

Better

Same

Worse

NA

61.2%

22.0%

14.4%

2.4%

46.1%

20.7%

22.0%

11.2%

Source: 1999 poli, Wall Street Journal, quoted by Rodrick (1999.) NA = No answer.

There are no easy answers to the difficulties and uncertainties faced by poor households in Latin
America. Macroeconomic policies promoting stable and broad-based economic growth certainly help in
reducing poverty, but they are not enough. This is why govemments in the región have increasingly been
concerned about establishing or strengthening systems of social protection and safety net programs that
help mitígate the potential impact of economic shocks before they occur, and help the poor cope with
these shocks after they have occurred. In this paper, we focus on publicly funded or mandated safety nets
functioning asrisk-copingmechanisms. What are the characteristics of good public safety nets? What
are the main types of safety nets currently in place in LAC countries? Which programs should be
protected or expanded during economic crises? Finally, what has been the actual record in terms of
protecting the poor through targeted public spending during crises? This paper provides tentative answers
to these questions. Section 2 of the paper introduces the reader to the main concepts used in the literature
on social protection and safety nets within the context of the various ways in which households are
affected by, and respond to shocks. Section 3 presents the main types of programs currently in place, with
their respective strengths and limits. Section 4 uses the experience of Argentina and México to assess the
actual record in protecting the poor through safety nets during crises. A brief conclusión follows4.

II.

The goal: Designing good safety nets and social protection systems

A recently drafted World Bank Social Protection Strategy Paper (World Bank, 1999a; see also
Holzmann and Jorgenson, 1999) places social protection in the context of social risk management. It
makes the point of dividing social risk management strategies into three types: risk reduction, risk
mitigation and risk coping. Under risk reduction fall macro economic management, regulations, and
institutional development policies that help prevent crises from occurring. While these prevention policies
are important, they are too general to be part of the social protection system which consists mainly of
mitigation and coping strategies. Risk mitigation strategies are developed before a shock to reduce the
impact of the shock once it occurs. They include for example income diversification and insurance
mechanisms, both formal and informal. Risk coping strategies are implemented after a shock to deal with
the impacts of the shock not covered by risk mitigation policies. Table 3 from Gilí (1999) provides an
overview of some of the main govemment and prívate sector policies.

4

Several authors have recently reviewed t e literature o social protection, safety nets, and crises. This includes a o g others
h
n
mn
Klugman (1999) for safety nets, Dar and Tzannatos (1999) for active labor market programs, Kami (1999) for u e po m n
nm ly e t
insurance, and Ezemenari and Subbarao (1999) for social assisatnce. The Inter-American Development Bank (in February 1999)
and the World Bank (in J n 1999) have held conferences o t e topics.
ue
n h
Sesión 4 - 3

Table 3
Classifying Government and Private Risk Management

Nature
of Policy
Mainly Prevention
(risk-reducing)
- regulatory
- or taxpayer funded
- universal
Mainly Insurance
(risk-mitigating)
- inter-temporal transfers
- premium funded
- non-poor and poor
Mainly Assistance
(risk-coping)
- within period transfers
- taxpayer funded
- targeted: focus on poor

Government
Led
Macroeconomic policies
Financial regulations
Infrastructure investments
Labor-related regulations
Human capital investments
Unemployment insurance
Severance funds
Job protection statutes
Public work guarantees
Public work programs
Means-tested cash transfers
Conditional cash transfers
Commodity transfers

Measures

Private
sector
Led
- Infrastructure investments
- Human capital investments
- Portfolio diversification
- Individual savings
- Sale of assets (e.g. land)
- Labor forcé participaron
- Inter-household transfers
- Community solidarity
- Support from NGOs
- Public-private partnerships

Source: Gilí (1999.)

In this paper, we focus on publicly provided or mandated instruments for assisting private
individuáis in coping with shocks (the bottom cell in the middle column in Table 3.) This does not mean
that we do not take into account private coping strategies5. There are of course linkages between private
and public coping strategies (see e.g. Ezemenari, 1997, for transfers.) As mentioned in the introduction,
some private short-term strategies may have permanent effects that make it difficult to reduce poverty in
the longer term. This is the case when the nutrition of certain family members suffers. It is also the case
when children are put to work. Although the substitution effects between work and schooling are likely to
be partial due to the possibility for the parents to reduce the time devoted by the children to leisure
(Ravallion and Wodon, 2000a), it has been estimated that on average, for six Latin American countries,
the reduction in the probability of going to school when a child is performing paid work varíes from 21 to
67 percent depending on the sample (Wodon and Siaens, 2000; the estimates include controls for a wide
range of other variables affecting the decisión to go to school and/or to work, so that they are net marginal
effects.) This substitution between work and schooling reduces the human capital endowment of working
children to the extent that working children may expect on average a loss of about 7 percent of their
discounted life-time earnings when they are put to work even after taking into account their positive
earnings when working as children and the higher level of experience accumulated because of work at an
early age. Clearly, the magnitude of the long term losses due to child labor calis for the design of
programs that help parents keep their children in school, especially during economic crises.
More generally, publicly funded or mandated social protection programs and safety nets represent
an attempt to protect the poor against therisksthat arise from shocks, whether the shocks are foreign or
domestically induced, and whether they are covariant or idiosyncratic. As already mentioned, an economic
crisis produces a covariant shock, whereby many people are affected at the same time. But even in
5

To cope with the income losses induced by shocks, the poor adopt a wide variety of coping strategies. These may include
movingfrom formal to informal sector employment; working longer hours and/or working at a second job; promoting the labor
forcé participation of additional family members such as spouses and children; selling (or consuming in the case of farmers)
productive and other assets, including stocks; migrating temporarily or permanently in order to search for employment
opportunities; reducing consumption pattems, including restricting the food intake of family members, taking children out to
school to reduce education expenditures, or postponing health care expenditures; relocating and/or restructuring households, for
example by having several families living under one roof; drawing on outside help both in kind and in cash, including support
from local communities, friends and relatives, and private institutions such as NGOs (this in turn highlights the role of social
capital.) One interesting paper about the gender dimensión of these strategies is Cunningham (1998.)

Sesión 4 - 3

normal times, households can be affected by idiosyncratic shocks such as a death, an illness, or a loss of
employment. One should differentiate between the social safety nets which should be in place at all times
to deal with idiosyncratic shocks, and the programs which are specifically designed to help large numbers
of poor people suffering from temporary adverse covariant shocks (World Bank, 1999a.)
There are typically more mechanisms available for coping with idiosyncratic as opposed to
covariant shocks, and thefiscal implications of both types of shocks are clearly different. However, even
among the programs designed to deal with covariant shocks, one can identify a number of alternatives.
These include:
•

Emergency employment programs involving public works, often using labor intensive methods,
commonly called workfare;

•

Social funds, which establish special programs, usually in rural areas, for financing small scale
public works identified by local community groups;

•

Nutrition and food interventions, particularly those targeted at vulnerable groups such as
children and pregnant women. These may take many forms, including food distributions, food
stamps, and food served in schools or community kitchens;

•

Systems of direct cash grants targeted to the poorest, which may be conditioned on favorable
behavior (such as school attendance and/or health center visits);

•

Other instruments, such as pensions and unemployment insurance, including systems of
mandatory severance payments upon terxnination.

Another way to organize the discussion is to look at the programs in terms of the age groups they
serve, on the basis of the fact that the different age groups have different needs.6 People in the youngest
age groups generally are at greater nutritional risk than other groups, while for adults, the principal
problem may be one of employment. For the elderly, the critical problems are maintaining sufficient
income to meet basic needs, and to have adequate health care access.
Within Latin America and the Caribbean, at present almost all countries have some mix of the
above programs in varying degrees. However, very few if any programs completely fulfill the criteria of
an ideal safety net which should have the following characteristics7:
•
•

It should be well targeted to the poor, with clear eligibility and termination rules, so that access
is simple and predictable;

•

It should be supervised by well functioning institutions already in place;

•

It should be counter-cyclical (i.e. receive more funding when there is an economic crisis), and
in some cases implemented automatically according to pre-agreed triggers such as ariseabove
some level in unemployment or poverty;

•

It should befiscally sustainable;

•

6

It should provide sufficient coverage of the population to be reached, particularly the most
vulnerable and excluded groups;

•

7

It should be based on a sound analysis of who is likely to be affected the most by crises, and
what kinds of coping mechanisms are normally used by those affected;

It should be able to provide benefits quickly, with as large as possible a share of the costs
resulting in net increases in incomes;

This idea comesfrom Ana-María Amagada, See Appendix table for details.
For a fuller discussion of criteria see Grosh, 1995.
Sesión 4 - 3 6

•

It should complement, not substitute for, prívate safety net programs and other social protection
mechanisms;

•

It should be scaled back when the crisis is over.

III.

The tools: Types of safety nets and social protection programs
A.

Workfare programs

Workfare programs provide employment through specifically designed public works projects. The classic
example is Trabajar in Argentina. In this program, projects are identified by local governments, NGOs
and community groups, and can provide employment for no more than 100 days per participant. Project
proposals are reviewed by a regional committee, and projects with higher poverty and employment
impacts are favored8. Workers hired by the project are paid by the Government, specifically the Ministry
of Labor. The other costs arefinanced by local authorities. Example of eligible projects include the
construction or repair of schools, health facilities, basic sanitation facilities, small roads and bridges,
community kitchens and centers, and small dams and canals. These activities are fairly similar to those
financed by social funds (see below.) One of the differences between a social fund project and a workfare
project is that the workfare project is likely to be supervised by local authorities, rather than by
independent agencies, and construction is typically not contracted to the prívate sector, but is carried out
by the sponsoring agency, which can include local and provincial governments, prívate groups, and
national organizations. Another difference is that workfare programs have the generation of employment
and income as their priority, while social funds focus more on the quality of the infrastructure produced.
The projects are limited to poor areas as identified by a poverty map. Moreover, wages are set to be
no higher than 90 percent of the prevailing market wage, so that the workers have an incentive to return to
prívate sector jobs when these are available. Thus, the program involves self targeting apart from
geographic targeting. Overall, targeting of the poor under Trabajar II (the second round of the project) has
been reported to be quite good, with 75 percent of the funds reaching the bottom 20 percent of the income
distribution, and 40 percent reaching the bottom 5 percent. However, the supply of jobs in the program
depends on budgetary allocations as well as the ability of local communities to identify viable projects. As
good as it is, Trabajar has provided employment to no more than 1 or 2 percent of the labor forcé, at a
time when unemployment has ranged from 13 to 18 percent of the labor forcé.
Large workfare program have also been implemented by the Government of Chile during the period
1975-88. The objective of these programs was to absorb workers displaced from the public sector, and to
reduce unemployment during the adjustment period. As for Trabajar, these programs provided
employment in emergency public works, including maintenance and repairs to roads and schools,
construction of parks, forestry projects, etc. The programs were administered by municipalities, and were
gradually built up to a peak in 1983, when they employed about 13 percent of the total work forcé (over
500,000 workers), while the unemployment rate was at 17 percent. The programs were gradually reduced
as prívate sector employment increased, andfinally phased out completely in 1988. The two largest
8

In a recent reform of Trabajar, several steps were taken to improve the performance of the program. The focus of the reform
was placed on increasing community participation and funding in the choice of the projects to befinanced. Trabajar now works
in collaboration with local community groups, NGOs, and municipalities who present projects for selection. Projects must first
be approved for technical feasibility. Next, they are selected on a points basis. More points are awarded to projects located in
poorer areas, yielding larger public benefits, benefiting from well-regarded sponsoring community groups or NGOs, and
reducing labor costs below the mínimum wage. These new features have improved targeting both at the geographic and
individual levels. The involvement of local groups has also improved the quality of monitoring and feedback for the projects.

Sesión 4 - 3

programs offered manual labor at very low wages to ensure self targeting. Most of the workers were
unskilled, and they received one-fourth of the then current minimum wage (about one-half of the market
wage.) One out of every four participant was a women. However, while the program was considered
successful in terms of reducing poverty and the social impact of unemployment, the quality of the public
works produced was notably low, particularly as the program expanded in size. In addition, it is not clear
that all the workers in the program would have been unemployed without the program. An evaluation
found that 32 percent of the participants had no work experience prior to participating in the program, and
46 percent had retired voluntarily before joining the program. Many of those with no experience were
women who went to work for thefirst time (Universidad de Chile, 1992.)
The advantages of workfare programs include their ability to expand quickly during a crisis, once
the basic mechanisms have been established, and to reach the poor through area targeting and, within poor
areas, through self targeting thanks to the low wages. But a problem with these programs is that the cost of
generating one dollar in additional income for the poor through public works is typically large, in the
range of three dollars or more. To understand why, the measure of cost effectiveness and its
decomposition proposed in Box 1 are useful.

Sesión 4 - 3

BOX 1
MEASURING THE COST-EFFECTIVENESS OF PUBLIC WORKS
Following Ravallion (1999), assume that without public works, an individual has a probability F* to find
employment at market wage W*. Expected earnings are F*W*. With public works, the individual earns
the public works wage W. If the individual can continué to search for private or self-employment while
participating in public works, with probability F of finding such employment, the expected wage with public
works is FW*+(1-F)W. The net wage benefit from the program for the worker is NWB = (1-F)W - (F* F)W*. If the worker gets unemployment benefits or a subsistence allowance S, the wage benefit is
reduced to NWB = (1 -F)W - (F* - F)W* - (1 -F*)S. If the program costs G to the Government per worker
employed, a measure of cost effectiveness is the share of public expenditures transferred to workers as
wage gain NWB/G. This measure can be decomposed as follows:

NWB
G
/

C(W + L)
~G
|

C
\

budget wage

W

NWB

(W + L) W
\

targeting

leverage share performance

proportionate
wage gain

The determinants of cost-effectiveness are a) the leverage ratio C/G, where C is the total cost per
worker including community funding; b) the wage share (W+L)/C, where W stands for wages paid to the
poor and L stands for leakage due to wages paid for the non-poor; c) the targeting performance W/(W+L)
which is the percentage of wages reaching the poor; and d) the proportionate wage gain NWB/W. This
model can be extended to take into account the benefits of the infrastructure built by public works, but
these benefits are not as immediate. [See the text for a discussion of the limits of this measure estímate of
cost effectiveness.]
The measure of cost effectiveness used is the share of total program costs which reaches the poor
through net increases in earnings. This share is a function of four parameters: the proportionate wage
gain, the targeting performance, the wage share, and the budget leverage. A reasonable valué for the
proportionate wage gain may be 0.5 because the workfare wages are low and the poor typicallyfind some
other way to generate resources, for example through part-time informal employment when they do not
have access to the programs. Because of the self-selection involved and the priorities given to poor areas,
targeting performance may be good, at about 0.8. The wage share can often be obtained from
administrative records by multiplying the number of work days created by the program by the wage rate,
and dividing this amount by the total cost of the program. In many cases, the wage share will not exceed
0.7. Finally, when the program is almost entirelyfinanced by the federal state (even though project
selection may be done at the local level), the budget leverage is equal to one (in the case of Trabajar, there
is budget leverage, but while this saves money for the central government, it still has to be paid by local
governments.) The measure of cost effectiveness is obtained by multiplying the various parameters9. It
thus typically costs three or more dollars to the national or federal government to transfer one dollar to the
poor in additional wages.
The notion that it costs three or more dollars to transfer one dollar of income to the poor through
workfare could be challenged, in that the benefits could be higher for two reasons10:
In our illustrative examples, this measure would be equal to 0.5*0.8*0.7=0.28, in which case the total cost of generating one
dollar in n t additional wage earnings for program participants is 1/0.28=3.6 dollars. For a example of the econometric
e
n
m t o s t a c n be used t m a u e with s m precisión t e net wage benefit of workfare programs using household surveys
eh d h t a
o es r
o e
h
(i.e. t e parameter NWB/W), see Jalan a d Ravallion (1998.)
h
n
10 For a fuller discussion of t e e points, see Maloney (2000.)
hs

9

Sesión 4 - 3

•

First, the method presented in Box 1 does not take into account the benefits of the public works
themselves, which can be substantial if the workers are put to good use. The problem, however,
is that these benefits will be enjoyed during the whole life of the infrastructure built, while what
the poor need in times of crises is immediate income support. If the poor have high discount
rates (which they do in general, but especially in times of crisis when their resources do not
provide for basic subsistence), the discounted valué of the benefits generated by the public
works may be quite low. Moreover, since the emphasis is on job creation rather than
investments, there may be a bias toward "make work" or prestige projects that may not be
highly valuable. This may be particularly true in a crisis, when a rapid expansión of the
program exhausts the backlog of viable projects.

•

Second, the method presented in Box 1 assumes that only the net proportionate wage gain must
be taken into account for measuring the program's impact. But in periods of high
unemployment, it could be argued that at least part of the difference between the public works
wage and what the program participant would have earned by himself without the program will
be available as eamings for another worker who does not particípate in the program and who is
also underemployed. At the extreme, the whole wage rate could be taken into account in the
cost-benefit analysis, which would greatly enhance the cost-effectiveness of such programs.

On the other hand, arguments could also be put forward to argüe that the net transfers to the poor
are lower than predicted by the decomposition in Box 5. For example:
•

First, since workers are paid by local authorities, the opportunities for corruption and political
bias are more pronounced. With Trabajar, there remains some evidence of political influences
in the choice of participants and gender discrimination (few women are selected in some areas.)

