Citizens, Politicians, and Providerssiteresources.worldbank.org/BRAZILINPOREXTN/...1.1 A Stylized Presentation of Accountability Relationships 3 2.1 Secondary Education Coverage for

Citizens, Politicians, and Providers

FMCitizens.qxd 2/16/05 12:39 PM Page i

FMCitizens.qxd 2/16/05 12:39 PM Page ii

Citizens, Politicians, and Providers: The Latin AmericanExperience with ServiceDelivery Reform

Ariel Fiszbein, Editor

FMCitizens.qxd 2/16/05 12:39 PM Page iii

© 2005 The International Bank for Reconstruction and Development / The World Bank1818 H Street, NWWashington, DC 20433Telephone: 202-473-1000Internet: www.worldbank.orgE-mail: [email protected]

All rights reserved

1 2 3 4 08 07 06 05

The findings, interpretations, and conclusions expressed herein are those of the author(s)and do not necessarily reflect the views of the Executive Directors of the International Bankfor Reconstruction and Development / The World Bank or the governments they represent.

The World Bank does not guarantee the accuracy of the data included in this work. Theboundaries, colors, denominations, and other information shown on any map in this workdo not imply any judgement on the part of The World Bank concerning the legal status ofany territory or the endorsement or acceptance of such boundaries.

Rights and PermissionsThe material in this publication is copyrighted. Copying and/or transmitting portions or allof this work without permission may be a violation of applicable law. The InternationalBank for Reconstruction and Development / The World Bank encourages dissemination ofits work and will normally grant permission to reproduce portions of the work promptly.

For permission to photocopy or reprint any part of this work, please send a request withcomplete information to the Copyright Clearance Center Inc., 222 Rosewood Drive, Danvers,MA 01923, USA; telephone: 978-750-8400; fax: 978-750-4470; Internet: www.copyright.com.

All other queries on rights and licenses, including subsidiary rights, should be addressedto the Office of the Publisher, The World Bank, 1818 H Street, NW, Washington, DC 20433,USA; fax: 202-522-2422; e-mail: [email protected].

ISBN 0-8213-6089-2 978-8213-6089-7e-ISBN 0-8213-6090-6

Library Cataloging-in-Publication data has been applied for.

FMCitizens.qxd 2/16/05 12:39 PM Page iv

Contents

v

Preface viiAcknowledgments viiiAbbreviations ix

1 Introduction 1Service Delivery Progress 1

Latin American Political and Institutional Context 4

Notes 6

2 Service Delivery and Social Outcomes: A Story of Successes and Failures 7Progress in Basic Service Coverage 7

Service Quality 11

Social Outcomes 12

Notes 14

3 LAC Service Delivery: Assessing Two Decades of Change 15Reforming the Compact 15

Enhancing Client Power 29

Enhancing State Accountability to Citizens 36

Notes 45

4 Conclusions 49Routes to Better Service Delivery 51

The Centrality of the “Fitting Process” 57

References 63

FMCitizens.qxd 2/16/05 12:39 PM Page v

vi CONTENTS

Boxes1.1 A Glossary of Terms for a Service Delivery Accountability Framework 3

3.1 Macro vs. Sectoral Reforms 16

3.2 Public Financial Management Reform in Guatemala 17

3.3 Haiti: Public Service Delivery in a Failed State 18

3.4 Lack of Civil Service Reform as a Health Reform Bottleneck 19

3.5 Executive Agencies in Jamaica 26

3.6 Higher Education Competition: Chile’s Reforms in the 1980s 30

3.7 User Fees 31

3.8 The Role of Community Participation and Cultural Adaptation in Service Delivery 35

3.9 Does Democracy Lead to More Social Spending? 37

3.10 ¿Cómo Vamos? Report Cards, Scorecards, and Citizen Monitoring of Service Quality 42

3.11 Opening Up to Social Accountability: Peru, 2001–04 43

4.1 Lessons for Donors and International Organizations 50

4.2 The Role of Evaluation 61

Figures1.1 A Stylized Presentation of Accountability Relationships 3

2.1 Secondary Education Coverage for Various LAC Countries, 1970 and 1999 8

2.2 Evolution of Water and Electricity Access in Selected Latin American Countries, by IncomeDecile, 1986–96 9

2.3 School Enrollment for Children 6–12 Years Old: Relative Gaps between Income Groups 9

2.4 Tertiary Education Enrollment Differences in Selected Countries 10

2.5 Percentage of Students Reaching Expected Levels, by Region and Language 11

2.6 Key Social Outcomes: Deviations from Expected Values Given Income 13

2.7 Inequality in Social Outcomes (Health): Bolivia, 1997, Peru and Nicaragua, 2000 14

3.1 Accountability Mechanisms in Presidential Democratic Systems 39

4.1 A Network of Influence and Accountability Mechanisms 49

Table4.1 Fitting Approaches to Country Conditions 59

FMCitizens.qxd 2/16/05 12:39 PM Page vi

Children regularly receiving health visits and education, the sick receiving proper andtimely health care, safe water flowing out of the tap, electricity reliably reaching homes andbusinesses—these apparently simple events are taken for granted in developed countries. InLatin America, despite two decades of social and infrastructure improvements, the poor andmany of the middle class make do with low-quality services. Far too many of the poorreceive no services. Improving service delivery to the poor is both a widespread politicaldemand and central to the realization of the Millennium Development Goals (MDGs).

This book interprets service delivery successes and failures in Latin America and pro-vides guidance to policymakers and development practitioners on shaping public action toprovide better-quality services for all. Its analysis builds on the accountability frameworkdeveloped in the World Bank’s World Development Report 2004: Making Services Work forPoor People, which emphasizes the behavior of people—from teachers to administrators,politicians, and rich and poor citizens—within the chain of interactions from demand toactual service delivery.

The report seeks to answer an essential question: If accountability relationships amongcitizens, policymakers, and service providers are key to effective service delivery and therehave been both systemic reforms (expanding national and local democracy) and an array ofspecific experiments (privatization, increased choice), why is service delivery in Latin Amer-ica still so inequitable and often of low quality?

Preface

vii

FMCitizens.qxd 2/16/05 12:39 PM Page vii

A World Bank staff team prepared this report. Led by Team Leader Ariel Fiszbein, the teamincluded, in alphabetical order, Cristian Baeza (health), Emanuela Di Gropello (education),Vivien Foster (infrastructure services), Jonas Frank (decentralization), Yasuhiko Matsuda(state reform), Laura Rawlings (social assistance), Mauricio Santamaria (service deliveryand social outcomes), Michael Walton (voice and empowerment), and David Warren (socialfunds).

Contributing to the report were, Kathy Bain, Lisa Bhansali, Jack Fiedler, Gillette Hall,Linn Hammergren, Amy Kirkley, Jerry La Forgia, Heather Marie Layton, EdgardoMosqueira, Harry Patrinos, William Reuben, Keta Ruiz, Mario Sangines, Jennifer Sara, RobySenderowitsch, and Joseph Shapiro.

The report benefited from comments from, Joachim von Amsberg, Ana-MaríaArriagada, Shanta Devarajan, Ron Myers, Guillermo Perry, Fernando Rojas, Elena Serrano,Luis Servén, and Eduardo Velez. Tomas Casas was responsible for the production of thereport.