*

Second, the poorest communities may not always be well positioned to submit proposals for
projects and/or to contribute to non-wage costs. In this case, the targeting performance of the
program may suffer, because the contribution of geographic targeting to overall targeting
performance will be reduced.

Apart from Argentina and Chile, the experience of México is also valuable for the assessment of the
strengths and limits of workfare programs. Mexico's rural areas the Programa de Empleo Temporal (PET)
provides off-season temporary employment below minimum wage through public works. Employment is
for up to 88 working days at 90 percent of the minimum wage. In 1999, 93 million work days and one
million jobs were to be created. PET is an example of a program which functions in normal times rather
than only during crisis, although it is restricted to certain periods of the year only. As expected, the
projects are labor intensive. Examples include irrigating land, paving roads, clearing land, improving
housing, and installing water and sewerage systems. The data suggests good targeting, with the
participants being poor and needing the temporary jobs more than non-participants because they do not
benefit from occupations that keep them employed all year long. Yet, the program does not reach the
smallest (and probably poorest) rural communities. On average, PET communities are almost twice as
large as non-PET communities. PET communities have better access than non-PET communities to
electricity (74 versus 60 percent), public phones (33 versus 19 percent), pre-schools (81 versus 67
percent), primary schools (89 versus 82 percent), and tele-secondary schools (22 versus 11 percent.) Part
of the problem may be due to the higher (e.g., administrative) cost of reaching very small rural
communities.
One might also consider job training programs as a safety net, particularly if they can be modified
during a time of crisis. In México, the Probecat program was implemented in 1986 as a response to the
growth in unemployment that followed the 1982 debt crisis and the subsequent structural adjustment
policies. Today, the program provides training to cióse to 500,000 beneficiaries per year in urban areas. A
new evaluation of the program suggests however that it does not have a statistically significant impact on
employment and wages (Wodon and Minowa, 1999; see Revenga et al., 1984 and STPS, 1998, for

Sesión 4 - 3

previous evaluations.) These disappointing results are not that surprisíng, because most retraining
programs in OECD countries have been found to have limited impacts. One reason for this may be that the
training is provided for too short a period of time (a few months) in order to provide skills valuable in the
long run. Some job training programs may in fact function as safety nets by providing temporary relief for
the unemployed with a self-targeting mechanism not unlike that of public works programs since
participants typically receive only the minimum wage. It is probably better to choose one goal or the other
(training versus social protection), rather than trying to meat both goals with a single program.

B.

Social investment funds

Social investment funds (SIFs) were the original World Bank response to the social aspects of adjustment
programs, and some of the earliest funds (e.g., Bolivia's Emergency Social Fund created in 1991) were
designed primarily to provide employment (Jorgensen, Grosh and Shacter, 1992.) In fact, SIFs were
started in part to avoid the problems associated with emergency public works (workfare) programs. Yet
almost all SIFs now have evolved into programs designed to provide small scale social and infrastructure,
particularly in rural and poor areas, using projects generated and executed at the local level. Therefore,
social funds are not safety nets per se. Note that unlike workfare programs, some social funds also finance
programs that do not involve construction or maintenance, such as nutrition programs, technical assistance
and micro credit. When construction work is involved, it is not rare to see social funds using a skilled
manpower paid at market wages. This is because the quality of the infrastructure built is considered as
more important than the provisión of employment for the poor. This is the case with the Honduras social
fund which originated from the transformation of an employment generation program in the early 1990s
but does not consider the objective of employment creation as its main priority nowadays. On the other
hand, most social funds are usually targeted to poor areas through the use of a poverty map (or, in some
cases, through the use of a map of non-satisfied and non-monetary basic needs.)
While both workfare and social fund programs build projects in the public sector, there are
important differences. The social fund finances the material and labor costs of a project, although some
local labor may be donated as a community contribution. This varíes across SIFs and countries. A
workfare program generallyfinances the labor cost of a project at the national or federal level, and asks
that local govemments or agencies provide for the material costs. Thus, there is a clear incentive in
workfare programs for the local agency tofind labor intensive methods of construction, and choose labor
intensive projects. Since SIF projects are bid out to the prívate sector, often the most modern and capital
intensive construction methods are used, although in some cases social funds specify minimum
employment levels to be attained in their operations.
Most social funds are agencies independent of line Ministries, often attached to the office of the
country's President, which reviews and funds projects submitted by NGOs, local govemments and other
sponsoring agents. Their strong points include local community involvement and the ability to respond to
local perceptions of needs, especially in rural areas where normal govemment expenditures often do not
reach the poor. Social funds also have a better ability to avoid corruption and "make work" projects. But
they are not very good at providing safety nets, and they do not normally expand during a crisis to provide
more employment. In fact, the amount of employment and income generation provided by social funds has
historically been low. For instance, a review of social funds found that for 10 major social funds provided
employment on average equal to only 4 percent of the labor forcé (Goodman et al., 1997.) Likewise, the
monetary contribution of wages was judged to be small, as well as the poverty reduction impact coming
from the projects themselves. Social funds are better at improving the supply of health, education, and
basic infrastructure services, with in some cases impacts on outcomes such as school enrollment rates,

Sesión 4 - 3

age-for-grade, or the incidence of illnesses (recent evaluations using household level data include Pradhan
et. al, 1998, for Bolivia, and ESA Consultores, 1999, for Honduras).
Still, one of the clear advantages of social funds is that they have strong organizations that exist
with relatively good systems for project management and monitoring. These organizations can be used in
times of crisis for the delivery of social safety nets. The existing social fund in Honduras, for instance, has
proven highly valuable in directing emergency assistance to local villages after Hurricane Mitch. One
possibility therefore is to work with social investment funds to modify their operations during a crisis,
such as by putting more emphasis on labor intensive projects, and by having the fund involved in new,
hard hit areas. It is thus a good idea to identify labor intensive projects in advance of a potential crisis, so
that these are ready for funding should a crisis come about. As was the case with Mexico's Probecat job
training program, social funds can be a valuable chaimel for special assistance during economic shocks or
after natural disasters.

C.

Nutrition and food programs

Nutrition and food programs take a variety of forms. Subbarao et al. (1997) have identified about thirty
countries using food policies with redistributive aims. Among these countries, price subsidies are used as
often as feeding programs and food for work requirements, and much more often than food quantity
rationing and food stamps. In fact, many workfare programs now providing wages in cash initially started
as "food-for-work" programs. As for direct feeding programs, they provide food to needy recipients,
through direct delivery of unprepared foods from a program warehouse, delivery of prepared food from a
community kitchen, or the provisión of a lunch or breakfast to children in school11. Evaluations of
nutrition programs generally indícate that there is only a small, marginal improvement in nutrition
compared to the case where the family receives an equivalent cash grant. This is because families may
substitute free food for their own purchases, and largely use the savings for other purposes. However, even
if a food program is roughly equivalent in effect to a cash grant, there are fewer possibilities of diversión
of funds, since food is less likely to be misappropriated than cash. In addition, food is more likely to go to
women, and be used to improve the welfare of the family, while cash is more likely to be used by men for
lower priority activities. An intermedíate alternative between cash and food are food stamps. Food stamps
have the added benefit of not requiring a complicated system of storage and transport of food, while
making use of the already existing prívate food distribution network. In Honduras, the food stamp
program used to also cover medicines and school books. In Jamaica, the food stamp program was
introduced in place of general food subsidies, and has proven effective in raising the incomes of the poor
(Grosh, 1992.)
A common way of targeting food programs is by linking distribution to a health program,
particularly maternal and child health care. In this way, the food serves as an incentive to attend the
program, and nutrition education can help improve the use of the food given out. As already mentioned,
giving food to women also lessens the possibilities of it being diverted for sale in the market. Food
programs also can be designed to be self targeted, if the food products given out are those consumed by
the poor and not by the middle class. Alternatively, food distribution can take place at centers located in
poor neighborhoods. Such approaches can reduce the administrative burden of targeting programs, but
they increase the possibility of leakage to the non-poor. School lunch and breakfast programs are also
difficult to target at the individual level if one is to avoid stigmatizing some students within the school as
being "poor." In many countries, schools from poor areas are targeted, but completely untargeted national
programs are also common. One benefit of these programs is the incentive given to keep the children in
school in order to have them fed, and to improve their learning abilities while in school (see Wodon and
11

School feeding programs are especially popular in Latin America (see, e.g., Phillips et al., 1995 on Honduras, Dall'Acqua,
1991, on Brazil, and Jacoby, Cueto and Politt, 1996, on Perú).

Sesión 4 - 392

Siaens, 1999a, for an evaluation of the Mexican school breakfast program.) From a nutrition point of
view, the prime beneficiaries of the programs can however be other members of the family, if the children
are not fed at home because the parents know that they will receive a school lunch.
Food subsidies are another way to help the poor by reducing the cost of their consumption bundle
(Besley and Kanbur, 1988.) Means-tested food subsidies tend to be more effective than other subsidies in
reducing inequality and improving welfare. For México for example, Wodon and Siaens (1999b) suggest
that universal subsidies do not perform well for the reduction of inequality and the improvement of
welfare. They compare three programs: the now defunct universal subsidy on Tortilla, a program
providing one kilo of free tortilla for households with income below two minimum wages, and a program
of subsidized milk, also means-tested. Their main results were as follows:
•

Food subsidies are better than non-food subsidies. Subsidies for basic consumption goods such as
tortilla reduced inequality, especially in urban areas, and more so than subsidies for Utilities such as
water and electricity. However, food subsidies generate price distortions and they are costly.
Furthermore, a universal subsidy on tortilla is less effective than would be a similar generalized
subsidy for the ingredients needed for making tortilla, such as corn flour.

•

Within food subsidies, means-tested subsidies are better than universal subsidies. The marginal impact
on inequality and welfare achieved with the universal tortilla subsidies does not come cióse to the
welfare gains achieved with the means-tested subsidies.

Still, food subsidies may not represent a sound and cost-effective investment for poverty reduction.
Food subsidies and distribution systems can have negative incentive effects on the supply of labor (Sahn
and Alderman, 1995.) They can be badly targeted, with high leakage among the non-poor (Grosh, 1994;
Comia and Stewart, 1995.) And, while self-targeting can be achieved to some extent by subsidizing goods
consumed in larger quantities by the poor than by the non-poor, this is no panacea (Tuck and Lindert,
1996.) There has been therefore a tendency to reduce funding for food subsidies in order to fund other
programs, for example in México (Levy and Davila, 1998.)

D.

Conditional cash transfers

Since food is fungible with money, one could argüe that cash grants are the simplest and most direct
way of providing safety net assistance. The use of cash grant also avoids the utility losses associated with
in-kind support. In developed countries, and some advanced Latin American countries, cash payments to
selected households are slowly becoming more common. Unconditional cash payments and similar forms
of social assistance are usually targeted to women with dependant children, the disabled, the aged, and
those unable to work. However, the problems of targeting and controlling cash payments makes this
approach problematic in the poorer countries that lack good administrative arrangement. For instance, in
Bolivia, the Bonosol program provided once a year cash grant equal to four weeks pay at the minimum
wage to adult citizens. The attractiveness of the grant and the lack of screening or identification
mechanisms resulted in widespread abuses, including double payments and payments to non-Bolivians.
An attractive alternative is to link cash grants with school attendance or other desirable behavior.
This has been introduced in various degrees in such countries as Brazil (Bolsa Escola), Argentina (Beca
Secundaria), México (Progressa), and Honduras (PRAF), among others12. These programs are not safety
nets properly speaking, or at least, they were not originally designed to function as compensatory safety
A rigorous evaluation of Mexico's Progresa is being prepared by the Progresa staff with supportfrom the International Food
Policy Research Institute. Preliminary results are available in Progresa (1999.) The International Food Policy Research
Institute is also going to help for the evaluation of Honduras's PRAF, with a design combining demand and supply-side
interventions.

12

Sesión 4 - 39

nets during crises. Yet the programs do provide valuable benefits which households can rely upon during
crises, and these benefits can be increased during a recession if need be. In other words, as was the case
for job training programs and social investment funds, existing programs providing conditional cash
transfers can be expanded and modified to serve as safety nets during a crisis. In general however, school
related grants will offer only a partial response to crisis situations, if only because the programs are
targeted to families with children already in school, so that some of the poorest who cannot afford to send
their children to school are excluded from the programs' coverage.
School-based conditional cash transfers programs reduce the opportunity costs for poor parents of
keeping their children in school. This opportunity cost is essentially the loss in child wages or in the valué
for the parents of the domestic work done by the children which cannot be enjoyed when the children go
to school. In many cases, this opportunity cost of schooling is difficult to estímate, and it is not obvious
that the grants must be equal to the opportunity cost for the parents to send their children to school
(Ravallion and Wodon, 2000a.) Indeed, it is reasonable to think that the parents have an intrinsic interest
in having their children go to school, either for altruistic motives, or for the future benefits that
intergenerational transfers provide once the children reach adulthood. In some countries however, the
level of the conditional grants appears to be high. For Progresa in México for example, to justify the
relatively high level of the Progresa grants, it has been argued that apart from providing incentives to
accumulate human capital, they also improve the families' overall quality of life. Yet there may be more
cost effective ways to improve the quality of life of the program's beneficiaries. But more work would be
needed to measure the trade-offs.
At what level of schooling should the grants be provided? This will depend on the characteristics
of the country. In Brazil and Argentina, the programs focus on secondary school, since these are the
children that are more likely to be pulled out of school during a crisis. In Honduras, the program focuses
on the last few years of primary school. In México, the program covers the end of primary schooling, and
the lower secondary school cycle. In Venezuela, the program covers primary school children. In some
cases, these programs are tied not only to attendance, but also to school performance, including passing on
to the next grade. While this may provide valuable incentives, one has to make sure that such conditions
do not exelude the poorest which may have more difficulties in succeeding at school.
How should the grants be targeted? The experience Progresa is interesting (Skoufias et al., 1999.)
The program uses a three stage targeting mechanism. First, poor rural localities are selected for
participation. Next, poor families are selected within participating communities using a multivariate
discriminant analysis. Third, local communities may review the Progresa's selection staff and reclassify
poor families as non-poor and vice versa. This targeting mechanism is basically sound, and the results
appear to be good. One concern is that the level of community involvement remains marginal. The
targeting process is centralized in part due to the desire to avoid political interference in the choice of
beneficiaries. Nevertheless, more efforts could be put to promote the role of communities in targeting.
Another related question relates to the need for targeting within poor communities. The higher the
proportion of the poor in a community, the less the need to target within that community, especially if
targeting is costly not so much administratively, but rather in terms of social cohesion (those who do not
get the program may envy those who benefit from it.) In Honduras, where the PRAF program is being
modified in part on the basis of Progresa's experience, it has been decided to provide support to all the
families residing in the poor communities that particípate13.
Conditional cash transfer programs can also be used to promote good health practices, including the
consultation of local health providers. This is again the case with México's Progresa where eligible
families receive a transfer (both in cash and in kind through a nutritional supplement) for health purposes
apart from the transfers related to schooling. That is, Progresa aims at providing a coordinated
It is important when feasible to use distributional weights in the evaluation of the targeting of social programs and other
interventions. For such an analysis, see for example Ravallion and Wodon (2000b.)

13

Sesión 4 - 394

intervention for education, health, and nutrition, with the hope that the impact of the whole program will
be larger than that of its individual parts. Of course, with Progresa as with the other programs, in
maximizing the impact of the school- and/or health-based interventions, it is important to take care of
supply side issues. For example, Progresa has been successful in raising school enrollment and attendance
at health care centers, but this has led to tensions on the supply side. Steps have been taken to coordinate
Progresa's action with that of other Ministries, such as the Ministries of Education and Health, but more
may be needed to optimize demand and supply side interventions.

E.

Other programs

There are several other types of social protection programs that exist, including pensions for the
elderly. Because many of the poor belong to the informal sector, they do not have access to the pensions
provided by social security systems (similarly, those in the informal sector often lack access to other state
organized benefits such as low income housing.) A discussion of pensión issues for the poor is provided
in Holzmann et al. (1999.)
Unemployment insurance is common to Europe and North America, but relatively rare in Latin
America, in part because of its high cost. On the other hand, the current labor legislation in many LAC
countries mandates a severance payment on termination that is a function of the number of years worked.
Normally, this is about one month's salary for every year of service, up to some máximum. This may be
supplemented by an unemployment insurance scheme (as in Argentina, Brazil for formal sector workers)
in which monthly contributions from payrolls and/or the employer entities the worker to a monthly
payment over a limited time horizon, with the payment set low enough to reduce disincentive effects. But
since the formal sector in many countries is less than half of the total work forcé, the coverage of
unemployment insurance/severance payments is far from complete, and the exclusión of the informal and
rural sectors means that these mechanisms miss those areas containing many of the poor. Moving from a
severance pay system to an unemployment insurance system could be beneficial if it reduces labor costs.
However, unemployment insurance can also create moral hazard problems by subsidizing unemployment
(as compared to workfare programs, which subsidize employment.) One important element here is to
ensure that unemployment benefits are not so generous as to discourage job search. For this, both the
level of the payments and the length of time workers can receive benefits must be monitored. There is a
large literature on these issues, but it cannot be reviewed here.

IV.

The constraints: Fiscal space and administrad ve capacity
A.

Are social protection programs counter-cyclical?

Good social protection programs should expand during an economic crisis as unemployment grows
and income levels decline. In other words, the programs should be counter-cyclical in nature, i.e. rising
when the economy falls. In some countries, the system of safety nets is supported by automatic
entitlements that forcé the public sector to spend more on those who need it. For instance, unemployment
causes an increase in the numbers eligible for unemployment benefits, and declining incomes can increase
the number of households eligible to receive cash or food assistance. But in most Latin American
countries, there are no such automatic triggers. This is in part because the need to have counter-cyclical
social protection programs can conflict with the need to impose fiscal austerity during a recession. An

Sesión 4 - 395

economic crisis, whether domestically or externally induced, leads to a drop in output, and consequently
in a fall in government revenues. But during a crisis, governments are reluctant to raise taxes. Thus
maintaining high expenditure levels can then lead to large budget déficits.
The extent to which LAC countries have provided counter-cyclical social protection is an empirical
question which has received surprisingly little attention. One problem is that the data on budget
expenditures normally do not typically identify social protection or safety nets as a separate activity.
Social protection programs are scattered in various sectors such as health, education, social security, and
welfare. However, in some countries there is data available. This is the case for Argentina and México on
which this sub-section is based.
Consider the data presented in Table 4. Both Argentina and México suffered an adverse shock in
1995 (México wasfirst, Argentina followed). Per capita GDP in Argentina decreased by 5.32 percent
between 1994 and 1995, while it decreased by 4.93 percent in México between 1994 and 1996. The share
of GDP devoted to targeted social spending decreased a bit in the two countries. The poverty rate
increased in the two countries, leading to an increase in the total number of poor people. The targeted
spending per poor person decreased much more than per capita GDP, yielding an elasticity to growth (in
this case to a recession) of targeted spending per poor person of aboutfive in the two countries. During
this recession, spending for the poor was thus highly pro-cyclical, while ideally it should have been
counter-cyclical in order to protect the poor from the adverse macroeconomic shock.

Sesión 4 - 396

Table 4
Targeted public spending per poor person, 1994-1996, Argentina and México

per capita
Share of Poverty
GDP
targeted
social (%)
spending
in
(1994
= 100)
GDP (%)

Real

Argentina
1994
1995
% change
México
1994
1996
% change

rate
Number
of Targeted
poor people spending
per
poor person
(million)
(1994
= 100)

100
94.68
-5.32%

1.24
1.21

21.6
27.2

7.5
9.6

100
63.12
-27.88%

100
95.07
-4.93%

1.36
1.23

46.95
60.93

42.04
56.51

100
67.30
-23.70%

Source: Wodon and Hicks (1999.)

A number of factors compound to work against the ideal of counter-cyclical safety nets funding
during a recession. First, as noted by critics of structural adjustment mechanisms, the share of GDP
devoted to public spending tends to decrease in order for fiscal restraint to restore macroeconomic
fundamentáis. Second, GDP itself is by definition reduced during a crisis, so that even if the share of
GDP devoted to public spending remains constant, there will still be less resources available for the poor.
These two factors tend to make aggregate targeted public spending for the poor pro-cyclical rather than
counter-cyclical. Third, poverty increases during a crisis, so that the available aggregate resources
targeted to the poor have to be distributed among a larger pool of applicants, yielding lower spending per
poor person.
The pro-cyclicality of safety nets in Argentina and México is also observed over longer periods of
time (Wodon and Hicks, 1999; see also Ravallion, 2000, for Argentina.) Table 5 provides estimates of the
elasticities of various categories of spending to GDP growth. Social spending is not only pro-cyclical, i.e.
falling when the economy contracts, but the elasticities are often higher for social protection programs
than for the social sector as a whole or total govemment spending. Henee, the evidence suggests that
social protection programs, instead of expanding quickly in a crisis, actually are contracting more rapidly
than other types of spending. For instance, in Argentina, the elásticity of total spending with respect to
GDP growth has been 0.94 over the period 1980-1997 (see Table 5), while the corresponding elasticities
for social spending and social protection (i.e. programs targeted to the poor) were 1.55 and 1.63.
Table 5
Estimated Elasticities of Expenditures with Respect to Changes in GDP
Total govemment spending
Social sectors spending
Social protection spending

Argentina

0.94
1.55
1.63

(1980-97)

México

1.88
1.90
2.73

(1987-1997)

Source: Authors' estimation. All estimates are statistically significant at the five percent level. However,
since the data for the two countries are not strictly comparable, it should not be inferred that social
protection spending is more pro-cyclical in México than in Argentina. Note that the Government of
México has made a substantial effort for increasing targeted funding for the poor in recent year, for
example by providing funding for Progresa. This is not yet reflected in the elásticity computed here.

Sesión 4 - 39

The elasticities in Table 5 can also be presented in terms of the responsiveness of the amount of
targeted public spending per poor person to GDP growth (Wodon and Hicks, 1999.) When doing so, it can
be shown that in both countries, a one percentage point decrease in per capita GDP leads to an estimated
three percentage point decrease in targeted public spending per poor person. One third of this impact is
due to the reduction in targeted public spending as a share of GDP that follows from fiscal discipline.
Another third is due to the reduction in per capita GDP itself, which reduces spending even when the share
of targeted spending in GDP remains constant. The last third comes from the increase in poverty due to
the crisis, i.e. public spending must be shared among a larger number of poor people.
The fact that targeted programs for the poor are not well protected may be surprising given that
these programs represent a small proportion of GDP (typically 2 percent or less), and a small proportion of
total spending as well. One reason for the lack of protection for targeted spending may be related to the
lack of bargaining power of the poor. Yet expenditure protection may also be difficult during an
economic and fiscal crisis when the country must honor its debt service and a number of entitlement
programs, with consequently low levels of discretionary resources. Said differently, if 30 percent of the
budget is protected because of its safety net and poverty aspects, and 50 percent is for debt service and
other programs that cannot be cut, a 10 percent reduction in overall spending would produce a 50 percent
cut in the remaining sectors. Given that the remaining sectors might include such sensitive areas as
defense, justice and administration, this may not be a feasible solution, thereby forcing the Government to
abandon the idea of protecting funding for key social protection programs.

B.

How can one select the programs to be protected or expanded during a crisis?

The objective during a crisis should be first to agree that key programs that benefit the poor will not
be cut back even though total government expenditures may be reduced. Next comes the difficult question
of which specific programs should be protected or even expanded. It is tempting to put on the list almost
all programs in basic health, education and nutrition. But not all programs in the social sectors are equally
effective in reducing poverty, and some programs producing long-term benefits for the poor may be
deferrable. In deciding which programs to keep and expand during a crisis, an important criteria should be
the cost-effectiveness of the program in quickly channeling income or its equivalent in kind to the poor.
As suggested in the case of workfare programs, the cost-effectiveness of programs depends on a number
of parameters. Knowing these parameters before hand helps in making a selection. The comparison of
the cost-effectiveness of alternative safety nets in smoothing consumption patterns for the poor via income
generation, in kind commodity distributions or commodity subsidies is one area where additional
analytical work should clearly be done.
One element of a viable strategy - actually, one of the key parameters of the cost effectiveness of
safety nets --, is to protect programs that have good targeting mechanisms for reaching the poor.
Untargeted or universal welfare programs are often fiscally impossible to sustain, particularly during a
crisis. It is true that targeting beneficiaries by complicated means testing, i.e. intensive questioning over
income and wealth, can be costly and not very reliable. This is why many World Bank funded programs
have relied on geographic targeting even though this can lead to substantial leakage. But an alternative
approach is the proxy means test, whereby likely recipients are interviewed and respond to questions
concerning their living conditions, such as the type of housing in which they live, the availability of water,
the types of appliances possessed by the household, etc. Wodon (1997) has shown that these indicators
are powerful for avoiding the two types of errors that can be committed in targeting, namely identifying as
poor a non poor household, and as non poor a poor household. For any given budget, good targeting also
helps for improving coverage among the poor (i.e., what percentage of the poor actually receive program
benefits), which is an important advantage in Latin America where coverage has often been low.

Sesión 4 - 398

Variants of means-testing systems have been used, among others, for the Fischa CAS in Chile,
SISBEN in Colombia, SISFAM in Argentina, and Progresa in México. In practice, people who apply for
assistance are interviewed, and the score on the questionnaire determines eligibility. The weights for the
various indicators included in the questionaire are based on estimated econometric relationships between
the poverty status of a representative sample of households and their indicators. Grosh (1994) has
estimated that the cost of targeting trough proxy-means testing or geographic targeting (poor areas) need
not be much higher than the cost of universal distribution (a reasonable mark-up would be 3 to 8 percent),
while producing substantial benefits. In a sample of untargeted programs, only 33 percent of the benefits
accrued to the lowest 40 percent of the population, whereas targeted programs had 72 percent of their
benefits reaching that population.
Beyond the need to protect the consumption of the poor during a crisis, it is also clear that some
existing health and education programs not necessarily targeted to the poor should be preserved. It is
possible that the demand for these programs might go up, if for example the losses in income due to
unemployment or underemployment cause people to shift from prívate providers and prívate health
insurance to public programs. Still in health, some programs not targeted to the poor should be maintained
because there are large extemalities involved. This would be relevant for disease surveillance,
immunization campaigns, malaria control, AIDS prevention, etc. Preserving access to basic (primary and
lower secondary) education is not a short-term safety net issue. But as discussed earlier, it does reduce the
costs of a crisis over the longer term by avoiding that cuts in education spending for the poor eventually
Iead to a reduction in human capital, and thereby lead to lower productivity.
To protect all the above programs, it will be necessary to identify some programs that can be
reduced, put on hold, or eliminated. For example, some of the worse cases of misspent social spending in
Latin America have come from large subsidized housing programs for the middle class, which are often
mandated through earmarked taxes. While it may be impossible for a government to change these
programs quickly without major legislative or even constitutional authority, in some cases the need to
reallocate spending due to an emergency or a crisis may facilítate fundamental changes. Other candidates
for expenditure reduction may include subsidies such as lifeline rates for Utilities with consumption ceiling
set too high, or price reductions for urban transport in relatively well off areas. Better cost recovery for
higher education (e.g., at the university level) and for some health care services may also help. Table 6
gives a generic ranking based on common practice and observation. While the priorities may change
depending on the country, Table 6 gives an idea of the programs to protect.

Sesión 4 - 399

Table 6
Priorities in protecting programs during crises: A possible hierarchy

Health

Education

Social

Protection/Others

High
Priority

- Immunizations
- Disease control
- AIDS prevention
- Pregnancy care and
early child development

- Means-tested stipends
- School breakfasts in
poor areas
- Special programs for
indigenous populations

- Workfare programs
- Means-tested pensions
for the elderly
- Means-tested nutrition
supplements

Moderate
Priority

- Primary health care
posts and centers
- Basic care in hospitals

- Basic education
(primary and lower
secondary)
- Scholarships for poor
students at universities

Low
Priority

- Advanced curative care
in hospitals

- Free/low cost university
education for non-poor
- research budgets

- Unemployment
insurance
- Social security
- Social funds
- Job training
- Food subsidies
- Subsidized housing
- Subsidies for
infrastructure services

Source: adapted from World Bank (1999b)

C.

Additional implementation issues

There are several additional issues which must be considered when implementing safety nets. First,
establishing new institutions that work effectively at the national level is difficult. Most social funds, for
instance, take 2 to 3 years to get up and running at a level that reaches a substantial number of poor
people. Under time pressure, it is much more easy to work with existing institutions by providing them
with support that enables them to expand their operations during a crisis. It is also often interesting to
expand new safety nets programs that have already started on a pilot basis, and that have been evaluated
and proven effective. As a corollary, it is useful to start new pilot programs every now and then on a pilot
basis, in order to anticipating their strengths and weaknesses and their ability to be expanded in case of
need. Also, as indicated earlier, it may be practical to use existing institutions running programs for job
training, social funds, or transfers to distribute emergency assistance when needed.
Second, not all govemments possess the same capacity to administer programs. In some countries, a
simple program that is not well targeted might be a better choice than a more complicated program.
Safety net programs need to strike a balance between administrative capacity, fiscal sustainability,
political acceptability, the scope of the intervention, and targeting efficiency. An additional problem is the
relationship between the federal and provincial level govemments. The staff of multilateral development
banks have traditionally established a dialogue with federal authorities, but a large part of social spending
takes place at the provincial level. It may be necessary to agree to additional support from the federal
budget to state, provincial, or municipal budgets during a crisis if safety net programs are operated locally.
However, while decentralization has a number of advantages, it is not necessarily pro-poor, so that serious
monitoring of sub-national authorities is needed.
Third, one problem encountered in many countries is the proliferation of small programs throughout
various parts of the govemment. Nutrition programs, for instance, are found in the Ministries of health,
education, social development and agriculture, as well as in autonomous agencies in the office of the
President. Training programs are run by the Ministries of education, Labor, and Economic development.
It is important to attempt some sort of inter-ministerial coordination in order to avoid overlaps and

Sesión 4

duplication. In the longer term, it may be possible to combine or consolidate programs, and to eliminate
the programs that are least effective. As mentioned earlier, a crisis can help in reorienting social budgets.
Fourth, in monitoring expenditure levels, it is in principie preferable to use actual expenditures,
rather than the amounts budgeted. However, actual expenditure levels are difficult to monitor in many
countries because of the long lag involved in receiving the information. By the time it is known that
actual expenditures have fallen, or that the available credits have not been used, it may be too late. To
avoid non-intentional budget cuts due to lags in spending, it is important after budget allocations have
been made to make some efforts in order to ensure that the budgets are set operationally and maintained at
adequate levels, with continuous monitoring of the outlays so that the plans are effectively being carried
out.

V.

Conclusión

Economic shocks may lead to a decline in real income caused by a loss of employment, a situation
of underemployment, or a shift to a less lucrative employment. Real wages may be reduced as well by
high levels of inflation. Changes in relative prices may have negative effects. The poor may also lose
their access to essential public services for health, nutrition and education as a result of reduced
government real spending. They may lose the valué of theirfinancial assets (however meager) if these are
not protected from bankruptcy by thefinancial system or by the public social security system. Or they
may simply have to sell these assets. In all these potential effects and in many others, the poor are more
vulnerable than the non-poor because their income may fall far more quickly below bare subsistence
levels.
It is common practice to recommend that governments protect key social protection programs
during a crisis and expand the social safety net. However, withfiscal resources shrinking at a time when
the number of poor are increasing, it may be extremely difficult to do this. Still, one would hope at least
that high priority programs will be less subject to cuts than other programs. This is feasible because social
safety nets represent a small part of existing budgets. Yet this does not appear to be the case at present.
Rather than being counter-cyclical, safety nets are actually more pro-cyclical than other sources of public
expenditures.
Assuming that there is an agreement to protect the poor from budget cuts during a crisis, the
question becomes: which programs should be protect? There are no easy answers to this question, and the
answer is bound to be country specific. Still, one important criteria for the selection of the programs to
protect or expand is the ability to quickly provide income support or its equivalent in kind for the poor.
Workfare may help, but policy makers should be aware that it typically costs more than three dollars to
generate one additional dollar in net earnings for the poor through these programs. Some nutrition and
food programs may work as well provided that they are well targeted, which is not always the case. Good
targeting is one of the key parameters of the cost effectiveness of safety nets. Also, existing social
programs which are not safety nets per se (e.g., job training programs, social funds, and conditional
stipends for school attendance in poor areas) can prove to be valuable delivery mechanisms for emergency
assistance when there is limited administrative capacity in the country. Apart from compensating the poor
from the likely loss in market-based income taking place during a crisis, it is also necessary for the
government to protect some universal programs. This would include primary health care and education, as
well as health programs with large externalities. But the most difficult task of all is to identify those
programs that can be reduced, put on hold, or eliminated. Candidates may include subsidized housing,
other subsidies for commodities not consumed mainly by the poor, some social security programs, and

Sesión 4 - 4 1

spending for higher education and advanced curative care. While it is often difficult politically to reform
these programs, the need to act during a crisis may facilitate the necessary fundamental changes.

References
Arriagada, A.-M., 1999, Managing Social Risk in Argentina, mimeo, World Bank, Washington, DC.
Skoufias, E., B. Da vis and J. Behrman, 1999, Evaluación del sistema de selección de familias
beneficiarías en Progresa, in Progresa, Mas Oportunidades para las familias pobres, México City.
Basu, K. and P. H. Van, 1998, The economics of child labor,, American Economic Review, 88: 412-427.
Besley, T. and R. Kanbur, 1988, Food Subsidies and Poverty Alleviation, Economic Journal, 98:701-19.
Besley, T., and R. Kanbur, 1993, The Principies of Targeting, in M. Lipton and J. Van der Gaag, ed.,
Including the Poor, World Bank, Washington, DC.
Comia, G. A. and F. Stewart, 1995, Two Errors of Targeting, in D. Van de Walle and K. Nead, eds.,
Public Spending and the Poor: Theory and Evidence, John Hopkins University Press, Baltimore.
Cunningham, W., 1998, Breadwinner versus Care Giver: Labor Forcé Participation and Sectoral Choice
Over the Mexican Business Cycle, mimeo, World Bank, Washington, DC.
Dall'Acqua, F., 1991, Economic adjustment and nutrition policies: Evaluation of a school-lunch
programme in Brazil, Food and Nutrition Bulletin, 13: 202-209.
Dar, A. and I. S. Gilí, 1998, On Evaluating Retraining Programs in OECD Countries, World Bank
Research Observer, 13: 79-101.
Dar, A. and Z. Tzannatos, 1999, Active Labor Market Programs: A review of the Evidence from
Evaluations, Social Protection Discussion Paper No. 9901, World Bank, Washington, DC.
De Janvry, A. and E. Sadoulet, 1999, Growth, Poverty, and Inequality in Latin America: A Causal
Analysis 1970-1994, mimeo, Inter-American Development Bank, Washington, DC.
Deutsch, R., 1998, How Early Childhood Education Can Reduce Inequality: An Overview of Recent
Findings, mimeo, Inter-American Development Bank, Washington, DC.
ESA Consultores, 1999, Evaluación Ex-Post del Fondo Hondureño de Inversión Social (FHIS 2),
Tegucigalpa.
Ezemenari, K., 1997, The Link Between Public and Prívate Interhousehold Transfers: Implications for the
Design of Safety Net Programs in Developing Countries, American Joumal of Agricultural Economics,
79: 666-671.
Ezemenari, K. and K. Subbarao, 1999, Social Assistance and Poverty-Targeted Programs: A Toolkit for
Latin America and the Caribbean, World Bank, Washington, DC.