Acknowledgments

viii

FMCitizens.qxd 2/16/05 12:39 PM Page viii

ix

ADERASA Asociación de Entes Reguladores de AguaPotable y Saneamiento de las Américas

AIN-C Comprehensive Attention to Childhood inthe Community program (Honduras)

CCT conditional cash transfer CEO chief executive officerCEPAL Comisión Económica para América Latina y

el Caribe (Economic Commission for LatinAmerica and the Caribbean)

CLAS local health administration committees(Peru)

COFOPRI Comisión para la Formalización de laPropiedad Informal (Peru)

EA executive agencies EAA Executive Agencies Act EDUCO Educación con Participación de la

ComunidadFONASA National Health Fund (Chile)FONCODES Fondo de Compensación y Desarrollo Social

(Peru)

IFMS Guatemala Integrated Financial ManagementSystem

ISAPRES a private health insurer (Chile)IT information technology LAC Latin America and the Caribbean regionLAL Latin American LaboratoryMDGs Millennium Development Goals NGO nongovernmental organizationOECD Organisation for Economic Co-operation

and DevelopmentPISA Program for International Student

AssessmentPPP purchasing power paritySIMCE standardized school scoresTIMSS Trends in International Mathematics and

Science StudyUNDP United Nations Development ProgrammeWDI World Development IndicatorsWDR World Development Report

Abbreviations

Note: All dollar amounts are U.S. dollars unless otherwise indicated.

FMCitizens.qxd 2/16/05 12:39 PM Page ix

FMCitizens.qxd 2/16/05 12:39 PM Page x

1

Introduction

1

Children going to school and learning, par-ents taking their children for regular healthvisits, the sick receiving proper and timelyhealth care, safe water flowing out of the tap,reliable electricity supplies—these appar-ently simple events are taken for granted indeveloped countries. However, in most ofLatin America (as in most developing coun-tries), only the privileged have reliable andquality access to such services, and theyoften achieve this by using high-cost privatesolutions. The poor and many of the middleclasses suffer low quality services, and, inthe poorer countries and regions, often noservice at all. How come? While technicaldifficulties may still be a binding constraintfor some sophisticated services, they areclearly not a bottleneck for the most essen-tial ones; it is not primarily a technicalproblem when children systematically failto learn to read and write or a communitydoes not receive reliable supplies of safedrinking water.

This report seeks to answer this ques-tion. In so doing, it aims to provide guid-ance as well to policy makers and develop-ment practitioners on how to shape publicaction to get better quality services for all.This challenge is both a widespread politicaldemand throughout the region and centralto the realization of the Millennium Devel-opment Goals (MDGs). This report’s majortheme is that to understand questions ofservice access and quality, we need to exam-ine people’s behaviors, from teachers toadministrators, politicians, and rich andpoor citizens. The big issue is whether thoseresponsible for designing and deliveringservices are accountable to the citizens whoare demanding the services—and also pay-ing the taxes and fees that finance services.This is fundamentally about relationships

between groups of people. But these rela-tionships are embedded in formal andinformal institutional structures. Influencepatterns on service delivery can workthrough two broad pathways: first, what istermed the “long route” from citizens topoliticians to service delivery agencies toservice providers; and second, the “shortroute” of direct interactions between indi-viduals, communities, and front-line work-ers. These considerations apply to somedegree to all public action areas—frommacroeconomic policy to policing. But theyare particularly important to service deliv-ery areas that involve intensive transactionsand discretionary behavior of providers—features typical of the full panoply of basicservices that are central to the lives of citi-zens, from schooling to water supply.

In this introduction we outline servicedelivery progress and introduce the prismthrough which we believe service deliveryshould be interpreted, particularly in theLatin American political and institutionalcontext. We then provide an interpretationof how this context can help understand thecurrent state of service delivery and sketchthe public action implications. The ideasand results are developed at greater lengthin the body of the report.

Service Delivery ProgressLatin America’s political and economic his-tory is associated with a tradition ofunequal service delivery, with the poor typ-ically suffering from a severe lack of decentquality services. Conditions have been bet-ter in the relatively richer societies, notablyin the Southern Cone and to a certainextent in Mexico. They are typically worsein poorer and more socially fractured soci-eties, such as Bolivia and Guatemala. They

c h a p t e r

01Citizens.qxd 2/15/05 3:46 PM Page 1

are worst in Haiti, the region’s poorest andinstitutionally weakest society. In particular,the 1980s debt crisis hurt progress in publicprovisioning in most countries in theregion.

However, there has been significantprogress in the past decade or so; chapter 2reviews this progress. Progress is most clearlyseen in service coverage, notably in educa-tion, health, and basic infrastructure waterand electricity services. These gains wereassociated with expanded public spending forsocial services.

CEPAL [Comisión Económica paraAmérica Latina y el Caribe (Economic Com-mission for Latin America and theCaribbean)] (2001) estimates that during the1990s, total per capita social spendingincreased by about 50 percent, if pensions areincluded, and 30 percent if they are not. Bycontrast, during the same period, publicinfrastructure services expenditures fell from1.2 percent of GDP to 0.6 percent of GDP,declines that were offset by surging privateinvestment in the middle of the decade, butnot since.

In the poorer countries, an importantagenda of increased coverage remains. Tovarying degrees, all countries suffer prob-lems of both inequality and quality. Withrespect to inequality, in Mexico, the averageperson in the bottom quintile of the incomedistribution has 3.5 years of schooling,compared with 11.6 years for a person inthe top quintile. In Brazil, a child born to ahousehold in the poorest 20 percent of thedistribution is three times as likely to diebefore reaching the age of five as a childborn into a household in the top fifth. AGuatemalan family whose income is in thebottom 20 percent of the distribution has a57 percent chance of being connected to thewater mains and a 49 percent probability ofhaving access to electricity, while a house-hold in the top 20 percent of the distribu-tion enjoys rates of 92 and 93 percent,respectively. Throughout the region, thebulk of social protection spending goes toformal social security systems that primar-ily benefit those in the upper part of theincome distribution. Such inequalities areparticularly noteworthy between indige-nous and nonindigenous people across

Latin America (Hall and Patrinos, forth-coming) and between Afro-Latinos andothers.

With respect to quality, the educationsector has specific evidence of low testscores at least relative to countries of theOrganisation for Economic Co-operationand Development (OECD). There are wide-spread reports of dissatisfaction with ser-vice delivery, especially from the poor,notably from country studies in the regionon the Voices of the Poor (Chambers et al.2000).

To interpret the Latin American experi-ence, this report builds on the frameworkdeveloped in the 2004 World DevelopmentReport (WDR). The WDR sought answersfor observed service delivery failures world-wide. It developed a framework thatemphasizes people’s behavior within thechain of interactions from demands of citi-zens to actual service delivery—whetherthis is by public, private, or nongovernmen-tal organizations (NGOs). For example, forchildren to learn how to write, there needsto be some type of school, the children needto be present, teachers need to show up andbe paid, and the teachers need to have theskills, resources, and motivation required togenuinely teach. Behind this there are awhole set of interactions from policy andbudgetary decisions, through the perfor-mance of ministries of education to recruit-ment and the influence of mayors and com-munities. When any link in the chain breaks(i.e., children do not come to school, theteacher is not present, there are no inputs,and so on), the entire process may fail.Whether a system works depends on howresponsibilities are assigned, what capabili-ties those in charge actually have, whataccountability mechanisms exist for thosein charge, and what incentives are in placeto make the accountability mechanismseffective in practice. All these factors tend tobe deeply embedded in social and politicalrelations.

The WDR conceptualized an approachthat focuses on accountability mecha-nisms and power relations between policymakers, providers, and citizens/users. Theapproach emphasizes the role of three suchrelationships:

2 Citizens, Politicians, and Providers: The Latin American Experience with Service Delivery Reform


1. The “political voice” of different citizengroups over policy makers in shapingpolicy design;

2. The “compact” between policy makersand service providers, whether they bepublic, private, or nongovernmental; and

3. The “client power” (either through theexercise of a more direct form of voice orthrough choice) of citizens vis-à-vis theservice providers.