Sesión 4 - 402

Fajnzylber, P. and W, Maloney, 1999, Labor Demand and Trade Reform in Latin America, mimeo, World
Bank, Washington, DC.
Ganuza, E., L. Taylor and S. Morley, 1998, Política Macroeconomica y Pobreza en America Latina y el
Caribe, PNUD-Ediciones Mundi-Prensa, Madrid.
Gilí, I., 1999, Managing Socioeconomic Risks in Latin America and the Caribbean, mimeo, World Bank,
Washington, DC.
Glewwe, P. and G. Hall, 1998, Are some groups more vulnerable to macroeconomic shocks than others?
Hypothesis tests based on panel datafrom Perú, Journal of Development Economies, 56: 181-206.
Goodman, M., S. Morley, G. Siri and E. Zuckerman, 1997, Social Funds in Latin America: Past
Performance and Future Role, Interamerican Development Bank, Washington, DC.
Grosh, M. E., 1992, The Jamaican Food Stamps Programme, Food Policy, 17: 23-40.
Grosh, M. E., 1994, Administering Targeted Social programs in Latin America: From Platitudes to
Practice, World Bank, Washington, DC.
Grosh, M.E. 1995. Five Criteria for Choosing Among Poverty Programs in Lustig, ed., Coping with
Austerity: Poverty and Inequality in Latin America. Brookings Institute, Wash. D.C.
Hicks, N., 1991, Expenditures Reductions in Developing Countries Revisited, Journal of International
Development, 3: 29-37.
Holzmann, R. and S. Jorgensen, 1999, Social protection as Social Risk Management: Conceptual
Underpinnings for the Social Protection Strategy Paper, Social Protection Discussion paper No. 9904,
World Bank, Washington, DC.
Holzmann, R. and T. Packard, 1999, Extending Coverage in Multi-Pillar Pensión Systems: Constraints
and Hypotheses, Preliminary Evídence and Future Research Agenda, mimeo, World Bank, Washington,
DC.
Ilahi, N., 1999, Children's Work and Schooling in Perú: Does Gender Matter?, mimeo, World Bank,
Washington, DC.
Jacoby, E., S. Cueto and E. Politt, 1996, Benefits of a School Breakfast Programme Among Andean
Children in Huaraz, Perú, Food and Nutrition Bulletin. 17: 54-63.
Jacoby, H. G. and Skoufias, E., 1997, Risk,financial markets, and human capital in a developing country,
Review of Economic Studies, 64: 311-335.
Jalan, J. and M. Ravallion, 1998, Income Gains from Workfare and their Distribution: Estimates from
Argentina's Trabajar Program, mimeo, World Bank, Washington, DC.
Jorgensen, S., M. Grosh and M. Schater, 1992, Bolivia's Answer to Poverty. Economic Crisis and
Adjustment: The Emergency Social Fund. World Bank, Washington, DC.
Kami, E., 1999, Optimal Unemployment Insurance: A Guide to the Literature, , Social Protection
Discussion Paper No. 9906, World Bank, Washington, DC.
Sesión 4 - 40

Karoly, L., et al., 1998, Investing in Our Children: What We Know and Don't Know About the Costs and
Benefits of Earlv Childhood Interventions. Rand Corporation.
Klugman, J., 1999, Social Safety nets and Crises, mimeo, World Bank, Washington, DC.
Levy, S. and E. R. Dávila, 1998, Subsidios y política social, Examen, 110:18-30.
Lustig, N., 1995, ed., Coping with Austeritv: Poverty and Inequality in Latin America, The Brookings
Institution, Washington, DC.
Lustig, N., 1999, Crisis and the Poor: Socially Responsible Macroeconomics, Presidential address to
LACEA 1999, Santiago, Chile.
Lustig, N. and O. Arias, 2000, Poverty and Inequality in Latin America: Economic Crises and Widening
Returas to Skills, Finance and Development, forthcoming.
Maloney, W, 1999, Does Informality Imply Segmentation in Urban labor Markets? Evidence from
Sectoral Transitions in México, World Bank Economic Review. 13:275-302.
Maloney, W., 2000, The Evaluation of Workfare and Condítional Transfer Programs as Emergency
Measures, and a Modification to Trabajar, mimeo, World Bank, Washington, DC,
Narayan, D., R. Chambers, M. Shah, and P. Petesch, 1999, Global synthesis: Consultations with the Poor,
World Bank, Washington, DC.
Patrinos, H. A. and G. Psacharopoulos, 1997, Family size, schooling, and child labor in Perú: An
empirical analysis, Journal of Population Economies. 10: 337-386.
Phillips, M., G. Saenz, J. Fielder, B. Rogers, P. Tatian, T. Sanghvi, and J. Berhman, 1995, The costs and
cost-effectiveness of school feeding and school bonos programs in Honduras, report prepared for USAID
under contract No. LAC-0657-C-00-0051.
Pradhan, M., L. Rawlings and G. Ridder, 1998, The Bolivian Social Investment Fund: an analysis of
baseline data for impact evaluation, World Bank Economic Review, 12:457-82 .
Progresa, 1999, Mas Oportunidades para las familias pobres, México City.
Psacharopoulos, G., 1997, Child labor versus educational attainment: Some evidence from Latin America,
Journal of Population Economies. 10:377-386.
Ravallion, M., 1993, Poverty Alleviation through Regional Targeting: A Case Study for Indonesia, in K.
Hoff, A. Braverman, and J. E. Stiglitz, eds., The Economies of Rural Organization, Oxford University
Press, Oxford.
Ravallion, M., 1999, Appraising workfare, World Bank Research Observen 14: 31-48.
Ravallion, M., 2000, On Protecting the Poorfrom Fiscal Contractions, mimeo, World Bank, Washington,
DC.

Sesión 4 - 404

Ravallion, M., and G. Datt, 1995, Is Targeting through a Work Requirement Efficient? Some Evidence for
Rural India, in D. Van de Walle and K. Nead, eds., Public Spending and the Poor: Theory and Evidence,
John Hopkins University Press, Baltimore.
Ravallion, M. and Q. Wodon, 2000a, Does Child Labor Displace Schooling?, Economic Journal,
forthcoming.
Ravallion, M. and Q. Wodon, 2000b, Banking on the Poor? Branch Placement and Nonfarm Rural
Development in Bangladesh, Review of Development Economics, forthcoming.
Ray, R., 1998, Analysis of Child Labor in Perú and Pakistan: A Comparative Study, mimeo, University of
Tasmania.
Revenga, A., M. Riboud, and H. Tan, 1994, The Impact of México's Retraining Program on Employment
and Wages, World Bank Economic Review, 8: 247-277.
Rodrick, D., 1999, Why is There so Much Insecurity in Latin America?, mimeo, Harvard University,
Boston.
Sahn, D. and H. Alderman, 1995, Incentive effects on labor supply of Sri Lanka's rice subsidy, in D. Van
de Walle and K. Nead, eds., Public Spending and the Poor: Theory and Evidence, John Hopkins
University Press, Baltimore.
Sen, A. K., 1995, The Political Economy of Targeting, in D. Van de Walle and K. Nead, eds., Public
Spending and the Poor: Theory and Evidence, John Hopkins University Press, Baltimore.
STPS - Secretaría del Trabajo y Previsión Social, 1998, Evaluación del Programa de Becas de
Capacitación para Desempleados, México DF.
Subbarao, K., 1997, Public Works as an Anti-Poverty program: An Overview of Cross-Country
Experience, American Journal of Agricultural Economics, 79: 678-683.
Subbarao K., A. Bonnerjee, J. Braithwaite, S. Carvalho, K. Ezemenari, C. Graham, and A. Thompson,
1997, Safety Net Programs and Poverty Reduction: Lessons from Cross-Country Experience, World Bank,
Washington, D.C.
Tuck, L. and K. Lindert, 1996, From Universal Food Subsidies to a Self-Targeted Program - A Case Study
in Tunisian Reform, World Bank Discussion Paper No. 351, World Bank, Washington, D.C.
Universidad de Chile, 1992, Estudio Sobre Los Programas Especiales de Empleo. Departamento de
Economía, Santiago, Chile.
Wodon, Q., 1997, Targeting the Poor Using ROC Curves, World Development, 25:2083-92.
Wodon, Q., 1998, The Cost Effectiveness of Job Creation Programs, mimeo, University of Namur,
Belgium.
Wodon, Q., 2000, Poverty and Policy in Latin America and the Caribbean, mimeo, World Bank,
Washington, DC.

Sesión 4 - 40

Wodon, Q., and N. Hicks, 1999, Protecting the Poor During Crises Through Public Spending? Framework
and Application to Argentina and México, mimeo, World Bank, Washington, DC.
Wodon, Q., K. Lee, P. Peeters, and C. Siaens, 2000, Poverty and Inequality in Latin America: 1986-1998,
mimeo, World Bank, Washington, DC.
Wodon, Q. and M. Minowa, 1999, Training the Urban Unemployed: A Reevaluation of México's
Probecat, mimeo, World Bank, Washington, DC.
Wodon, Q. and C. Siaens, 1999a, Child Labor and Government Programs in México, mimeo, World Bank,
Washington, DC.
Wodon, Q. and C. Siaens, 1999b, Food Subsidies and Consumption Inequality in México, mimeo, World
Bank, Washington, DC.
Wodon, Q. and C. Siaens, 2000, Estimating the Cost of Child Labor, mimeo, World Bank, Washington,
DC.
World Bank, 1999a (R. Holtzman and S. Jorgensen), Social Protection Sector Strategy Paper, draft,
Washington DC.
World Bank, 1999b (K. Lindert), Panama Poverty Asssessment, draft, Washington, DC.

Sesión 4 - 406

Appendix: Key Risks by Age Group, and Role of Social Protection in Addressing these Risks
Age Group
0-4

Main Risks

Role for Other Sectors

Role for Social
Protection: Insurance

Role for Social
Protection: Assistance
- Early Child
Development

- Stunted child
development

- PHC Services
- Pre-schools

5-14

- Low education quality
- Late entry to school
- Late age by grade

- Quality education
- Earlier entry
- Reduced repetition

- Scholarship and
schooling incentives

15-24

- Low secondary school
completion rates
- Teen pregnancy
- STDs

- Access/quality of
secondary education
- Remedial education
- Reproductive health

- Scholarship and
schooling incentives

25-64

- Low income
(unemployment/
underemployment)

- Labor-intensive growth
(SMEs emphasis)
- Labor market reforms

65+

- Low income
(no pensión, no assets)

AU

- Low access to and
quality of health care
- Low housing quality
- Low access to basic
infrastructure
-Insecure tenancy
- Risk of flood

-

-

Source: Adapted from Arriagada (1999.)

Sesión 4 - 407

- Workfare
- Cash transfers
- Job training
- Job search assistance

- Social security
(contributory pensions)
- Better provisión of
health services
- Affordable housing
- Investments in basic
infrastructure
- Titling programs

- Unemployment
insurance

- Income transfer
(non-contributory
pensions)

- Health insurance

- Housing subsidies
- Relocation to safe areas

— '

1

LABOR MARKETS AND INCOME SUPPORT:
WHAT DID WE LEARN FROM THE CRISES?

by
Gustavo Márquez
Office of the Chief Economist - Inter-American Development Bank

This paper has been prepared for the Conference on Social Protection and Poverty. The author wants to
thank Carola Alvarez (IDB), Gaurav Datt (IFPRI), Nora Lustig (IDB), and three anonymous reviewers for
insightful comments and suggestions.
Thefindings, interpretations, and conclusions are those of the author and should not be attributed to the
Inter American Development Bank.

Sesión 4 - 40

Table of Contente

I. Introduction
II. The failure of traditíonal income support mechanisms in Latin America
III. The "emergency" income support mechanisms in 1995
A.
Unemployment insurance
B.
Employment generation programs
C. Training as income transfer programs
D. Expenditure in income support devices
IV. The pieces of an income support system
A. Cash transfers
B.
Unemployment insurance
C.
Employment generation programs...—
D.
Short term training programs
V.
Putting the pieces together
Bibliography
Annex

Sesión 4 - 410

411
412
415
416
418
419
421
422
424
424
426
428
429

Abstract
The adoption of emergency employment and short term training programs to
provide income support to the population affected by the episodes of economic
volatility since 1995 reveáis the failure of the traditional, Labor Law-based
income support mechanisms in the región. These emergency programs are not
additional and transitory, but rather have become a permanent component of
social policies. As such, they present policy makers with the opportunity to build
a more encompassing and better designed mechanism to protect workers
against the effects of macroeconomic volatility. The paper argües for an income
support system centered on unemployment insurance and complemented with
additional mechanisms to provide effective income support to workers during
normal times, and discusses what changes are necessary in the emergency
programs to make them work as part ofthat system.

I.

Introductíon

The high level of macroeconomic volatility that has historically affected Latin American economies
has not diminished with the adoption of growth-oriented sustainable economic policies. Economy wide
shocks like the Tequila effect in 1995 and the renewed instability since 1997 have induced governments to
revisit the ideas around the building up of mechanisms that could protect the population from income
losses.
Structural reform and macroeconomic stabilization have increased therisksthat workers face given
the vulnerability of domestic production to changes in the international capital and goods markets. At the
same time, the traditional legally mandated severance payment mechanisms established in the Labor laws
have become inrelevant in this new environment, given the narrow scope of its coverage.
Therefore, countries in the región did not have the option of expanding "normal" income support
programs when hit by economy-wide shocks, but had to adopt emergency responses. The programs that
resulted from that constitute more a mosaic of often unrelated initiatives than a coherent and well thought
system of income support. However, the fact that these emergency programs are in place creates an
opportunity to build a more permanent system of income support that solves the failure of the traditional
mechanisms.
The question this paper purports to address is how a more comprehensive income support system
can be built starting with those emergency programs. In other words, what is required to transform this
mosaic of programs in a coherent set of mechanisms that can provide income support to workers both in
an economy-wide shock, and in the normal process of job churning that flexible and dynamic labor
markets experience. Efforts to cope with economy-wide shocks are doomed to failure if we do not count
with mechanisms that can provide income support in the normal course of business, and that can expand
to help workers cope with economic downturns.
The paper does not attempt to extrapólate from this small and haphazard set of programs a complete
social risk management framework. Not only extrapolation would be unwarranted given the limited size
and scope of the programs in place, but also socialriskmanagement must by necessity include programs
that do not operate through the labor market (as subsidies for school attendance of school-age children,
social investment funds, and demand side subsidies to health service consumption, to mention just a few).

Sesión 4 - 4 1

What the paper does attempt is to draw some conceptual generality from the practical (and very
real) particularities of the implementation of labor market based income support devices in the región. By
doing so, it can help policy makers to put emergency measures in the broader and more structured
framework of the long term objective of helping workers cope with both economy wide and idiosyncratic
shocks that result in income losses and unemployment.
In thefirst section, we will discuss how the traditional legally mandated severance payment system
fails to work both because it covers a dwindling fraction of the population, and because it creates rigidities
that hinder the ability of firms to operate in more open and contested markets. The second section will be
dedicated to a characterization of the labor market-related income transfer programs that 7 countries in the
región had in place around 1995, and that were used to attempt to mitígate the impact of the Tequila crisis
on their labor markets. There we will argüe that these emergency programs were not designed in such a
way as to act as counter-cyclical devices, and in some cases created important labor market distortions.
The third section will present a discussion of the changes needed in each of these programs to mitígate the
negative extemalities they generate, and to enhance the positive ones. Finally, in the fourth section we will
present some ideas on how to develop from these emergency programs an income protection system
centered on a well-designed unemployment insurance system, and complemented by employment
generation and short-term training programs to address the needs of workers who do not have access to
unemployment insurance.

II.

The failure of traditional income support mechanisms in Latin America

The high level of macro-economic volatility of Latin American economies has generated a strong
social demand for mechanisms to protect the working population from the resultingriskof income losses.
Traditionally this demand has been met by the enactment of employment security regulations that penalize
terminations either through high severance payments when terminations are allowed, or through direct
prohibition of terminations. The región has had until the mid-nineties high levels of employment
protection, even relative to those enjoyed by workers in more developed OECD countries (Márquez 1997
and IPES 1997). For workers in regulated contracts severance payments are quite high, and employment
protection regulations are strictly enforced both in practice and in legal courts.
The "social justice" rationale behind these employment security regulations is quite well known.
Firms enjoy rents that arise from high tariffs and other forms of privileged access tofinancial and nonfinancial resources in the context of so called "crony capitalism", and employment security is just one
form of sharing those rents with their employees. In a sense, the system works as privately implemented
unemployment insurance with coverage limited to those workers in a regulated employment contract.
These workers are protected both because thefirm has a positive cost associated with termination (that
ensures that lay-offs andfirings will be used sparsely as adjustment mechanisms), and because workers
who end unemployed receive an income transfer through the payment of severance.
This pattern of employment security that protects the job but not the worker after losing his or her
job is associated with the "polarized" model of wage setting (Banuri and Amadeo, 1992). For the puipose
of protecting against income volatility, this model of regulation provides effective protection against
income losses to a minority of highly organized and vocal workers, while leaving a large fraction of the
working population uncovered either in non regulated employment contracts or in self-employment.
In the cozy environment of a semi-closed economy this was possible because demand-boosting
govemment policies, however unsustainable, protected the domestic market against extemal shocks.
Workers who lost their jobs received severance payments and couldfind or invent an alternative job in the
unregulated sector of the economy. In an inflationary environment real wages could be adjusted by raising
Sesión 4 - 4 1

nominal wages slower than the general price level. Therefore, employment in the regulated sector was
quite stable, unemployment was low, and the adjustment variable in the labor market was the real wage.
To the extent that demand-boosting government policies deepened the misallocation of resources
resulting from protectionism and cronyism, the growth performance of the región deteriorated during the
eighties. As a consequence, employment in the regulated sector remained stagnant and employment in
non-regulated jobs without any form of protection against income loss grew to cover almost half of the
working population (see Márquez [1997]). Therefore, the. protection against income volatility ended up
covering just the fraction of the labor forcé employed in the bigger and more closely regulated firms.
In the more dynamic environment that resulted from economic opening and stabilization at the
beginning of the nineties, the relevance of this model of protection is questioned by the increase in the
pro-cyclicality and level of unemployment. What caused these changes in unemployment? First and
foremost, structural reforms and macro stabilization changed the relative prices of capital and labor in
favor of the former (Lora and Olivera [1998]). This in and by itself increased the use of physical capital
and slowed employment generation. But macro stabilization also implied important changes in the
employment behavior of the public and private sectors.
Public sector retrenchment resulted in the displacement of workers that increased the stock of
unemployed. More importantly and of larger consequences, the maintenance of low or zerofiscal déficits
required byfiscal discipline makes it impossible for governments to act as employers of last resort, as they
had done directly or indirectly in the crisis of the eighties (Lora and Márquez [1998]).
The reduction of inflation reduced the latitude that private (and public) employers had enjoyed to
reduce real wages through nominal wages lagging behind inflation. The increased nominalrigiditymakes
employment —and not real wages— the adjusting variable in the labor market during economic
downtums. Therefore, shocks transíate themselves into more unemployment, rather than into a collapse of
real wages (Márquez and Pages [1998]). Once this higher level of unemployment is reached, slower
employment generation makes its reduction a more protracted and prolonged process over time (Márquez
[1998]).
From a world where unemployment was almost absent, and where low quality jobs in the nonregulated sector were the crux of the problem, we jumped into a brave new world of growing
unemployment. By the end of 1998 a record number of countries in the región were experiencing double
digit unemployment rates (Argentina, Colombia, and Venezuela the most notorious cases) and, more
concerning, unemployment remains high when the economy recuperates after each shock (see Graph 1).
Although macro andfinancial policies are used to limit the domestic vulnerability to external shocks, the
fact remains that the recessionary environment of 1999 is translating itself in most countries in more
unemployment and falling wages, particularly for the less skilled workers.
But perhaps the most damaging blow to the relevance of severance payments-based income
protection has been the recognition that privately insuring workers against income losses may hurt the
ability of firms and workers to profit from the enhanced opportunities associated with a more open and
competitive economic environment. Colombia in 1990, Perú in 1991, Nicaragua and Argentina1' in 1995,
and Venezuela in 1997 have enacted reforms of labor regulations aimed at creating more flexible forms of
employment contracts, both through the reduction of firing costs and through the introduction of
promotional contracts with lower payroll taxes.

i/ Some of the Argentina reform measures of 1995, however, were partially reversed in 1998.
Sesión 4 - 4 1

Graph 1
Unemployment and growth in Latin America, 1980-1996

(Population-weighted

80

81

82

83

84

85

90

91

averages

92

93

94

95

96

for

17 countries

97

Source: Growth of GDP at constant 1990 USS from IADB Statistical data base; unemployment data from
ILO Panorama
Laboral.

One of the impacts of these reforms has been a further reduction in the share of full-benefit
contracts in total employment, and a corresponding expansión of more precarious forms of employment.
The introduction of more flexible forms of employment contracts has not in fact weakened the
employment protection enjoyed by workers in the oíd full-benefit, full-protection employment contracts.
With relatively few exceptions, the labor reforms enacted during the nineties have aimed at flexibilizing
hiring andfiring conditions at the margin, without diminishing the legal protections enjoyed by workers
already employed. Making the changes binding only on new labor contracts has in fact grandfathered
employees already in the job. However, as the new contractual forms have cost advantages for firms,
employers have substituted workers in full-benefit contracts by workers under the moreflexible contracts.
In Argentina, for instance, promotional employment contracts generate lower payroll taxes and do not
generaterightsto severance payment upon termination. According to official information2 the number of
new jobs created under promotional contracts more than doubles the number of new full-benefit contracts,
while the opposite is true regarding the number of workers terminated under each type of contract.
In summary, at the mid-nineties wefind ourselves in a situation where unemployment is higher and
more sensitive to economic downturns, and where more than half of the employed do not have any
protection against theriskof income loss associated with unemployment. Economic downtums may result
in very large income losses for both the unemployed and workers in more precarious employment
contracts, and society does not offer them any formal mechanism to insure against that risk.
21

Encuesta de Indicadores Laborales, Min. de Trabajo y Seguridad Social, October 1998.

Sesión 4 - 414

III.

The "emergency" income support mechanisms in 1995

In order to at least partially compénsate for the adverse labor market effects of the Tequila effect in
1995, governments in the región adopted and/or expanded labor market policies and programs aimed at
supporting incomes and/or expanding employment opportunities for particularly vulnerable groups. The
renewed volatility in international capital markets since the end of 1997 has maintained the interest of
policy makers in the adoption of income support programs and suggests that these programs are more
permanent than what the "emergency" label indicates.
The enactment of these programs has created an opportunity to test some new ideas and to
disseminate innovative approaches to oíd problems. Labor intensive public work programs, youth-aimed
training programs, and semi-universalized unemployment insurance systems all became acceptable ideas
to help sustain the fragüe political coalitions behind the economic modernization process.
These programs were designed in a crisis context, when incomes were falling, unemployment was
increasing, and the risk of loosing political support for the reform efforts was high. Therefore, the ability
of these mechanisms to expand and reach the groups atriskas fast as possible took priority over optimal
design considerations. As a consequence, we ended up with a mosaic of programs and mechanisms that
lack the coherence and size necessary to fulfill the role of income support to wider groups of the
population, that often introduce labor market distortions, and that have difficulties adjusting countercyclically.
On the positive side, these programs are generally well suited to the institutional layout of the
countries and have demonstrated capability to deliver resources fast. The experience of the región with
income support programs shows than they can, at least to some extent, help to compénsate for the effects
of economy-wide shocks on workers. A varied array of programs has been used to help workers cope with
one or another of the damaging effects of economy-wide shocks. This battery of programs will be the
backbone of any income support system, given organizational, political, and resource constraints that limit
the ability of the government to create new programs.
Given the urgency of coping with the effects of the crisis, the quest for mechanisms that could be
set up quickly to transfer income to the poor in the most targeted way possible dominated any optimal
design consideration. In a perfect world, countries would have automatic mechanisms as unemployment
insurance to protect vulnerable groups during economic shocks. In reality, however, few countries in the
región have a system of unemployment insurance and, even in those countries, the protection of
unemployment insurance is limited to a relatively small fraction of the total workforce.
What the countries did was to put in place or expand a wide array of actions, programs, and policies
that were considered suitable as income transfer devices and that could be expanded on short notice. Labor
intensive public works programs executed in a decentralized fashion by sub-national governments or
NGOs were widely used. Short-term training program with scholarships and some form of job search
assistance were also used to provide an immediate source of income for the beneficiaries.
Public investment programs were not used as employment programs, in spite of their important
direct and indirect employment effects. Part of the reason for this was the lack of fiscal resources to
finance their execution; it was often the case that investment programs were used as adjustment variable in
equilibrating the fiscal budget. Besides, there is no technical reason why (and often many why not) a
large-scale investment in sanitation, for instance, should be executed with labor intensive technology.3'

y

Another factor that probably explains this is that the redesign of investment programs to incorpórate labor-intensive
technology takes a long time and requires a complete re-engineering of the investment project, not to speak of the needed
institutional redeployment.

Sesión 4 - 41

The particular responses of different countries varied widely both in function of tradition and
history of labor market programs implementation, and in terms of the political equation that made diverse
groups more or less visible as objects of income protection. Generally speaking, governments reacted
along two main lines. Where organizational and political constraints so permitted, labor intensive public
work programs were implemented. Where the perception was that youth unemployment was a particularly
important problem, training programs that targeted unemployed youth as clientele were implemented.
Some governments in the región opted for developing both lines of action simultaneously, and yet some
others opted for an ever-wider variety of programs that also covered subsidization of prívate sector
employment generation.

A.

Unemployment insurance

In Table 1 we present a summary description of unemployment insurance systems in the región. As
can be seen, very few countries in the región have legally and/or administratively enacted unemployment
insurance systems, and even fewer have working unemployment insurance schemes. This is a
consequence of the very weak incentives that exist for the development of unemployment insurance and
other more socialized forms of income protection, given the fact that severance payments work as
privately provided income insurance for workers in full-benefit employment contracts that are laid-off.
In those countries that do have unemployment insurance systems, coverage is limited to workers
that have contributed while employed to thefinancing of the system. In other words, only workers in fullbenefit employment contracts and working in payroll tax-paying firms enjoy the benefits of the
unemployment insurance system. The level of benefits provided and its duration is low relative to more
developed countries unemployment insurance systems. Replacement rates are normally in the order of 5060% of last wage, with tops linked to the minimum wage for higher salaries. Benefits are granted for
periods typically not longer than four months.
The unemployment insurance system in Argentina is quite limited in number of beneficiaries, and
has remained so in spite of strong increases in the number of unemployed workers. Mazza 1999 reports
that the number of beneficiaries has remained stable around 100-125 thousand workers, more than 70% of
whom are prime age males, and more than 50% are not household heads. She also reports that an analysis
of beneficiaries in their personal and previous job characteristics shows that there is a definite trend
towards serving younger and middle class displaced workers. This suggests that unemployment insurance
is not fulfilling a safety net role for the poor in the case of Argentina.
Brazil has the biggest unemployment insurance system in the región, with around 300-400 thousand
beneficiaries. Mazza 1999 reports that unemployment insurance in Brazil is also serving younger (more
than 50% of beneficiaries are younger than 30) and more educated (45% of beneficiaries have completed
eight grade or better) workers. As IPEA 1998 assesses, the unemployment insurance system reflects wage
inequality, in the sense that benefits accrue to the middle deciles in the distribution of income.
In Venezuela, the unemployment insurance system is enacted in the law since 1989, but was never
implemented. The system has been reformed last year. The new system will protect beneficiaries through
a mix of individual and collective insurance operated by competitive insurance providers, but
implementation has not yet begun as of this date. Given that only workers with regulated, tax-paying
contracts are entitled to benefits, it is likely that the pattem of distribution of beneficiaries will be very
similar to that of Argentina and Brazil discussed above.
México and Uruguay have unemployment insurance programs operated by the social security
system. In both cases coverage is quite limited, and in the former it is just an advance payment of oíd age
pensions for a máximum period of five years. In the Barbados case, the unemployment insurance system
Sesión 4 - 41

is very small in coverage, though quite well adapted to the needs of an island economy with frequent but
short episodes of unemployment concentrated in workers in the tourism industry (Mazza 1999).

IV. FUN
DING

Law

Argentina

Barbados

Brazil

Chile

Ecuador

1991 reform. Worker: 1 % wage
Empr.: 1.5 %
95
payroil

1982

1986
1990

° rWw9a : e 5 %
a

®
Empr.: 1.5%
payroil

Replacement
rates < a )

Benefit
Benefits
duration Min/Max

Coverage

Requirements

4-12
meses

Employees

1(12), 2,3

Employees 1664 yo

1(6)

.
Employees

.,„_
c _
4(36,4), 5,6

Employees

2, 4(12,2), 5,

Employees

1(24), 7(30)

60%

weeks

4 0 O / o 1 6 w e

1-3 minimum

1981

Government

37$ monthly
for the first 6
months, 18$
last 6 months

1958,1988

Worker: 2% sal.
Empr.: 1 % payroil

26 weeks
¡na 52
week
e k s
period

wages

,

... „ _
months Min: 1 m.w.

4

Max. 1
year < b )

One time subsidy,
amounts decided each
year.

Social Security

1981

1989 reform.
Venezuela
1998

Min: 1 m.w.
Max:4 m.w.

6 0 % 1 0

FAT (.65% tax on
total sales)

México

Uruguay

W

Table 1
UNEMPLOYMENT INSURANCE IN THE REGION

95 % pensión

5 years
max

Contributions to
Social Security

Up to 50%

6 months

Worker: .7 %
wages
„
Empr.: 1.5%
payroil

Up to 60%

13-26
.
weeks

Employees 60- Between 60-65
65 yo
yo
Min:0.5m.w Employees in
. Max: 4 commerce and
m.w.
industry

Max. $44

Employees
K
1

1 (6), 5, 3, 8.

1(12), 2

Taken from Lora E. and C. Pages, 1997
Source: S.S. Programs throughout the World -1995, US Dept. of Health and Human Services
Notes: ( a ) % last wage.
^ Beneficiaries receive also family support, medical and maternity benefits.
Requirements: 1 (s) - Be employéd s months before receiving subsidy; 2 - Availability to work; 3 - Does not receive; other social
security benefits; 4 {s, j ) - Not having received more than s months of benefits in the last j years;
5 - Unemployed for reasons outside the conduct and willingness of worker; 6 - Subject to economic need; 7(x) - Waiting period of x
days; 8 - At least 12 months between periods of receiving subsidy.

Sesión 4 - 4 1

B.

Employment generation programs

In Table 2 we present a summary description of the employment generation that 7 countries in the
región had in operation at the end of 1995. The list was extracted from a joint ILO-IADB volume on
active labor market policies in Argentina, Brazil, Chile, Costa Rica, Jamaica, México, and Perú. These
countries represent a wide spectrum of variation in terms of policy development, operational capabilities,
and exposure to international capital markets volatility. Program description and characteristics were
summarized in Verdera (1998) and a more thorough discussion of programs is presented in the national
reports contained in that volume. An itemized description of the programs is presented in Table A-l in the
Appendix.
Table 2
EMPLOYMENT GENERATION PROGRAMS
IN 7 COUNTRIES IN THE REGION

Argentina
Brazil
Chile
Costa Rica
Jamaica
México
Perú

Beneficiaries
% of total
1.000S
forcé
9.31%
892.2
221.8
0.49%
0.10%
4.3
0.71%
8.1
0.61%
6.0
4.42%
1,024.0
0.93%
27.8

labor
million

Expenditure

249.2
1,188.8
1.4
3.3
21.2
1,802.0

US$ % GDP
0.09%
0.21%
0.00%
0.04%
0.50%
0.51%
0.19%

100.0
Source: data from Verdera 1998, modified by the author. For a complete listing, see
Table in Appendix.

Argentina is the country with the most varied set of employment generation programs, comprising a
combination of public works and subsidies to private employment. Public subsidies to private sector
employment, in the form of subsidies to firms that increase the number of employees, were widely used
under a variety of mechanisms. Workers displaced from the public sector and unemployed workers
receiving unemployment insurance were given vouchers that employers could use to pay tax liabilities.
Firms could opt for tax rebates if hiring particular groups of workers (young, women, ex-combatants, etc.)
under promotional contractual forms. Also, firms in particular activities (like reforestation) were
subsidized if they hire new workers. But the most visible mechanism of subsidization was the use of
"promotional employment contracts" established in a series of decrees in 1995. These promoted contracts
were more precarious than regular full-benefit contracts, did not originaterightsto severance payments,
and payroll tax liabilities were lower.
Argentina federal government also financed labor intensive public works as an employment
generation device. The Trabajar and similar programs were financed and supervised by the Federal
government using the Fondo Nacional de Empleo (a fund financed through pay-roll taxes). The resources
were used to build small scale and labor-intensive public works (in many cases social infrastructure, but
also roads and small sanitation works), with the works being executed by a wide variety of agencies,
including from local and state governments to NGOs.
The PROGER program in Brazil is a contrasting mechanism for employment generation. The
program operates through the establishment of credit lines offered through the national development
banking system to smaií enterprises, cooperatives, NGOs, and other civil society associations. This
mechanism serves to circumvent the sub-national governments for works execution, in order to avoid the

Sesión 4 - 4 1

creation of budgetary entitlements. Partial and incomplete evaluations of PROGER, however, are not too
optimistic about the results in terms of employment generation.4'
Chile does not have any employment generation program as such, though it has a number of very
small and narrowly targeted programs to address living conditions that may hinder the labor market
insertion of particular groups.
Costa Rica uses public works, wage subsidies, and credit to small enterprises as mechanisms to
promote employment generation. Credit to promote employment generation in smallfirms is also widely
used in Jamaica in a battery of programs, some of which also include a form of short-term training.
Jamaica has a training and temporary employment program for unemployed youth, aimed at easing their
labor market insertion.
México uses public works (rural roads and other social infrastructure) as employment generation
devices. The programs arefinanced by allocations from general revenues (not from payroll taxes) in the
Federal Government budget, and states and local governments execute the works.
Finally, Perú uses legal incentives, a social investment fund, and a micro and small enterprise credit
program as tools for employment promotion. The Labor law reform of 1991 introduced a number of more
precarious forms of employment contracts, allowingfirms to hire workers without generatingrightsto
severance payments underfixed term contracts. FONCODES, a social investment fund, is also used as a
employment generation device which can be quickly adjusted to the situation of local labor markets.
However, it is not clear how much capacity or interest FONCODES management has in employment
generation as opposed to the physical execution of civil works (Verdera 1995).