Figure 1.1 illustrates these relationships,and box 1.1 provides a glossary of terms rel-evant to this framework.

Within such a framework there are twoinfluence pathways for service delivery forthe poor (or nonpoor). The so-called “longroute” involves the influence of citizens on

Introduction 3

PUBLIC, PRIVATE,CIVIC PROVIDERS;

FRONT-LINE WORKERS

CITIZENS, RICHAND POOR, POWERFUL

AND WEAKclient power

POLITICIANS ANDPOLICYMAKERS

voice/politics

compact/ policies

Figure 1.1 A Stylized Presentation of Accountability Relationships

Source: World Bank 2003a.

Actors: Individuals, households, communities,firms, governments, and other public or non-governmental organizations that finance,produce, regulate, deliver, or consume services.

Accountability: A set of relationships amongservice delivery actors with five features: theunderstanding that a service or goodsembodying a service will be supplied (dele-gating); paying for the service (financing);supplying the service (performing); obtain-ing relevant information and evaluating per-formance against expectations; and beingable to impose sanctions for inappropriateperformance or provide rewards for appro-priate performance (enforcing).

Client power: The accountability relationshipconnecting clients to the front-line serviceproviders, usually at the point of servicedelivery, based on interactions throughwhich clients monitor suppliers and expresstheir service demands.

Clients/citizens: Citizens participate in politicalprocesses to shape and attain goals. Clientsreceive services to satisfy their householddemands. Individuals can act as both clientsand citizens

Clientelism: The tendency of politicians aspatrons to respond to political competitionby excessively favoring one group of clients over another in return for a politicaladvantage. Politicians often practice clien-telism by providing narrow supportergroups with free public services or public employment.

Compacts: The broad, long-term accountabilityrelationship that connects policy makers toorganizational providers.This relationshipcan, but need not, take the form of acontract.

Long and short routes of accountability:Clients may hold service providers account-able for performance in two ways. First, theshort route connects clients and providersdirectly.When such direct client power isweak or not possible to use, clients must usevoice and voting as citizens to hold politi-cians accountable. In turn, politicians andpolicy makers must use their compacts withservice providers to hold those providersaccountable.The long route of accountabil-ity combines the actions of clients pressur-ing politicians and policy makers/politicianspressuring providers.

Organizational providers: Public, private, andnonprofit entities that provide services.These can range from a government min-istry with thousands of employees to a com-munity-run school serving a village.

Politicians and policy makers: The servicedelivery actors that discharge the state’s leg-islative and regulatory responsibilities. Politi-cians may be elected or may reach theirpositions through undemocratic means.They can also be policy makers—for exam-ple, a general who is president, but also runsthe military. Policy makers are morecommonly the highest unelected officials,either from the civil service or appointed.Politicians set the general policy direction,

while policy makers implement those direc-tions and also set and enforce theconditions under which service providersoperate. Accountability subrelationshipsbetween politicians and policy makers areusually derived from the constitution oradministrative law.

Service delivery framework or chain: The fourservice-related actors—citizens/clients,politicians/policy makers, organizationalproviders, and front-line professionals—andthe four relationships of accountability thatconnect them:

• Voice and politics: Connects citizens andpoliticians

• Compacts: Connects politicians/policymakers and providers.

• Management: Connects provider organiza-tions with front-line professionals

• Client power: Connects clients withproviders.

Voice and politics: This is the most complexaccountability relationship. It connects citi-zens and politicians and includes formal andinformal processes such as voting, electoralpolitics, lobbying, propaganda, patronage,clientelism, media activities, informationaccess, and others. Citizens delegate topoliticians the functions of serving citizeninterests and financing governmentthrough taxes. Politicians perform by provid-ing services. Citizens enforce accountabilitythrough the short and long routes discussedabove.

B O X 1 . 1 A Glossary of Terms for a Service Delivery Accountability Framework


policy makers—working through politicalinstitutions—who then provide the policyand resources for providers (organizationsand individuals) to run services, withimportant relationships of “management”within providers, especially with respect toinfluences on front-line workers’ behavior.When such a long route works, it is theclassic model of an effective Weberianbureaucracy that is ultimately accountableto citizenry through an elective democ-racy. The so-called “short-route” involvesthe more direct influence of the ultimatebeneficiaries, when clients directly controlhow services are delivered or have theeffective capacity to choose from whomthey get services.

Such a framework helps understand ser-vice delivery problems. When governmentsdo not feel the pressure to respond to citi-zen demands, when they are incapable ofenforcing basic performance rules on thosedirectly responsible for service delivery, andwhen citizens have no control or choiceover service providers, typically services willfail. Unless there is an enabling environ-ment (one that provides opportunities for,rewards and recognizes “good” behavior,that supports those that try, that controlsabuse of various types, that learns frommistakes, and has the chance to adapt),individual effort by committed providersand users will go to waste. The (public pol-icy) trick is to help establish such anenabling environment.

Latin American Political and Institutional ContextWithin this type of framework for inter-preting service delivery, it is fundamental tounderstand the broader political and insti-tutional context where delivery mecha-nisms operate. And the last two decadeshave seen a deep transformation in LatinAmerica and Caribbean (LAC) countries.First, the region has moved from being a setof countries where democracy was theexception, to one in which it is the rule. In1980, a majority of LAC countries did nothave democratically elected governments.Today, most of these countries are politicaldemocracies with national authoritieselected in free and competitive elections,

with freedom of expression and association.The shift from authoritarian to constitu-tional rule and more competitive electoralsystems constitutes a major development,particularly when seen in historical per-spective. The spread of democracy has pro-vided a large number of citizens, hithertoexcluded from the political decision-mak-ing processes, with mechanisms throughwhich they can voice their demands for bet-ter services, though whether governmentwill actually respond to these demands isstill subject to a range of political and socialfactors that often limit actual exercise of cit-izens’ rights.

Second, governments in the LAC regionhave become significantly more decentral-ized in political, administrative, and fiscalterms. This is best understood as a particu-lar dimension of democratization. Thenumber of popularly elected subnationalgovernments has expanded throughout theregion, achieving what has been called a“silent revolution” (Campbell 2003). Thesegovernments have acquired growing servicedelivery responsibilities and channel alarger share of total public spending on keyservices’ provision. This has also entailed agrowing transfer of revenue authority tolocal levels, in the form of transfers, taxes,or access to debt.

Third, there was a process of redefiningthe state’s role that involved increased con-tracting out to private and nongovernmentservice providers, wholesale privatization(especially in the infrastructure sectors),and more generally, a movement towardenhancing the role of competition in manyeconomic activities, including the provisionof key services.

Fourth, there has been rising activity bycivil society actors, from NGOs to identity-based movements (notably amongst indige-nous groups), both in the delivery of ser-vices1 and as actors seeking differentpolicies. This has sometimes been associ-ated with efforts by national or local gov-ernments to open up government decision-making to citizens and civic organizations.

Interpreting service delivery successes andfailures in Latin America. If accountabilityrelationships between citizens, policy mak-



ers, and service providers are key to effectiveservice delivery, and there have been bothsystemic reforms (expanding national andlocal democracy) and an array of specificexperimentation (privatization, increasedchoice), why is the current service deliverycondition still problematic, especially interms of inequality and quality? Chapter 3systematically explores this issue.

The general message is that institutionalreform is a long-term endeavor, and thereare no magic bullets. The political and insti-tutional changes noted above are importantand are broadly positive for service delivery.However, history has bequeathed the regionweak and unequal formal and informalinstitutions. While conditions vary consid-erably across countries, institutional struc-tures are often clientelistic and, especially incountries with ethnic and social cleavages,socially exclusive. Throughout the region,the poor are typically disempowered.Greater formal democracy and decentral-ization have opened political opportunities,but procedural democracy is only one ele-ment in the shift to broadly accountablegovernments, and decentralization some-times only leads to relocating patronageand the consequent capture of benefits bylocal elites.

An important corollary of this analysis isthat service delivery’s effectiveness for bothpoor and nonpoor people is highly depen-dent on the institutional context where thedelivery takes place. Similar designs andspending levels on delivery of education,health, roads, water, electricity, or other ser-vices can have very different effects depend-ing on how they work in practice, bothacross countries and across localities withincountries. For example, decentralized deliv-ery with local participation may lead togreat service quality for the poor in one sit-uation, but be captured by clientelistic prac-tices of local elites in another, or renderedineffective by unmotivated and underpaidfront-line workers in yet another. When weturn to specific experiments in servicedesign, we see considerable variety. Privati-zation can lead to greater service efficiency,coverage, and quality for the poor in onecase, but corruption and increased prices inanother.

We also analyze these experiences throughan additional prism that examines the extentto which the poor (and the nonpoor) are“empowered” in their relationships with bothpolicy makers and service providers—whether service-providing organizations orfront-line workers. In essence, we see poorpeople’s empowerment as the product of theinteraction between what we term the“agency” of the poor (i.e., their capacity toexercise choice, to envision alternativecourses of action, and to act upon theirchoices) and the social and political opportu-nity structure (i.e., the institutional contextwithin which poor actors operate) thatenables the exercise of that agency. Thisopportunity structure depends on a varietyof factors, including the openness of formalor informal institutions, the power and cohe-siveness of elite groups, and the state’sbureaucratic capacity. The extent of empow-erment is fundamental to both the overallservice coverage and quality, whether underpublic, private, or nongovernment auspices,and the workings of particular institutionaldesign experiments.

While context is central—and there areno magic bullets—it is possible to interpretsuccess and failure patterns. Overall, we findthat “long-route” reforms have broughtmixed results when used to strengthen bothpolicy makers’ responsiveness to citizensand the internal performance of servicedelivery mechanisms. Such deep institu-tional reform is proving to be a muchlonger-term endeavor than was originallyenvisioned, especially when the poor havehistorically been not empowered or havenot been incorporated in local or nationalpolities through clientelistic structures. Insome cases, however—within some socialfunds or privatization experiences, forexample—there are important positiveexperiences, but this greatly depends on thepolitical and institutional context. By con-trast there are many positive cases of “short-route” mechanisms that have brought gainsthrough either increased choice of clients(for example, as a product of increasedcompetition for telecommunications) orincreased client power (for example, inexperiments in greater parental influenceon schooling).

Introduction 5


In chapter 4 we turn to implications forfuture policy design and formulation. Draw-ing on the interpretation of past successesand failures, we argue for approaches thatseek both to work through the long path to strengthen the long route of greateraccountability to all citizens and strengthenthe workings of core government services,and continued experimentation with a vari-ety of short-route mechanisms. These tworoutes are seen as highly complementary. Inboth it is fundamental to link specific tech-nical designs with measures to empower thepoor, both through strengthening theiragency and improving the institutionalopportunity structure where they live andpotentially influence policy makers, organi-zations, and front-line workers. In designingpolicies, it is important to shape designs andapproaches in light of the country or localcontext—the best changes in Chile may bequite different from those in Guatemala,

and those in Brazil’s southeast differentfrom the northeast.

Finally there are important conclusionsfor the policy design process. Integratedwithin democratic debate, systematic explo-ration and evaluation processes are key forboth the “technical” design issues and forthe broader process of empowering all citi-zens and making decisionmakers moreaccountable. External actors can play avaluable complementary role, but thisneeds to be shaped to support internal pol-icy experimentation and accountabilityprocesses, by exploring opportunities tosupport institutional change, strengthenevaluation and debate processes, and facili-tate systematic sharing of experiences.

Notes1. Salamon et al. (1999) present a quantitative

assessment of the service delivery role thatNGOs played in Argentina, Brazil, Colombia,Mexico, and Peru.



2

Service Delivery and SocialOutcomes: A Story of Successes and Failures

7

Latin America has a long history of eco-nomic, political, and social inequality thathas had a potent influence both on thequestion of service delivery for the poorand on overall economic and social perfor-mance. The last two decades, though, havebeen a period of change for most countriesin the region. There has been a dramaticallydifferent scene in country after countrystemming from the overwhelming shiftfrom authoritarian to constitutional rule,the movement to more decentralized gov-ernments, the rising activity by civil societyactors outside the state, the growing role ofNGOs (both for- and not-for-profit) indelivering a variety of services, and theincreased activity of various LAC countriesin the world economy.

During this period, coverage of key ser-vices (education, health, water and sanita-tion, electricity, and telecommunications)increased—often substantially—in mostcountries in the region. However, progresshas been unequal between countries,regions, services, and social groups: largeinequities in access to services remain a fea-ture in most LAC countries. The improve-ments in social outcomes resulting fromservice coverage expansion have not beenlarge and spread out enough to break withthe resilient pattern of inequality that hascharacterized the region.

Often, recorded service delivery progressmasks large inefficiencies and deep inade-quacies in the service quality and relevance.The noted service coverage expansion wasoften achieved by relying on increased pub-lic spending from a very low base relative tothe region’s income level. Furthermore, the“service delivery reform” agenda that LACcountries face can be appropriately charac-terized as a moving target; as coverage gaps

in very basic services are closed, new ser-vices become a binding constraint for equi-table development in the region (for exam-ple, secondary and tertiary education). Inaddition, more complex challenges, besidescoverage, become apparent, the mostimportant being the quality of the servicesoffered. In summary, while there has beensignificant progress, there are enough sig-nals suggesting that there is still room forsubstantial improvement; service deliveryreform remains an open agenda.

Progress in Basic ServiceCoveragePrimary and secondary education coverageincreased significantly between 1980 and2000 in the region. Considering the 20largest countries, average primary enroll-ments of the relevant age group went from81 to 92 percent, while the correspondingfigures were 40 and 58 percent for sec-ondary education. Most of the countriesanalyzed displayed primary education cov-erage rates close to or above 90 percent in2000. Despite the noted increases, sec-ondary education coverage remains below70 percent in most countries in the region.In addition, coverage differences betweencountries are larger for secondary than forprimary education. For example, the mini-mum secondary education coverage in thesample corresponds to Guatemala (26 per-cent), while the largest corresponds toArgentina (79 percent). This gap is almostthree times larger than the correspondingone for primary education. Turning tointernational comparisons, only a couple ofLatin American countries underperform(considering their income levels) in termsof primary education, while in the case of

c h a p t e r

02Citizens.qxd 2/16/05 11:56 AM Page 7

secondary schooling, about half the coun-tries in the region experience gaps relativeto their income levels.

Figure 2.1 provides a summary view ofthe secondary education case. Three findingsare immediately apparent: (i) the absolutecoverage increases have been remarkable; (ii)most of the region still lags behind the mid-dle income countries, and the gap remainedalmost unchanged throughout the periodunder study; and (iii) very few countries(Argentina and Chile) come close to the cov-erage levels of high-income countries,including that of the United States. Althoughthe gap closed in the majority of the cases,this is due in part to the low relative startinglevel observed in the Latin American coun-tries in the sample, relative to their incomelevels.