C.

Training as income transfer programs

In Table 3 we present a summary description of the training programs that were being used as
income transfer devices in 7 countries in the región by the end of 1995 (Verdera (1998)). An itemized
description of these programs is presented in Table A-2 in the Appendix. Training programs were widely
used as a mechanism to transfer income, particularly to unemployed youth, through scholarships during
the classroom-training period (normally 3 to 6 months) and in some cases through job search assistance
and/or apprenticeship stages in prívatefirms. In most cases these training programs werefinanced by the
government and delivered by prívate and NGO training providers, with little or no intervention of the
traditional National Training Institutions
The basic operational technology of these training programs was based on Chile Joven, a pioneering
youth training program that combined an scholarship for classroom-training with a three month paid
apprenticeship in a prívatefirm. Instead of direct purchasing of training services, resources are used to
create a fund that is managed by a central government agency. The managing agency requests proposals
for training projects, and funds are granted through open biding. The proposals must describe the content
of the courses to be taught, and must include a commitment from prívate sectorfirms to accept the trainees
as apprentices for a period of time (normally three months). The provisión of scholarships served as an
income transfer to beneficiaries, took them out of the unemployment queue, and gave them some labor
market experience during the apprenticeship. These three beneficial effects of the Joven program were

4/

Relatorio da forca-tarefa sobre políticas de emprego - Diagnóstico e recomendaciones, Min. do Trabalho, Brasilia, Agosto
1998.

Sesión 4 - 4 1

quite adequately suited to situations characterized by high youth unemployment rates5'. However, other
countries in the región emulated the contracting methodology of the Joven program to cater to the needs of
other population groups.
Table 3
TRAINING PROGRAMS IN 7 COUNTRIES IN THE REGION
Beneficiaries

1.000s
Argentina
Brazil
Chile
Costa Rica
Jamaica
México
Perú

133.0
740.5
36.6
13.1
43.5
410.3
1.5

% of total
labor forcé
1.4%
1.6%
0.8%
1.2%
4.4%
1.8%
0.1%

V.

EXPENDI
TURE
million US$
% GDP
95.6
310.2
18.3
60.6
18.6
135
5.0

0.04%
0.06%
0.03%
0.73%
0.44%
0.04%
.01%

Source: data from Verdera 1998, modified by the author. For a complete
listing, see Table in Appendix.

Among the countries in the study reported here, Argentina, Chile and Perú have programs inspired in
the Chile Joven design, targeting low income unemployed youth. Argentina has also used the contracting
mechanisms of the Joven program to develop training programs for other groups of the population, and
granted subsidies to prívate employers who hire apprentices under promotional employment contracts.
Brazil also uses competitive biding for training provisión, but the program operates in a highly
decentralized way. The PLANFOR program isfinanced through the FAT, a payroll taxfinanced fund, and
funds are allocated to states and local governments, who in tum hire different providers (both prívate and
public) through competitive bidding. States must present annual training plans to the PLANFOR
administration, and funds are allocated in proportion to the state's share of the total workforce. This method
of allocation is presently being changed to reflect the state level of poverty and education, and past
experience with the execution of annual training programs. It is interesting to note that the national training
institutions (in the case of Brazil the SENAI-SENAC system) particípate in the bidding process as another
provider of training services, thus creating an interestingfinancial and institutional dynamic in the overall
training system.
Costa Rica used instead the national training institution (INA) as a channel for delivery of training
services to semi- and skilled unemployed workers. Thus, INA schedules and delivers training programs for
low income workers in marginal urban areas, for displaced public sector workers and for handicapped
workers using its own facilities and instructors. A special line of action was established to enable INA to
contract out other training institutions, but no special targeting mechanism has been used.
Jamaica uses a number of programs to provide training for unskilled and young unemployed workers,
but the mechanism for income transfer is temporary jobs rather than scholarships during training.

The contracting mechanism of Chile Joven was in f c a way to create incentives for training providers to deliver good quality
at
and labor market-relevant content of their courses. This created pressures for a institutional and content revamping of the
n
training system, a firms accepting apprentices acted as controllers and gatekeepers of the relevance a d adequacy of t e
s
n
h
training provided. The program was therefore rightlyperceived as a tool to modemize and connect t e training s s e with
h
yt m
real productive activities.
Sesión 4 - 4

México has the biggest training with income transfer program in the región, and it has been effectively
used as a protective device for unemployed and displaced workers, and expanded and contracted according
to the economic cycle. The PROBECAT program provides a scholarship for the beneficiaries, and the state
offices of the Labor Ministry organize a variety of training programs that are delivered locally. Different
program evaluations have found that the program has been somewhat successful as a training program,
increasing incomes and likelihood of employment for beneficiaries, even though positive effects tend to
increase with higher levels of education of the beneficiary.6'

D.

Expenditure in income support devices

In general terms, the 7 countries in our sample invest a smaller fraction of GDP in unemployment
insurance, employment generation, and training-cum-income-transfer programs that a wide sample of more
developed OECD countries (Table 4). On average, the countries in our sample spend less than half a point
of GDP in these programs, while the average for the OECD countries is 2.4% of GDP. Only the US, Japan,
and Greece spend a smaller fraction of GDP in these programs than the LAC average.
México is the country that spends the most resources in employment generation (one half of a
percentage point of GDP) to benefit around 4% of the total workforce, while Argentina reports expenditure
below one tenth of 1% of GDP to benefit around 9% of the total workforce. Resources invested in training
programs7' are of the same order of magnitude than those dedicated to employment generation programs,
though the number ofbeneficiari.es seem to be somewhat bigger.
It is also interesting to note that countries in the región that do have unemployment insurance systems
(Brazil and Argentina) spend less in unemployment compensation than even the less spendthrift countries in
the OECD.
These comparativefigures suggest that there is room for expansión of the expenditure in income
support programs. Even doubling the expenditure in these programs (an increase well beyond any
reasonable expectation) would not make countries in the región big spenders in international terms.

11

Capacitación y Empleo: Evaluación del programa de becas de capacitación para desempleados, Sec. De Trabajo y Prev
Social, México DF, Agosto 1 9
95
These figures exelude t e expenditure of t e traditional National Training institutions.
h
h
Sesión 4 - 4 1

Table 4
EXPENDITURE ON LABOR MARKET PROGRAMS
OECD AND LATIN AMERICAN COUNTRIES

(% of GDP)

Training for
unemployed
Australia (1994-95)
Austria (1995)
Belqium (1995)
Canada(1994-95)
Denmark (1995)
Finland (1995)
France (1995)
Germany (1995)
Greece (1995)
Italy (195)
Japan (1994-95)
New Zealand (1994-95)
Spain (1995)
Sweden (1994-95)
US (1994-95)

0.23
0.13
0.24
0.29
0.86
0.60
0.67
0.44
0.04
0.39
0.03
0.44
0.33
0.98
0.07

Argentina (1995)
Brazil (1995)
Chile (1995)

0.04
0.06
0.03

Employment
generation
OECD countries
0.21
0.05
0.68
0.07
0.36
0.68
0.40
0.44
0.09
0.69
0.06
0.15
0.31
0.90
0.01
LAC countries
0.09
0.21
-

Unemployment
compensation

Total

1.62
1.30
2.11
1.50
3.06
3.57
1.43
2.08
0.44
0.68
0.39
1.26
2.46
2.51
0.35

2.06
1.48
3.03
1.86
4.28
4.85
2.50
2.96
0.57
1.76
0.48
1.85
3.10
4.39
0.43

0.14
0.19

0.27
0.46
0.03

-

Costa Rica (1995)
0.73
0.04
0.77
Jamaica (1995)
0.44
0.50
0.94
México (1995)
0.04
0.51
n.a.
0.55
Perú (1995)
0.01
0.19
0.20
Notes: Training for unemploved includes training for the unemployed adults and those at risk, plus measures
for unemployed and disadvantages youth and support of apprenticeship programs.
Employment generation includes all forms of subsidized employment, plus direct job creation by public and
non-profit sector.
Unemployment insurance includes all expenditure on benefits, independently of source of financing.
Source: For OECD countries, Table J, ps. 211-218, Employment
Outlook,
June 1998, OECD.
For LAC countries, Verdera(1998).

IV.

The pieces of an income support system

The repeated bursts of economic instability during the nineties induced govemments to revisit the
ideas around the building up of mechanisms that could protect the populationfrom income losses. These
efforts, described in the previous section, were conceived as emergency measures aimed at transferring
resources as fast as possible to vulnerable groups of the population affected by an economy-wide shock.
Much more attention was brought to the capacity of these programs to expand rapidly than to their ability
to expand or shrink in response to demand. This trait made these programs a semi-permanent component
of social sector expenditures, therefore raising all types of questions regarding their role in a
comprehensive social policy.
One question that has been overlooked in the discussion about income support programs is what
they tell us about the failure of the established mechanisms in our societies to protect workers (including
the poor) from the effects of both economy-wide and idiosyncratic shocks. Because of the failure of the
Sesión 4 - 4

traditional income support mechanisms (i.e., the severance payment system) there is an absence of
mechanisms to deal with the income losses associated with the idiosyncratic shocks normal in the process
of job churning. Therefore, when hit by the crisis countries in the región did not have the option of
expanding "normal" income support programs, but had to adopt emergency responses. On the positive
side, however, these emergency programs create the opportunity to build a more permanent system of
income support that solves the failure of the traditional mechanisms.
This system of income support should operate through the labor market, because labor is the only
asset for an overwhelming majority of the population and because most of the transmission mechanisms of
shocks operate through labor demand (Lustig and Walton, 1998). Programs and policies that facilítate the
use of labor in income earning activities will be the most effective to mitígate the impact of cyclical or
shock-originated downtums on workers. However, when workers fall into unemployment their
consumption capacity needs to be protected (be it by generating employment, or by providing other forms
of income support) without generating a dependency trap.
From a design point of view, these mechanisms should provide a mínimum income guarantee to all
workers. To be feasible infinancial and economic terms, they need to fulfill at least three requirements:
•

Their design should not introduce labor market distortions and, in particular, should not induce
workers to reduce their search effort or increase their reservation wages.

•

Their coverage should be as wide as possible, given that the risk of unemployment affects all
workers, including those in non-regulated and precarious forms of employment contracts.

•

Their budget allocation should be adjusted counter-cyclically, expanding in economic downtums
when unemployment increases, and contracting in expansions when it decreases.

How well do the programs and mechanisms already in place in the región comply with this
requisites? Table 5 presents an attempt to summarize some characteristics of these programs relative to
those criteria. In the most general possible terms wefind that an effective and relevant income protection
system should be a composite of programs offering different mechanisms targeted to different groups in
function of their human capital levels and labor market insertion.
Table 5
SOME CRITERIA FOR EVALUATION OF INCOME PROTECTION PROGRAMS IN THE REGION
Program

Labor market distortions

Coverage

Counter-cyclically

Cash transfers

Very important
Reduce incentives for labor
market participatíon

Requires very thorough and
expensíve targeting system to
target on the poorest.

Low
Create entitlements,
contraction is politically very
expensive.

Unemployment
insurance

Low if well designed (individual
contributions, low replacement
rates, short period of benefits)

High productivity workers in
regulated employment contracts
Requires keeping of administrative
records

w. .
Mign

V.
EMPLOYMENT GENERATION
PROGRAMS
•
Labor intensive
public works
Wage subsidies
Scholarships for shortterm classroom
training,
apprenticeships, and
job search assistance

V
L

'

a u t o m a t l c

VII.

Politically expensive if the
Low if well designed (wages
Local labor markets
program generates budgetary
below average wages of poor
Self-selection into the program
entitlements of sub-national
deciles)
governments
High
Narrowly defined groups
Potentially very high, though
Induce substitutíon of nonRequires supervisión capacity and program may create
subsídized by subsidized groups keeping of administrative records entitlements

Low
May increase participatíon rates
of target group
Increase the efficiency of market
exchange mechanisms

Targeting ¡s easy on observable
characteristics, self-selection
More adequate for unemployed
youth and new-entrants, not very
useful for displaced workers

Sesión 4 - 4

High.
Requires maintenance ofthe
network of non-government
providers.

A.

Cash transfers

The most immediate and direct way to protect unemployed workers' income is through cash
transfers to families that fall below a predetermined income level. Even though the criteria for receiving
benefits from the program is formulated in terms per capita family income, low family income levels are
associated with either very low wages (Hausmann and Szekely (1998)) or unemployment.
Cash transfer programs are usually targeted to the poorest segments of the population, which cannot
obtain a minimum survival income level in the labor market. In many cases, a cash transfer is part of a
more comprehensive program aiming at protecting and furthering the ability of low income families to
maintain and accumulate human capital and, therefore, to "gradúate" from the program.8/ Because these
families are poor to begin with, economy wide or even idiosyncratic shocks can put into question their
ability to sustain minimum consumption levels. Therefore, a cash transfer can help them smooth their
consumption levels.
However, cash transfers induce very important labor market distortions, by increasing reservation
wages and creating incentives against work. Because they do not require any counterpart work effort
(making participation in the program effectively a free good), these programs also require a very
sophisticated targeting system and a complex system of verification and enforcement, to avoid fraud by
inclusión of non-needy families and to avoid exclusión of needy ones. Cash transfers also tend to create
strong entitlements among the beneficiaries, making it extremely hard to adjust either the number of
beneficiaries or the amount of benefits in a counter-cyclical fashion.

B.

Unemployment insurance

Though differences in design, coverage, and benefits makes it quite difficult to present an overall
assessment of the importance of unemployment insurance systems as part of a comprehensive income
support mechanism, there are some common traits that deserve comment. In thefirst place, unemployment
insurance is normally a benefit additional to severance payments. Once the worker is fired, he or she
receives severance payments. In addition to that, the worker has therightto unemployment insurance as a
supplementary source of income during his or her search for a new job. Therefore, income protection by
the unemployment insurance system is targeted to workers that have had full-benefit employment
contracts. As we mentioned above, this exeludes from the protection a sizable fraction of the work forcé
that work in the unregulated segment of the labor market, presumably those who because of their human
and social capital déficits are the neediest in terms of income protection.
In the second place, unemployment insurance systems generally lack connection with other labor
market intermediation and placement services. Even in the cases where the UI system is operated through
the Labor Ministry (as in Brazil) workers are not required to register in the intermediation service, and
payment of the benefit is not contingent on verification of search effort. On the one hand, this lack of
connection generates opportunity for fraud. Even if it is illegal to have a job and receive UI payments
simultaneously, most operators complain of their lack of capacity to control what is perceived to be

8/

The rationale for programs as Bolsa Escola in Brazil, Programa de Asignaciones Familiares (PRAF) in Honduras, and Beca
Alimentaria en Venezuela is to avoid perpetuating a vicious circle of poverty. Thus, these programs require keeping children
in school or attending primary health care facilities during pregnaney, as m c a i m to prevent t e transmission of poverty
e h ns s
h
to t e next generation.
h
Sesión 4 - 4 4

widespread fraud and collusion between firms and workers.9' On the other hand, this lack of connection
with labor market intermediation services makes the system a puré income transfer that does not ease the
transition of the unemployed into a new job.
In the third place, most unemployment insurance systems arefinanced through payroll taxes, which
are already quite high in the región. This partly explains why coverage is limited, replacement rates are
low, and periods of coverage quite short. Any expansión of the system to cover hitherto unprotected
segments of the population is likely to face substantial opposition by its own present beneficiaries and by
firms operating in the regulated sector of the economy. Only in the case of Brazil some expansión to new
groups have been made (to traditionalfishermen and to workers affected by the drought in the Northeast),
but the expansión has been temporary andfinanced through the use of excess funds. If unemployment
insurance is to work as part of the safety net in a crisis, the expansión of coverage would have to be
produced just when the flow of benefits to already protected workers is highest, creatingfinancial strains
on the system and the need for additional funding. The question is whether this effort is worth doing
through the unemployment insurance system, or by creating alternative mechanism for income transfer
better suited to the needs of workers with different labor market insertions.
In the fourth place, it has been argued that the implementation of unemployment insurance requires
considerable institutional resources in terms of accounting and record keeping. However, it should be
noted that mandatory savings-based schemes in place in countries as Brazil, Colombia, Ecuador, and Perú
also require considerable institutional resources of the same type. Furthermore, pensión system reforms in
a number of countries in the región10' have created a network of institutions that hold individual worker's
accounts that can be used for record keeping in the unemployment insurance system with little additional
costs.
In conclusión, the design and target population of unemployment insurance makes it suitable to
protect workers who have full-benefit employment contracts and that acquire rights to it through their
contributions while employed. In terms of labor market distortions, the low level of benefits and their
short duration apparently do not create incentive against search. In fact, the reports on fraud in Argentina
and Brazil rather suggest that workers use unemployment insurance as a means to obtain additional
income while in a new job. As Hopenhayn and Nicolini (1999) show it is possible to design optimal
unemployment insurance schedules that do not induce reduction in search efforts. Furthermore, schemes
of unemployment insurance based on nominative contributions to individual accounts that can be rolledover into retirement funds can minimize negative impacts on search effort.11'
In terms of ability to expand and contract counter-cyclically, unemployment insurance expenditure
is an ideal mechanism. By definition outlays increase when unemployment is raising, and contract with
the recuperation of employment.
The most problematic aspect of unemployment insurance, however, is related to its coverage.
Workers must bear at least part of the cost of insurance to prevent moral hazard problems. For high
productivity workers wages are high enough to make the benefits of paying for unemployment insurance
(the expected valué of benefits when unemployed) higher than the current income foregone by paying the
contribution. However, for low productivity workers the utility gain from an increase in current income
will be big enough as to generate incentives to negotiate with employers a contract without benefits in
exchange for a higher current income.