Coverage of health services also increasedsignificantly during this period. For exam-ple, for the same group of countries the per-centage of births attended by a trained pro-fessional increased by 10 percentage pointsand immunization rates for children up toone year of age by 16 points. By 1999, thesecountries displayed higher than expected—

for their income level—immunization ratesand professionally attended births.

Figure 2.2 shows that there have alsobeen significant increases in coverage forbasic infrastructure services (water, sanita-tion, electricity, and, in some cases, tele-phone service). In the decade from 1986 to1995, coverage in a sample of 12 countriesincreased from 76 to 82 percent for thewater service, 67 to 80 percent for sanita-tion, and 82 to 89 percent for electricity,while telephone penetration doubled from5 to 10 percent. However, the figure alsoshows that coverage is still unequal withlarge differences between householdsaccording to their income levels.

Importantly, coverage increases have fre-quently been larger for the population’smost disadvantaged segments. This is par-ticularly true for services where coveragerates were relatively high to start (see, forexample, Foster 2003). Figures 2.2 and 2.3show these results for water, electricity, andprimary education. Despite this progress,about half the countries considered showedcoverage rates below those predicted bytheir income level.


0

10

20

30

40

50

60

70

80

90

100

Guatemala

Nicaragua

El Sal va

dor

Dominican Republi c

Costa Rica

Paraguay

Ecuador

Venezuela, RB

ColombiaMexic

o

PanamaPeru

Upper midd le in

comeBoliv

iaBrazil

Guyana

JamaicaChile

Argentina

United State

s

High incom

e

cove

rage

(%)

19701999

Figure 2.1 Secondary Education Coverage for Various LAC Countries, 1970 and 1999

Source: World Development Indicators (WDI).


The many successes of LAC countries inexpanding coverage of basic services overthe last two decades have not alwaysresulted in leveling the playing field amongdifferent groups in society; many coverageand quality gaps are reproduced for lessbasic services, which, increasingly, becomebinding constraints to equal opportunities.The most notable example is in highereducation. As basic and secondary educa-tion gaps have diminished in many coun-tries, higher education has become, per-

haps, the main differentiating factoramong the region’s social groups. In partic-ular, the last decade has seen a sharpincrease in the returns to higher educationin the majority of the countries. Thisdevelopment is responsible for a high shareof the region’s recent rise in incomeinequality.1 However, access to higher edu-cation continues to be highly restricted inmost countries and associated directly withincome status. Furthermore, access gapsbetween the rich and the poor appear to

Service Delivery and Social Outcomes: A Story of Successes and Failures 9

85

87

89

91

93

95

97

99

1986 1989 1992 1995 1996

acce

ss r

ates

(%)

60

65

70

75

80

8590

95

100

1986 1989 1992 1995 1996

acce

ss r

ates

(%)

1 3 5 7IncomeDecile

IncomeDecile

ElectricityWater

1 3 5 7

Figure 2.2 Evolution of Water and Electricity Access in Selected Latin American Countries, by Income Decile, 1986–96

Source: Foster 2003.Note: The figure includes Argentina, Bolivia, Brazil, Chile, Colombia, El Salvador, Guatemala, Honduras, Mexico, Paraguay, Uruguay, and Venezuela.The lines represent coverage rates for the different years included and for the income deciles shown in the lower part of the figure.

0.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

40.0

45.0

JamaicaPeru

Chile

Uruguay

Argentina

Panama

Costa Rica

Venezuela

Brazil

Mexico

Ecuador

Paraguay

ColombiaBoliv

ia

Honduras

El Salvador

rela

tive

gap

(%)

Around 1990Around 2000

Figure 2.3 School Enrollment for Children 6–12 Years Old: Relative Gaps between Income Groups

Source: De Ferranti et al. 2004.Note: The relative gap is defined as (E1-E0)/E0, where E1 and E0 are enrollment rates, respectively, for top and bottom income quintiles.


have increased over the last decade in mostcountries (see figure 2.4).

Another increasingly important gap canbe seen in social protection mechanisms inthe region. The traditional social insuranceinstruments that command the largest shareof public budgets in many countries in theregion have, by design, been associated withformal sector employment. As a result, thepoor, who often lack formal employment,have tended to be excluded from the benefitsof pensions, social health insurance, unem-ployment, and family allowance systems,among others. Today, the so-called “trun-cated welfare state” constitutes one of thekey bottlenecks to inclusive development inthe region (De Ferranti et al. 2004). Whilethe 1990s saw an expansion in targetedsocial assistance programs, these remainsmall in size relative to the other programs,which appear to be particularly resilient (asin the case of higher education) to change.2

Beyond all the noted limitations and dif-ferences across countries and services, it isfair to say that LAC countries have madeimportant efforts to expand coverage of key

services over the period considered. In manycases, such efforts were associated withchanges in the way services are delivered.Chapter 3, below, analyzes those changes.But more generally, the service coverageexpansion was typically associated withchanges in the spending level and composi-tion. For social services, this was largelyachieved through a substantial increase inpublic resources, while for infrastructureservices, public expenditure declined andthe key mechanisms were private financingand improved cost recovery.

CEPAL (2001) estimates that the increasein total per capita social spending during the1990s was about 50 percent, if pensions areincluded, and 30 percent, if they are not.Public education spending for the 20 largestcountries in the region grew from 3.3 per-cent of GDP in 1975 to 4.2 percent in 1999.The increase in health expenditures is evenmore marked. Several countries have largepublic spending increases (above 50 per-cent) in these two services, includingArgentina, Bolivia, Brazil, Chile, Colombia,Nicaragua, Paraguay, and Uruguay. In most


0

1

2

3

4

5

6

7

8

enro

llmen

t Q5

/enr

ollm

ent

Q1

enro

llmen

t ra

te (

%)

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Around 1990Around 2000Enrollment Q5Enrollment Q1

Nicaragua

Bolivia

Honduras

Paraguay

EcuadorPeru

El Salvador

Colombia

JamaicaBrazil

Panama

Argentina

Venezuela

Costa Rica

Chile

Uruguay

Mexico

Figure 2.4 Tertiary Education Enrollment Differences in Selected Countries

Source: De Ferranti et. al. 2004.Note: The bars represent the enrollment ratio between the richest and poorest 20 percent of the income distribution for the age group 18–23. Countries are orderedaccording to the 2002 GNP per capita as published by the WDI (lowest to highest). The lines are the country enrollment rates for the top and bottom quintiles for thelatest period (around 2000).


cases, the expanded coverage of social ser-vices was done fundamentally by increasingbudgets (with little or no change in thefinancing of some large ticket items in gov-ernment budgets that continue to benefitrelatively well-off groups in society). Thissuggests the presence of large inefficiencies,which, in some cases, may question the sus-tainability of the overall efforts.

During the 1990s, public expenditures oninfrastructure services fell from 1.2 percentof GDP to 0.6 percent of GDP. During themid-1990s, this decline was more than offsetby surging private investment. However,these tended to be concentrated in middleincome countries and the telecommunica-tions and energy sectors. Even at their peak,private capital flows covered barely half ofestimated infrastructure investment needsand less than one-tenth of needs in the waterand sanitation sector (Fay 2000). Moreover,the decline in private capital flows since1997 has led to an overall reduction in theresources available to expand these services.This is a serious concern, particularly giventhat remaining coverage gaps tend to beconcentrated in commercially unattractiverural and peri-urban areas, where it is unre-alistic to expect that private initiative willsolve the coverage problem without somedegree of public subsidy.

Service QualityThe evidence on quality is much less system-atic than that on coverage, but there are clearsignals that suggest both poor average qual-ity of services and significant differencesamong social groups. In education services,for example, the few Latin American coun-tries participating in international educationassessments such as PISA (Program forInternational Student Assessment) andTIMSS (Trends in International Mathemat-ics and Science Study) have shown averagelow scores (suggesting overall poor educa-tion quality) and, sometimes, high disper-sion (suggesting unequal access to qualityeducation). Recent PISA examination resultsindicate that in many LAC countries, educa-tion quality is lower than what would beexpected given their income levels. The LatinAmerican Laboratory (LAL) for Assessmentof Quality in Education’s First InternationalComparative Study of Education suggeststhat only a minority of students achieve sat-isfactory results (for example, read, recognizemeanings, and understand and interpretinformation) in most countries in the regionand those tend to be concentrated in thelarger cities; towns and rural areas show asignificantly worse performance. Figure 2.5presents the reading results; similar resultswere obtained for mathematics.


Mega-city

93.15

Urban

90.83

a. Level ISimple Literal Reading

Minimum expected percentage

Rural0

10

20

30

40

50

60

70

80

90

100

Mega-city Urban

75.84

64.45

b. Level IILiteral Reading Using Paraphrasing

Rural

48.93

Mega-city Urban

44.63

c. Level IIIReading Leading to Inference

Rural0

10

20

30

40

50

60

70

80

90

100

31.81

54.16

82.48

Figure 2.5 Percentage of Students Reaching Expected Levels, by Region and Language

Source: Laboratorio Latinoamericano de Evaluación de la Calidad de la Educación (Latin American Laboratory for the Evaluation of Qualityof Education).


The data from the Laboratorio Latino-americano de Evaluación de la Calidad de laEducación (Latin American Laboratory forthe Evaluation of Quality of Education)show that test results are strongly associatedwith a few school-related factors that can beseen as more direct quality measures: educa-tion level of teachers, instruction hours,teacher-student ratios, and degree of schoolautonomy/participation. Unfortunately, fewcountries collect such indicators systemati-cally. However, there is evidence that LatinAmerican countries tend to have yearlyhours of instruction, which are below theOrganisation for Economic Co-operationand Development (OECD) average. Giventhe high incidence of teacher absenteeism inmany LAC countries, the gap in actual classhours is likely to be even larger. Similarly,LAC pupil-teacher ratios are above theOECD average.

Of course, the poor quality problem isnot unique to education. There is scatteredevidence that quality is low in the provisionof some infrastructure services in LatinAmerica when compared to countries withsimilar development levels. For example,waiting time to get a telephone line or thenumber of faults in this service per unit oftime in Latin America is more than doublethose observed in East Asia. Relatively highwater coverage rates often hide seriousinadequacies in both the potability and ser-vice continuity. Inequity in the quality ofsome infrastructure services, especiallysewage and telephone service, also seems tobe high (Estache, Foster, and Wodon 2002).There is anecdotal evidence that the healthservices that the poor receive are of muchlower quality than those received by thepopulation’s better-off segments (see, forexample, World Bank 2002b). This is espe-cially true in regard to physician trainingand experience, the quality of the infra-structure where health services are offered,and waiting times to get appointments(DNP 2002, for the case of Colombia).

Social OutcomesThe trend of pronounced improvements inaccess to basic services has contributed tosignificant progress in the evolution of keyindicators of well-being and capabilities

(social outcomes), such as average educa-tional attainment, literacy rates, lifeexpectancies at birth, and child mortalityrates.3 For example, for the largest 20 coun-tries, life expectancy increased from 59.7 to70 years between 1970 and 2000, and mostof the countries show greater than 15 per-cent growth in this category. In addition,the countries with the lowest initial level(Bolivia, Ecuador, El Salvador, Honduras,and Peru) show the highest increases, catch-ing up with the regional average. By 2000,the dispersion in this indicator haddeclined. A very similar pattern emergeswhen analyzing infant mortality rates. Allthese countries increased their average edu-cational attainment in a significant mannerin the 1970–2000 period. It is also notewor-thy that the countries that experienced thehighest increases were the ones lagging in1970. Thus, in 2000 there is less dispersionin educational and health outcomes inLatin America. However, relative to itsincome level, many Latin American coun-tries are underperforming. In particular, theinfant mortality levels and rates of access tosafe water in many countries are below theirexpected levels; in some cases, the differ-ences are quite large (see figure 2.6).

Unfortunately, the noted service cover-age increases have not been sufficient toeliminate the very large inequalitiesobserved in many key social indicators. Forexample, in Mexico, the average person inthe bottom quintile of the income distribu-tion has 3.5 years of schooling, comparedwith 11.6 years for a person in the top quin-tile. In Brazil, a child born to a household inthe poorest 20 percent of the distribution isthree times as likely to die before reachingthe age of five as a child born into a house-hold in the top fifth. In Bolivia, that ratio isgreater than four, with children in the bot-tom fifth experiencing under-5 mortalityrates as high as the South Asian average. AGuatemalan family whose income is in thebottom 20 percent of the distribution has a57 percent chance of being connected to thewater mains and a 49 percent probability ofhaving access to electricity. The corre-sponding probabilities for a household inthe top 20 percent of the distribution are 92and 93 percent, respectively. Figure 2.7 por-



trays some of the most important inequali-ties in outcomes quite clearly.

Such inequalities are particularly note-worthy between indigenous and nonindige-nous people across Latin America (Halland Patrinos, forthcoming). Indigenouspeople in Mexico are four times as likely tobe illiterate than the nonindigenous. InGuatemala, the average indigenous adulthas just 2.5 years of schooling, compared to5.7 years of schooling among nonindige-nous adults. In Bolivia, indigenous peoplehave 5.9 years of schooling, but nonindige-nous people have 9.6 years of schooling.

One in four indigenous Ecuadorians neverattended school, compared to only one in20 nonindigenous Ecuadorians. Comparedto nonindigenous Ecuadorian children,indigenous Ecuadorian children are twiceas likely to be chronically malnourished. InMexico’s predominantly indigenousmunicipalities, the infant mortality rate is41 per 1,000, compared to 24 per 1,000 inpredominantly nonindigenous municipali-ties. The child mortality rate for Ecuador’sindigenous population is double the ratefor the nonindigenous population (10.5and 5.1 per 1,000, respectively).


–80%

–60%

–40%

–20%

0%

20%

40%

60%

80%

100%

120%

0

2000

4000

6000

8000

1000

0

1200

0

1400

0

1600

0

1800

0

GDP/capita, PPP adjusted

–15%

–10%

–5%

0%

5%

10%

15%

20%

25%

0

2000

4000

6000

8000

1000

0

1200

0

1400

0

1600

0

1800

0


–50%

–40%

–30%

–20%

–10%

0%

10%

20%

30%

40%

0

2000

4000

6000

8000

1000

0

1200

0

1400

0

1600

0

1800

0


–40%

–30%

–20%

–10%

0%

10%

20%

40%

30%

0

2000

4000

6000

8000

1000

0

1200

0

1400

0

1600

0

1800

0


a. Percent Difference from Predicted Infant Mortality b. Percent Difference from PredictedLife Expectancy

d. Percent Difference from Predicted Water Accessc. Percent Difference from Predicted Adult Literacy

Figure 2.6 Key Social Outcomes: Deviations from Expected Values Given Income

Source: Hicks and Todd 2003.Note: PPP = purchasing power parity. 1The regressions include 32 LAC countries with data taken from the WDI. The vertical line denotes weighted average GDP/capita, PPP adjusted, year2000. The worst results are found in (i) infant mortality: The Bahamas, Barbados, Brazil, and Mexico; (ii) life expectancy: The Bahamas, Brazil, Guyana,and St. Kitts; (iii) adult literacy: El Salvador, Haiti, Guatemala, and Nicaragua; and (iv) water access: Argentina, Belize, Haiti, and Jamaica.