Mazza 1 9 reports t a s m efforts have been m d in Argentina to detect if workers receiving unemployment insurance
99
ht o e
ae
were working by using a c m o taxpayer identification number. It was f u d t a a sizable number of workers were not
o mn
o n ht
only working, but also contributing to t e social security in a new job while continuing to receive the UI payment.
h
l0/ Most notably Chile, but also Argentina, Uruguay, Perú, a d Venezuela.
n
n / For a proposal of a u e po m n insurance s s e
n nm ly e t
y t m along these lines see Cortazar et al. (1995) and t e Venezuelan Social
h
Security law of 1 9 .
97

91

Sesión 4 - 4

C.

Employment generation programs

Employment generation programs are a natural reaction of govemments to increasing
unemployment. Politically they show the concern of the govemment with the workers' plight and, by
providing jobs, they directly attack unemployment. For analytical purposes it is convenient to separate
labor intensive public worksfrom wage subsidies to the prívate sector.
1)

Labor intensive public works

Labor intensive public works have been the tools of choice to deal with economy wide shocks. The
number and variety of programs in place in the región shows that govemments choose to spend more
additional resources in employment generation than in other mechanisms to provide income support to
unemployed workers. One of the main advantages of these programs is that they are self-targeted (Grosh
1994, Ravallion 1998) and, therefore, can be implemented without the delays necessary to implement a
targeting mechanism.
Three characteristics of labor intensive public works are crucial in their success as income support
mechanisms. In thefirst place, these programs arefinanced by the central govemment and executed by
local organizations (be it local govemments or NGOs), who normally are in charge of selecting the works
to be performed and the selection of beneficiaries. Thus, labor intensive public works require an extensive
and solid network of institutions at the local level, with the technical and operational capacity to choose
the works to be done, to organize the production process, and to channel resources to the needy poor. A
large part of the success of these programs hinges on how well structured is the relationship between the
central govemment and the executing agencies. There is not a unique way of designing this relationship.
To mention just two examples, Argentina choose to finance works that are approved by a central
govemment agency and executed mostly by local govemments, while Brazil choose to allocate resources
semi-automatically on a regional needs base and have works selected by the sub-national govemments. In
any case, what is important is that the design of the relationship betweenfinancing and work execution be
adequate to the institutional and political structure of the country. More federalist countries should respect
local autonomy in work selection and allocate budgets on objective criteria, while more centralized
countries will be more able to select works and distribute resources at the central level while keeping
responsibility for execution at the local level.
In the second place, the wage level and the criteria for selection of beneficiaries are set at the central
level, while local organizations are in charge of the selection of beneficiaries. Thus, there is a certain
degree of tensión between the criteria set at the central level and the local political and social reality
within which the selection of beneficiaries take place. There are múltiple ways to solve, or at least mitígate
the consequences, of this tensión. Community participation is useful to oversee that resources are not
diverted through political favoritism or other forms of corruption, but there is no guarantee that the needed
level of community participation will exist. A useful complement to community participation is a system
of random sampling of projects and beneficiaries by the central govemment agency in charge of
overseeing the program to check whether or not resources are being diverted. This implies a non-trivial
investment of resources in sampling and supervisión, but these resources will pay for themselves in more
transparency and better targeting of beneficiaries.
In the third place, the virtue of self-targeting has the vice of low wages. In order to target resources
on needy groups and to avoid inducing distortions in local labor markets, wages in labor intensive public
works arefrequently set below the market wage of the relevant labor market. The literature on work-fare
in the developed world suggests that this targeting mechanism is not without costs in terms of stigmatizing
workers who particípate in the program (Lightman 1995), and in terms of political and social
Sesión 4 - 4

discrimination among workers by program administrators (Rose 1994). There is no solution to this
problem, short of raising wages to market levels, which in most cases will be impossible given resource
constraints.
In summary, labor intensive public works do not generate important labor market distortions to the
extent that they offer wages below the relevant market, and can provide a source of income to workers
temporarily unemployed. Their coverage depends on the amount of resources allocated to the program,
but there is no intrinsic reason why coverage of low skill workers could not be as ampie as needed to
reduce unemployment to the target level. This same property, however, brings ais to the problem of their
counter-cyclicallity. Because the amount of resources dedicated to the program is a political decisión there
is no way of guaranteeing that the program will move in sync with the economic cycle, expanding in
downtums and shrinking in upturns. In fact, the experience in the región shows that once the programs are
in place it is very difficult to reduce their size. In the well known cases of programs that were phased out
during the eighties (PEM and POJ in Chile and PAIT in Perú), the closing seems to have been mostly a
reaction to widespread problems of design and political manipulation (Graham (1995)).
2)

Wage subsidies

The experience of subsidizing prívate sector jobs is much less widespread than labor intensive
public works programs. Argentina is the only case where wage subsidies were widely used, and even there
the scope of these programs has shrunk recently due to criticisms from the unión movement.
Wage subsidies work through reducing the payroll tax and/or severance payments in employment
contracts for particular groups of workers (youth, women, ex-combatants, etc.). This characteristic makes
them suitable for the introductíon of moreflexible (or precarious) employment contracts in a process of
labor market regulation reform. In fact, this was the role these programs fulfilled in Argentina in 1995.
But at the same time, this makes them the center of a political debate on labor market flexibilization,
which in large measure explains why these programs were faced out in front of unión opposition in 1998.
On the other hand, because they are targeted on particular groups, they change the relative prices of
different types of workers in favor of the target group and induce large labor market distortions, not the
least of which is the substitution of subsidized by non-subsidized workers12/. In order to mitígate this
problem, there is normally an "additionality" requirement, by which subsidies are granted only for new net
hires that expand the payroll. In turn, this requires the determination of a baseline number of employees
and a control on new hires. Theoretically this is a task that Ministries of Labor fulfill in the normal course
of their business. In practice, the Ministries are extremely weak and have a very low enforcement
capability. This weakness makes impossible the task of determining base lines and controlling hires of
subsidized workers, therefore making worker substitution a widespread problem. As a consequence, is not
clear whether or not these programs really create more jobs than those that would have been created
without the subsidy.
In summary, these programs tend to generate large and important labor market distortions by
attempting to change the relative salaries of different types of workers. Because they have to be explicitly
targeted by design, they require a comprehensive and often non-existent enforcement apparatus, making
the problem of targeting the program an intractable one. In terms of their counter-cyclicallity, expanding
and shrinking the program requires an administrative decisión. To the extent that these programs are often
perceived as a mechanism to introduce moreflexible (or more precarious) employment contracts, they can

w

More formally, deadweight effects appear w e t e subsidized jobs would have been created anyway without the subsidy,
hn h
while substitution effects appear w e subsidized workers replace non-subsidized (Calmfors (1994)). The additionality
hn
requirement address t e deadweight effect, while substitution effects are only prevented at the margin.
h
Sesión 4 - 4

become the center of an often ardent political debate and make decisions about program implementation
politically very costly. This has been the experience of Argentina, where these programs were faced out
jointly with the rejection of more far reaching labor regulation reforms during 1997.

D.

Short term training programs

Short term training programs work as an income support device through the provisión of
scholarships to trainees during the classroom training and apprenticeship períods, normally between 4 and
6 months. The scholarships are below the relevant market wage, and the apprenticeships are developed in
privatefirms with which the training providers sign an agreement. The short duration of the classroom
training makes these programs more adequate to provide young new entrants to the labor market with job
search skills, than to the needs of skill updating or upgrading of workers displaced from declining sectors.
The main challenge in the design of these training programs arises from the existence of a national
training institution, normally a monopolistic public provider of trainingfinanced through a payroll tax
with no incentive whatsoever to adapt the nature of its activities and clientele to the challenges of high
unemployment. In order to circumvent this obstacle the programs are organized through the setting up of a
separate pool of resources managed by a specialized agent at the central government level. This agent in
turn bids out resources to private providers that execute the training programs in a decentralized fashion.
As we mentioned, these decentralized providers must enter into agreements with private sector firms to
ensure that trainees will have an apprenticeship stage, making privatefirms the effective gate-keepers of
the quality and relevance of the training programs. Another interesting by product of this process is the
development of stronger connections between firms and training providers, which make the latter
effectively providers of job search assistance services.
Training programs tend to be more expensive on a per beneficiary basis than labor intensive public
works, given that a larger part of the resources goes to pay the training provider. However, calculations of
benefits should include the long term change in the structure of the training system and the development
of job search assistance services, which are very large positive externalities of these programs.13'
The organization of the programs makes it easy for the program organizer to administratively target
groups of the population, and the programs have been quite successful in attracting unemployed youth.
However, it should be noted that the programs can be too effective in attracting the target group: in
México in 1996 youth participation rates increased so much that even if the employment rate of the group
rose, so did its unemployment rate. Although there is no formal proof that this was the result of the
expansión of training programs (particularly PROBECAT) in that year, there is a suggestive association
between expansión of these programs, decline in school enrollment rates, and increase in labor forcé
participation and employment of the target groups.
In summary, these programs tend to generate positive labor market externalities beyond the training
process itself, by easing the insertion of young workers and creating experience in the operation of labor
market intermediation mechanisms (job search assistance). In terms of coverage, the nature of the training
provided makes them suitable for unemployed youth and, as with any training program, one should not
expect them to create new jobs but rather to provide new entrants with some labor market experience.
Because youth unemployment is a permanent problem in the labor market, one should not think about
these programs as counter-cyclical devices, but rather as permanent features of a well-functioning labor
1/ T .
3 U
These emeigency training programs have created t e opportunity to introduce institutional innovation into a training system
h
characterized by the monopolistic power of payroll taxfinanced institutions. Disseminating these innovations to t e
h
mainstream vocational training system will m k it m c m r successful in addressing t e needs of skill upgrading of
ae
u h oe
h
workers caught in t e normal process of job churning.
h
Sesión 4 - 4

market intermediation system, that could be expanded and contracted following demand in a countercyclical way.

V.

Putting the pieces together

Macroeconomic volatility seems to be a permanent feature of the Latin American economies, and
the adoption of growth-oriented sustainable economic policies has not reduced it. In the brave new world
of higher and more pro-cyclical unemployment the traditional legally mandated severance payment
mechanisms have failed to provide the income support needed by unemployed and displaced workers.
These traditional mechanisms have become irrelevant in the new environment of more open and
competitive economies, given the narrow scope of its coverage. This same failure has led governments,
faced with shaip economic instability since 1995, to develop a series of attempts at the setting up of
mechanisms to support the incomes of groups of the population hurt by unemployment and declining
incomes.
The pieces of the new income support system are there in most countries. However, because they
were developed as emergency programs, they are more a mosaic of often unrelated initiatives than a
coherent and well thought system. As such, their ability to expand rapidly was a dominant concern. Little
attention was paid to the potential labor market distortions they introduced, and even less to their ability to
expand and contract in a counter-cyclical fashion. Moreover, the scarcity of resources available at the
moment of their implementation called for narrow targeting on the poor, as the programs were thought as
emergency devices to protect the population that did not have the resources to protect themselves. The
underlying thought was that the traditional income support mechanism based on severance payments
served to protect the non-poor.
The starting point for a new income support system is the recognition that workers in general, and
not only the poor, need to count with mechanisms to help them cope with both economy wide and
idiosyncratic shocks that result in income losses and unemployment. The protection of the severance
payment system covers today not more than half the labor forcé, and has a high cost in term of labor
market distortions. As has been argued extensively elsewhere (IPES 1997, for instance) more socialized
forms of income protection, as unemployment insurance, can provide effective protection and cost less in
terms of labor market distortions. In this sense, some of the recent labor market reforms (exemplified by
the Venezuelan reform of 1997), that reduce legally mandated severance payments and create a system of
unemployment insurance, represent a step in the right direction.
However, coverage of unemployment insurance will not be wide enough to protect all workers, and
particularly the poorer ones. Part of what differentiates the poor from the non-poor is the nature of their
labor market insertion and, therefore, the mechanisms that need to be devised to insure them against the
risk of income loss. Poorer, low productivity workers must relay on alternative mechanisms to protect
themselves as they cannot afford the cost of unemployment insurance, or are in employment situations
(self-employment or causal work in unregulated contracts) that make it unsuitable as insurance
mechanism. For those workers, a menú of altematives needs to be provided based on the existing
mechanisms on income support.
A veiy general visión of the components of an income support system is presented in Table 6. At
the center of the scheme, a well-designed unemployment insurance system covers the group of workers
that, given their employment contracts and their productivity, can "buy" the insurance. Both individual
accounts and collective insurance are needed in order to prevent moral hazard and to extend as much as
possible the coverage of the system. The rationale for the implementation of the collective insurance

Sesión 4 - 4

component is that workers that suffer morefrequent or more prolonged unemployment episodes can
"borrow" against future deposits. The Chilean Ministry of Labor proposed a good example of this type of
unemployment insurance in 1993 (see Cortazar et al. (1995)), and the recent reform of the social security
system in Venezuela contemplates an unemployment insurance system along the same lines.

Table 6
THE DESIGN OF AN INCOME SUPPORT SYSTEM
Program
Unemployment
insurance
Scholarships for
short-term classroom
training,
apprenticeships, and
job search assistance

Financing

Targeting

Institutional requirements

Non-targeted, covering all
workers in regulated contracts
(in lieu of severance
payments).

Financed from workers' and
firms' contributions.

Independent financial
institution(s).
System connected to pensión
system.

Unemployed youth

Financed from training system
payroil tax.

Opening of the market for
training services.
Network of non-government
providers.

VIH. Employment generation programs

A solid network of local
institutions able to apply
selection criteria and to
implement works' execution.

•
Labor intensive
public works

Self-targeting.

Financed from general
revenues, strongly countercyclical.

•

Administrative.

Exemption of payroil taxes,
financed by general revenues.

Sophisticated enforcement and
verification system from labor
authority.

Very narrow, based on family
income below the poverty line.

Financed from general
revenues.

Sophisticated targeting system.

Wage subsidies

Cash transfers

Mandatory-savings based schemes (the so called "separation funds" in Brazil, Colombia, Perú, and
Venezuela) mimic a number of the desirable characteristics of unemployment insurance, and can serve as
an alternative basis to protect the income of workers in regulated employment relations.
For workers that do not have access to unemployment insurance a variety of mechanisms need to be
put in place, depending on the reasons for their lack of access. In thefirst place, short term training
courses could provide low skill new entrants with the necessary training, labor market experience, and job
search assistance necessary to access a regular job. Scholarships should be set at a level low enough as to
not discourage search in the local labor market 14/ , and targeting should concéntrate resources on out of
school unemployed youth. Careful attention should be given to attract exclusively individuáis out of the
school system, and therefore to deter school desertion pulled by the program. Financing of this program
should be obtained from the existing payroil tax earmarked for training, maintaining the system of
decentralized provisión with apprenticeships. Funding should be provided in sync with demand for these
services, higher during economic downturns, lower during expansions. However, to the extent that the
training system unifies around the decentralized model of provisión, a continuous level of funding should
be provided to help all workers deal with normal job churning, while the youth component adjusts to the
economic cycle in function of the evolution of the youth unemployment rate.
For displaced low skill workers, labor intensive public works are the tools of choice in order to
provide them with jobs at the local level. As we mentioned above, in the design of these programs it will
be necessary to take into account the level of development of local govemments, and the nature of its
financial and political relationship with the central govemment. Many of initiatives to set up employment
141

The main criteria here is that the scholarship should be below the market wage that equivalent workers obtain in the labor
market, and not in function of the overall average wage.