Overall, the evidence reviewed is indica-tive of the growing importance that govern-ments in the region have given to expand-ing social services. However, these trendsraise many questions about challenges forthe future. There are a number of countriesthat still show insufficient commitment ofpublic resources to financing key services.But overall, further progress is unlikely tohappen if governments keep relying only onincreasing resources and if more effectivedelivery systems are not developed. This isthe case particularly when fiscal constraintslimit the capacity to respond by simplyadding more resources. The same is true forservices for which ensuring quality and rel-evance is essential and, thus, deeper changesmay be needed in the type of service pro-vided. It is in this context that we stress theimportance of assessing how well today’sservice delivery systems are positioned toaddress the many remaining challenges fac-ing LAC countries, particularly in terms ofserving the poor. The next chapter will pro-vide an assessment of the changes intro-duced in LAC service delivery systems overthe last two decades.

Notes1. In fact, secondary education returns remained

almost unchanged during the 1990s, while thosefor tertiary education were on the rise, implyingthat the skill premium grew importantlythroughout the region. Santamaría (2003) findsthat this increase in the wage gap between thosewith high education levels and the rest of thelabor force explained most of the increase inoverall income inequality in Colombia over the1990s.

2. In 2000, CEPAL (2001) estimated that pensionspending was around 5.3 percent of GDP as anaverage for the region and that it representedmore than half of total social spending. When theregion’s 15 largest countries are considered, onaverage, only 7 percent of that expenditure bene-fits the poorest 40 percent of the population.

3. Of course, many other (macro, micro, andstructural) factors have contributed to theobserved progress in those outcomes. For exam-ple, child mortality reductions tend to be associ-ated not only with health services, (for example,distribution of hydration tools to poor house-holds), but also with improved parental educa-tion, increased coverage of water and sanitationservices, higher incomes, and reduced fertility.


Figure 2.7 Inequality in Social Outcomes (Health): Bolivia, 1997, Peru and Nicaragua, 2000

years of age

Under 5 mortality is 143

Poorestquintile

Richestquintile

0

household income quintile

perc

enta

ge o

f chi

ldre

nsh

ort f

or th

eir a

ge

Poorest1 2 3 4 50 0

10

20

30

40

50

50

100

150

200

2nd 3rd 4th Richest

a. Bolivia 1997 b. Percent of stunted children by income quintile

PeruNicaragua

Source: De Ferranti et al. 2004.


3

LAC Service Delivery:Assessing Two Decades of Change

15

The last two decades have seen some radicaltransformations in state organization andstructure and in the relationship between thestate, its citizens, and markets. The spread ofdemocracy, state decentralization, a redefinedstate role, and increased activity by non-governmental actors represent importanttransformations that create opportunities for change in accountability relationshipsbetween policy makers, provider organiza-tions, front-line providers, and citizens.

This chapter seeks to explore the influencethat these various changes have had on thedelivery of services to poor people in LatinAmerica. We review the evidence on how key accountability relationships were trans-formed during this period and the influencesuch transformations had on service delivery.First, we look at changes in the relationshipsbetween policy makers and public and pri-vate service providers. Next, we reviewchanges in the relationships between clientsand providers that have sought to increase“client power.” Finally, we consider changes in the relationships between citizens and pol-icy makers and their impact on overall stateaccountability. In doing so, we will seek tohighlight the many connections betweenthese accountability relationships and iden-tify remaining bottlenecks to improve servicedelivery systems in Latin America.

This chapter draws on available studiesand assessments of specific country andsector service delivery experiences. Thegrowing emphasis on conducting system-atic evaluations of policies and programsamong both academics and developmentpractitioners provides us with a goodempirical basis for our review. Nevertheless,important knowledge gaps remain (forsome types of services, groups of countries,and service delivery approaches). In our

assessment, we have made an effort to rec-ognize and identify such gaps in the twodecades of change in service deliveryapproaches in Latin America.

Reforming the CompactLatin America has long been characterized bya general tendency for its states to be “weak”and “unequal,” specifically when deliveringkey public goods (especially in terms ofmacroeconomic stability and universal guar-antees of property rights and of citizenship)and provisioning quality basic economic andsocial services to all (De Ferranti et al. 2004).Where public goods are inadequately pro-vided, wealthier groups either resort to pri-vate mechanisms (for example, with privatesecurity or holding savings abroad to dealwith macroeconomic instability) or actuallybenefit from weaknesses (for example, byinfluencing justice systems through privateinfluence). Where basic goods are underpro-vided, it is typically the poor who are servedlast and where quality is low, richer groupshave greater means to opt out in favor of pri-vately provided services. To a great extent,such weakness has been associated with badlyfunctioning state bureaucracies in charge ofservice delivery—even though the underly-ing causes of such deficient functioning havetypically been linked to the social and politi-cal structures within which such bureaucra-cies operate. Thus, a key aspect to improvingservice delivery is associated with the incen-tives and capacity of the policy makers toestablish effective compacts—broad, long-term accountability relationships—with theorganizations (state and nonstate) in chargeof service provision.

The last two decades have seen manyattempts at reforming the state’s internalorganization, its structure, and its roles in

c h a p t e r


Latin America. Some of these attempts hadthe specific purpose of improving state ser-vice delivery effectiveness, while othersresponded to different policy goals (such asthe state decentralization process). In thissection we review the most important ofthese reforms, searching for empirical evi-dence on whether and how they improvedservice delivery. First we discuss the effortsto reform government bureaucracies. Next,we explore the effects of administrativedecentralization that is part of the multifac-eted state decentralization process that LACcountries have been implementing. Finally,we look at a broad range of initiatives thatinnovated service delivery by establishingstronger, more performance-based com-pacts with actors outside the mainstreamstate bureaucracies.

LAC government bureaucraciesremain generally weak, ineffective,and largely unaccountable for resultsA mix of reforms emerged in the regionover the last two decades, whereby govern-ments have sought to modernize and adapt

their structures and administrative systems1

and introduce new public managementapproaches. Many countries significantlyimproved capacities for core economicmanagement (see box 3.1). However, theeffects of the modernization efforts onimproving service delivery have been lim-ited to date. Implementing fundamentalreforms of government bureaucracies tomake them more results-oriented hasproven to be a long, arduous path rife withpolitical and technical difficulties.

Initial public sector reform effortsfocused nearly exclusively on fiscal adjust-ment objectives and the broad redefinitionof the state’s roles. Attention graduallyshifted to modernization of core govern-ment functions, such as macroeconomicplanning, financial management, and taxadministration. By the end of the 1990s,state institutional reform was squarely onthe region’s development agenda.

These reform efforts tended to empha-size improvements in internal processes andinstruments. Aided with technologicaladvances in the 1990s, integrated financialmanagement systems have become popu-larized as potent instruments for strength-ening government financial management(especially expenditure control during thebudget execution stage and ex post report-ing). Integrated financial management sys-tem reforms are sometimes seen as buildingblocks or as a catalyst for broader institu-tional reforms of the public expendituremanagement systems; better informationcan contribute to developing better institu-tions that facilitate better decisions andeventually, better outcomes.

The introduction of these improvementsin many countries has been among the mostsuccessful reform efforts in the region’s pub-lic administration realm in the 1990s. As seenin Guatemala’s reform (see box 3.2), it is pos-sible to use such systems in a way to deriveimmediate service improvement impacts(Dorotinsky and Matsuda 2001). Overall, thecontribution of the region’s financial man-agement reforms has been to establish betterinput control as a fundamental buildingblock of good public administration.2

LAC countries have also begun developingthe technical foundations for performance-


Centralized approaches to strengtheningcore economic management functionshave been relatively effective, and manycountries now boast professional centralbanks and finance ministries aided by com-petent technocrats and information tech-nology (IT)-based systems to better man-age public monies.Why was this reformapparently more successful than others,especially those in service delivery? We sug-gest two reasons why this domain was dif-ferent from that of most front-line servicedelivery programs:

1. First, the huge costs of macroeconomicmismanagement in the 1980s shapedpolitical will and practical learning tomake changes in core governmentmanagement functions.