Sesión 4 - 4

generation programs have failed for not taking into account the organizational and political dimensión of
the separation betweenfinancing and execution. Again, wage levels in the programs should be low enough
as not to crowd out alternative employment opportunities. Funding should be provided from general
revenues and not, as is normally done, from payroll tax revenue15'. Again, funding should be provided in
sync with demand, with expenditure adjusting according to the evolution of the general or, if available,
local unemployment rates.
The other member of the family of employment generation programs, wage subsidies, should be
used sparingly if at all, in spite of the conceptual attractive of the idea of generating prívate sector "real"
jobs (as opposed to the "make work" jobs of labor intensive public works). The distortions caused by
meddling with the relative wages of different types of workers are important enough as to counsel caution
in this area. Besides, the sophisticated enforcement and supervisión system needed to mitígate the
deadweight and substitution effects of these subsidies is not present in most of the countries.
Finally, cash transfers to poor families are the last resort mechanism to support those families that
"fall through" the other support mechanisms. As a mechanism of last resort a signal of good
implementation is low coverage. If coverage grows up to a well-defined poverty head count there is a
clear signal that other components of the income support system are failing, and too many workers are
"falling through". The complication of their design and implementation are well known16', and extreme
caution is recommended given that the program can create dependency traps for beneficiaries.
This set of programs and mechanisms centered on unemployment insurance, and encompassing
other complementary mechanisms to protect more vulnerable workers is a quite comprehensive system of
income support. As such, the system as described is more a statement of purpose than a guide for action.
Political realities and implementation abilities will determine what part of a labor market based income
protection system can be implemented in a given country at a given time. As a visión of purpose it could
be useful to put emergency measures into a more structured framework of long term objectives. Most of
the components of this comprehensive system exist in one form or another in most countries in the región.
Setting some of them (as replacing legally mandated severance payments by unemployment insurance or
saving-based separation funds) will require far-reaching political agreements, but these agreements can
only be facilitated by sharing a visión of the end product as a system to protect all workers against income
volatility. In other cases (as in designing emergency employment programs) the pitfalls of designing in the
context of an emergency can be to some extent mitigated by a visión of how the emergency program will
fit in the context of a new and more effective income protection system.

1S/
w

The use of payroll tax resources (which increase the cost of labor and therefore reduce employment generation) to generate
jobs is somewhat contradictory in terms.
For a summary of state-of-the-art arguments around cash transfers see Family allowances program, Fase II, Loan Proposal
(H00132), IADB, Project Report nr. 2354.

Sesión 4 - 4 1

Bibliography
Banuri,T. and EJ. Amadeo 1992. Worlds within the Thirld World: Labour market institutions in Asia and
Latin America. In Economic liberalization: no panacea, ed. T. Banuri, Oxford, Clarendon Press.
Besley, T. and S. Coate, 1995. The design of income maintenance programs. Review of Economic Studies,
62: 187-221.
Calmfors, L., 1994. Active labour market policy and unemployment - aframework for the analysis of
crucial design features. OECD Economic Studies, no.22, París.
Cortazar, R. et al., 1995. Hacia un nuevo diseño del sistema de protección a cesantes. Colección Estudios
CIEPLAN, nro. 40, Santiago, Chile.
Grosh, M., 1994. Administering targeted social programs in Latin America. Washington DC, World Bank
Regional and Sectoral Studies.
Ham, J. et al, 1998. Unemployment and the social safety net during transitions to a market economy:
evidence from the Czech and SlovakRepublics. American Economic Review, v.88,5: 1117-1142.
Hausmann, R. and M. Szekely, 1998. Inequality and the family in Latin America. OCE Working Paper,
Interamerican Development Bank, Washington DC.
Hoppenhayn, H. and J.P. Nicolini, 1999. Heterogeneity and optimal unemployment insurance, presented
at the Conference of social protection and poverty, Interamerican Development Bank, Washington
DC
IPES, 1997. Latin America after a decade of reforms, Interamerican Development Bank, Washington DC
Lightman, E.S., 1995. You can lead a horse to Water, but...: the case against workfare in Cañada. In
Helping the poor: a qualified case for workfare, Richards, J. et al., Toronto, C.D. Howe Institute.
Lustig, N. and M. Walton, 1998.
Washington, DC.

Crises and the poor: a template for action. Draft. IADB-WB,

Lora, E. and C. Pagés, 1997. La legislación laboral y el proceso de reformas estructurales de América
Latina y el Caribe, in Empleo y distribución del ingreso en América Latina, Hemos avanzado?,
Cardenas, M., Bogotá, Fedesarrollo
... and G. Márquez, 1998. The employment problem in Latin America: perceptions and stylized facts.
Interamerican Development Bank, Washington DC
... and M. Olivera, 1998. Macro policy and employment problems in Latin America. Interamerican
Development Bank, Washington DC
Márquez, G., 1997. Protección al empleo y funcionamiento del mercado de trabajo: una aproximación
comparativa, in Empleo, Flexibilidad laboral y protección social, proceedings of the II Technical
Meeting, Círculo de Montevideo, Montevideo, Uruguay, UNDP.
... and C. Pages, 1998. Ties that bind: employment protection and labor market outcomes in Latin
America. Interamerican Development Bank, Washington DC
Mazza, Jacqueline, 1999. Unemployment Insurance: Case Studies and Lessons for the Latin American and
Caribbean Región, Technical Study RE2/S02, Interamerican Development Bank, Washington DC
(forthcoming)
Ravallion, M., 1998 Apprasising workfare programs, Technical Study, SDS/POV, Interamerican
Development Bank, Washington DC.

Sesión 4 - 4

1999. Is more targeting consistent with less spending?, International Tax and public finance
(forthcoming).
1999. Monitoring targeting performance when decentralized allocations to the poor are observed,
The World Bank, Washington, DC.
... and J. Jalan, 1999. Income gains from workfare and their distribution, The World Bank, Washington
DC.
Rose, N.E., 1994. Put to work. Relief programs in the Great Depression. New York: Monthly Review
Press.
Verdera, F. 1998. Análisis comparativo de los programas de empleo e ingresos en América Latina y el
Caribe. In Programas de empleo e ingreso en América Latina y el Caribe, eds. Márquez, G. and D.
Martínez, Lima, Perú, Banco Interamericano de Desarrollo and Organización Internacional del
Trabajo.

Sesión 4 - 4

Annex
Employment

Table
generation

A-1
programs

Argentina

in 7 countries

Beneficiaries
% of total
1.000s
labor forcé
9.31%

in the región

Expenditure

10 6 USS

% GDP
0.09%

Public worksfinanced with public resources
1. Programa de Asistencia Solidaria (PROAS)
Unemployed household heads ¡n public works executed jointly by
Sec. Desarrollo Social and state Govemments
2. Programa de Entrenamiento Ocupacional (PREÑO)
3. Programa de Servicio Comunitario (ASISTIR)
Female household heads in community development activities
4. Programa Trabajar
Unemployed household heads in public works executed by local
govemments and NGOs.

260.0

2.7%

54.5

0.020%

94.0

1.0%

20.0

0.007%

25.0

0.3%

2.6

0.001%

233.0

2.4%

44.9

0.017%

254.0

2.7%

42.4

0.016%

21.0

0.2%

4.4

0.002%

Prívate sector employment promotion

5. Programa de Empleo Privado para Pequeñas y Medianas
Empresas
Subsidy for new jobs for unemployed workers in firms with less
than 100 employees
6. Programa Nacional de Forestación Intensiva (FORESTAR)
Subsidy for new jobs for unemployed workers in new
agricultural/forestry firms.
7. Programa de Reinserción Laboral
Subsidy to workers that find a job while receiving unemployment
insurance.
8. Programa de Movilidad Geográfica
Subsidy to workers that have to move from place of residence to
keep the job.
9. Bono para la creación de empleo privado (BOCEP)
Fiscal credit for workers displaced from states' payroil. New
employer can use as collateral for credit from public banks.

IX. BRASIL

(1)

1. Programa de Generación de Empleo e Ingresos (PROGER)
Special credit lines to MSMEs, cooperatives and informal sector.

X.

CHILE

COSTA

n.a.

n.a.

n.a.

5.2

0.1%

73.4

0.027%

221.8

0.49%

1,188.8

0.21%

221.8

0.5%

1,188.8

0.21%

XI.
0.
10%

4.3

1. Trabajadoras temporeras
Child-care and educational sen/ices for children of ag. Temporary
workers
2. Programa de desarrollo del microempresario indígena
Strengthening of economic networks of indigenous groups
through ME creation and support

XIV.

n.a.

RICA

4.3

0.1%

n.a.
XV.

1

4

1.

1.2

XIII. 0.0
0%
0.002%

0.2
8.

XVI. 0.
71%

1. Programa nacional de generación de empleo
Transfer of a min. wage to unemployed workers who particípate in
2.1
0.2%
construction of social services infrastructure and sen/ice delivery.
2. Pro Trabajo
- Incentivos
para la reinserción
laboral
y el empleo
temporal
Subsidy of 50% of min. wage for on-the-job-training for
3.4
0.3%
unemployed/vulnerable workers.
- Ideas productivas
0.2%
2.6
Support to ME creation.

Sesión 4 - 434

XII.

XVII. 3.
3

XVIII. 0.0
4%

0.1

0.001%

2.1

0.026%

1.1

0.013%

Beneficiaries

Expenditure

, 0 t ° f t ° t a l 10 6 US$
labor forcé
XXII. 21
XX. 6.
XXI. 0.
0
61%
.2
1.000.

XIX.

JAMAICA

1. Micro Investment Development Agency (MIDA)
Credit for ME development
2. The Government of Jamaica/Government
of the
Netherlands Micro Enterprise Project (GoJ/GoN MEP)
Credit for ME development
3. The Government of Jamaica/European Union Programme
Credit for ME development
4. Mel Nathan Institute for Development and Social Research
(MMI)
Community development services
5. Enterprise Development Trust (EDT)
Credit for ME development
6. The Women's Construction Collective (WCC)
Training and credit for female construction workers
7. ASSIST Ltd.
Credit for ME development
8. Bee Keeping and Honey Bee Project
Training and employment for youth in bee-keeping activities.
9. SESP
Training and temporary employment for unemployed workers

XXIV.

MEXICO

2. Programa de Conservación de Caminos Rurales
Rural public works for unemployed youth, Federal Govt. financed,
works organized by State and local govts.
3. Progr. de Construcción de Infraestructura Física y Obras
de Empleo Productivo
Social infrastructure public works for unemployed youth, Federal
Govt. financed, works organized by State and local govts.
4. Programas Sociales Privados Club de Leones y Rotarios

XXIX.

PERU

(2)

1. Progr. de autoempleo y microempresa (PRODAME)
Training and credit for ME creation and support
1. FONCODES
Social investment fund builds small public works using local workforce.

6.0

0.6%

7.6

n.a.

% GDP
XXIII. 0.5
0%
0.181%
0.000%

n.a.

1.4

0.034%

n.a.

1.6

0.038%

0.2

0.004%

0.0

0.0%

n.a.

n.a.

0.1

n.a.
XXV. 1,
024.0

XXVI. 4.
42%

0.007%

10.0

n.a.

0.002%

0.3

n.a.

0.237%

XXVII. 1,
802.0

XXVIII. 0.
51%

712.0

3.1%

350.0

0.099%

312.0

1.3%

1,452.0

0.410%

n.a.
XXX. 27
.8

n.a.
XXXI. 0.
93%

n.a.
XXXII. 0.
1

n.a.
XXXIII. 0.
19%

4.2

0.1%

0.1

0.000%

23.6

0.8%

100.0

0.002%

(1) PROEMPREGO is excluded from the Brazilian list of employment generation programs, as it is an investment program, with
obvious employment consequences, but with the primary objective of improvements in sanitation, environmental infrastructure, urban
transport, etc. through BNDES lines of credit.
(2) Perú has also implemented a number of its employment generation programs as labor intensive investments sub-projects,
complementary to the normal investment activities of institutions such as Instituto Nacional de Desarrollo (INADE), Programa
Nacional de Asistencia Alimentaria (PRONAA), Empresa Nacional de Edificaciones (ENACE), Fondo Nacional de la Vivienda
(FONAVI), Instituto Nacional de Infraestructura Educativa y de Salud (INFES), SEDAPAL, CORDECALLAO, CORDELIMA, INABIF,
Fondo de Compensación Municipal, PROMANACHCS (Min. Agricultura), Ministerio de Transporte.

Sesión 4 - 4

Training

Table
programs

A-2
in 7 countries

in the región

XXXIV. EXP
ENDITURE
% total labor forcé 106 us$
% GDP
1.4%
95.6
0.04%

Beneficiaries
Argentina
1. Proyecto Joven
Scholarships and stage in temporary job for low-income,
non-skilled, unemployed youth.
2. Proyecto de Microemprendimiento
Entrepreneurship training for experienced, unskilled
workers
3. Programa Imagen (Orientación para el empleo)
Job-search assistance
4. Programa de Talleres Ocupacionales (PTO)
Support to NGOs on setting up training institutions
5. Programa de Capacitación Ocupacional
Training of unemployed and SMEs personnel
6. Programa de Capacitación para el Empleo
Scholarships and stage in temporary job for low-income,
non-skilled, unemployed and displaced workers
7. Programa Aprender
Financing of health and accident insurance for young
workers hired under Contratos de Aprendizaje
8. Programa Emprender
Financing of training cost for workers in new firms.
9. Programa de Crédito Fiscal
Tax exception for training firms.

lO O
.Os
133.0
53.0

0.6%

71.7

0.027%

5.4

0.1%

6.5

0.002%

27.0

0.3%

1.2

0.000%

1.
80

0.2%

4.2

0.002%

24.0

0.3%

7.3

0.003%

17
.

0.0%

2.3

0.001%

19
.

0.0%

-

2.0

0.0%

2.4

0.001%

n o
ii» a .

1.6%

310.2

0.06%

340.8

0.7%

1 98
4.

0.03%

399.7

0.9%

1 94
5.

0.03%

36.6

Brazil

n.a.
740.5

0.8%

1.
83

0.03%

1.
79

0.4%

1.
04

0.019%

1.
50

0.3%

4.9

0.009%

0.1

0.0%

0.3

0.001%

0.2

0.0%

0.8

0.002%

13
.

0.0%

0.3

0.001%

XXXV. PLANFOR

FAT-financed training program executed at t e federal
h
and state level by independent training institutions.
1. Programas Federales y Estatales

Federal and state programs for vulnerable groups.

2. Programas de Emergencia

Emergencies from drought and declining/restructuring
sectors

Chile

1. Programa Chile joven
Stipend and stage in temporary job for low-income, nonskilled, unemployed youth.
2. Programa de apoyo a mujeres jefas de hogar de
escasos recursos
Training, day-care, health and other services to improve
labor market insertion of poor women
3. Mujer y microempresa (Capacitación e gestión
n
empresarial con perspectiva de género)
Entrepreneurship training for female household h a s
ed
with some education
4. Programa de reinserción laboral
Job search and relocation assistance to displaced workers
f o carbón and textile workers.
rm
S. Programas regulares de becas.
Scholarships for training at official institutions for
vulnerable groups (temporary workers in ag., ports a d
n
fishing)

Sesión 4 - 4

Beneficiaries

XXXIV. EXP
ENDITURE
106 us$
% GDP

l.OOOs

% total labor forcé

0.1

0.0%

0.1

0.000%

0.6

0.0%

0.9

0.002%

0.2

0.0%

0.1

0.000%

1.2

0.0%

0.4

0.001%

Costa Rica

13.1

1.2%

60.6

0.73%

Contracting out of training activities by the public training
institution (INA)

n.a.

6. Proyecto apoyo a los progr. de inserción laboral
para personas con discapacidad

Policy formulation and pilot program for labor market
insertion of handicapped workers

7. Programa de capacitación e inserción laboral para
personas con discapacidad.
8. Programa de rehabilitación, capacitación e inserción
laboral para personas discapacitadas.
9. Programa de formación y capacitación para el
trabajo.
Adult training program privately operated.

3. Llave en mano.

n.a.

5. Formación y reconversión para los movilizados

1.5

0.1%

6. Talleres públicos

6.2

0.6%

7. Formación profesional para el desarrollo socio
laboral de personas con discapacidad

1.0

0.1%

n.a.

S. Prog. de becas de capacitación para el empleo del
Ministerio de Trabajo y Seguridad Social

4.4

0.4%

0.1

Jamaica

43.5

4.4%

18.6

0.44%

Training for out-of-school and unskilled unemployed
workers

40.0

4.1%

11.4

0.3%

0.6

0.1%

4.6

0.1%

Training for displaced public sector workers

Training of low-income workers in marginal urban areas

Training of handicapped workers

Scholarships for training of workers with secondary
education.

1. Skills 2 0
00

2. Special Training Empowerment Programme (STEP)
Youth training

n.a.
n.3.

3. Strategies to Rehabilítate Inner Cities through
Viable Enterprises (VIABLE)

n.a.

4. National Youth Service (NYS)

2.9

0.3%

2.6

0.06%

México

410.3

1.8%

135

0.04%

1. Programa de Becas de Capacitación para
Desempleados (PROBECAT)
Training and scholarships for unemployed workers

410.3

1.8%

135

0.038%

Peni

1.5

0.1%

5.0

0.01%

1.5

0.1%

5.0

0.01%

Urban youth training

Training temporary employment for unemployed youth

n.a.

1. Programa de Capacitación Laboral Juvenil
(PROJOVEN)
Scholarships and training for unemployed youth

Sesión 4 - 4