2. Second, macro policy is an area that canbe effectively run by small cadres at thenational level. Moreover, once the pop-ulist macro policy’s failures as a redistri-bution instrument became manifest, itbecame an area that offered relativelyfew patronage opportunities and thusis more easily delegated to technocrats.

By contrast, large service deliverydomains are intrinsically transaction-inten-sive and involve discretionary decisions byfront-line bureaucrats that are difficult tomonitor and control centrally. It is this com-bination that makes it particularly prone toclientelistic abuse and thus challenging toreform (World Bank 2003a; Pritchett andWoolcock 2002).Widespread citizen discon-tent with government performance hassomehow failed to translate into a strongdrive for results-oriented public sectorreforms.This contrasts with the evolution ofreform efforts in OECD countries, where fis-cal pressures led not only to the generalmovement toward tighter fiscal policy andexpenditure control (Atkinson and van denNoord 2001), but also to a greater expendi-ture efficiency focus (for example, anemphasis on “value for money”) and conse-quently to deep public sector managementreforms.We might then speculate thateither the public sector bureaucracy’s poli-tics or culture has to change for fiscal pres-sure to translate into concerted efforts toimprove service delivery institutions.

B O X 3 . 1 Macro vs. Sectoral Reforms


based budgeting and management with theuse of performance indicators and the intro-duction of program evaluations. Severalcountries, including Bolivia, Brazil, Chile,Colombia, Mexico, and Peru, have embarkedon efforts to develop a results-oriented man-agement framework even though, with thepossible exception of Chile, no country hasyet developed a fully functioning perfor-mance management system. Often theseefforts have suffered wavering political com-mitments by different governments. In somecases, the work done so far appears to bemostly technocratic exercises with limitedlinks to actual government decision-makingand management processes.

Arguably the most important publicsector reform from the viewpoint ofimproving service delivery, especially inthe labor-intensive social sectors, is per-sonnel management reform. After theemphasis on public sector retrenchment inthe fiscal adjustment context, the secondhalf of the 1990s saw the emergence of civilservice professionalization, or the estab-lishment of a merit-based civil service, as acommonly-agreed-on priority in a num-ber of countries. Yet deep reform of the

civil service systems has proven politicallydifficult and technically intractable. Mostcountries in the region have yet to estab-lish professional “Weberian” bureaucra-cies, which many analysts see as a neces-sary precursor to introducing modernmanagerial approaches (Shepherd 2003).Far from introducing performance-ori-ented personnel management practices asa necessary complement to strengtheningthe public sector’s service delivery capac-ity, several countries in the region are stillstruggling to keep the aggregate wage billunder control and lack sufficient capacitiesto control deployment and attendance, letalone performance, of their public ser-vants. In very few cases, this has reachedthe level of failed states (see box 3.3 ).

Aside from the technical and fiscal com-plexities (for example, how to resolve thewell-known dilemma of controlling theaggregate wage bill for sound fiscal policyon the one hand and paying individual civilservants competitive salaries on the other),civil service reforms are riddled with politi-cal difficulties. Introduction of a merit-based civil service is difficult because stake-holders in the political system actively resist

LAC Service Delivery: Assessing Two Decades of Change 17

At one point or another, most countries in theregion have attempted to reform their financialmanagement systems (normally budgeting,treasury, accounting, and debt management) tomake public expenditure more effective andefficient.The Guatemala Integrated FinancialManagement System (IFMS) stands out as oneof the most successful. Emerging from threedecades of civil war, Guatemala had a desperateneed not only to improve service delivery, butalso to restore confidence in the public sector’sintegrity and efficiency.

Some examples that illustrate Guatemala’ssuccess and its potential impact on the cover-age of public sector services are:

• Elimination of the “floating debt” that arosefrom inadequate budgetary controls, whichused to lead the Ministry of Finance to devotevirtually the full first quarter of the year topaying commitments from the previous year.It turned out that many of thesecommitments were not legitimate. Close to$100 million was saved through the controlsimplemented as part of the IFMS.

• By decentralizing purchases through theIFMS, control of the procurement processeswas improved and arrears with supplierswere eliminated.This reduced the cost of cer-tain wholesale purchases for hospitals andschools by 10 to 70 percent.

• As part of the IMFS deployment, paymentsare electronic and checks are no longerissued to suppliers, with the exception ofteachers’ salaries. Once these begin to be paidelectronically, an estimated savings of $2.5million per year will be achieved simply byeliminating the use of physical checks.

• As a complement to the IFMS, the project isdeploying an online procurement system(“Guatecompras”), which has been widelypraised both inside and outside the country.It is expected that in the medium run, up to a20 percent savings in public sector purchaseswill be achieved through increasedtransparency and more effective competition.

• Legislators, private citizens, NGOs, and otherstakeholders can access proposed and exe-cuted national budgets at several computer

sites in Guatemala and through the Internet(www.siafsag.gob.gt).

• Physical and financial indicators are nowincluded in the budget to better measureperformance and promote a result-orientedfocus, and line ministries have timely andaccurate information to improve physical pro-gramming and service provision.

• At the municipal level, the project is also imple-menting IFMS platforms that share the nationalsystem’s functionality at a smaller scale.Thesesystems will also generate substantial savingsthat can be directed to improving municipalgovernment service delivery.

By themselves, back-office systems such asthe Guatemala IFMS cannot directly affect thequality of final public service delivery; however,they can help increase coverage by reducinginput costs, be it through improvedprocurement practices, speedier payments tosuppliers, or reduced corruption opportunities.In addition, they provide real-time financialinformation that can be critical for making pol-icy decisions in all sectors.

B O X 3 . 2 Public Financial Management Reform in Guatemala


it. Civil service unions are often opposed tochange initiatives that are designed toimprove their performance incentives andthus also affect their entrenched (and usu-ally legally sanctioned) privileges and bene-fits. Unions typically oppose these mea-sures, especially those geared toward fiscaladjustment purposes (which result in joblosses), better internal controls (whichresult in reduced room for opportunisticbehavior), and managerial flexibility (whichcan result in both job losses and greater per-formance exigency). In Brazil, for example,the 1998 administrative reform was intro-duced with the aim of introducing greaterflexibility to public sector personnel man-

agement and to introduce greater perfor-mance orientation. However, the reformhas stalled partly due to strong union pres-sure (Rinne 2001). In Chile, union opposi-tion has been cited as one of the reasons forrelatively limited progress in civil servicereform aimed also at introducing greaterflexibility in the public sector personnelmanagement regime (Egaña 2003).

Union opposition to performance-enhancing reforms tends to be particularlyproblematic in social sectors. For example,teachers unions typically negotiate for jobsecurity (as well as, obviously, for betterwages) and against results-based qualityassessment systems and local control.3 Incountries ranging from Bolivia, Costa Rica,the Dominican Republic, Peru and Vene-zuela, just to name a few examples, unionsopposed reform measures to strengthenmeritocracy and teachers’ accountabilityfor learning outcomes or to introduce pri-vate provisions so as to generate competi-tive pressure for public providers and offer“exit” options to parents (Gray Molina,Pérez de Rada, and Yeñez 1999; Maceiraand Murillo 2001). It is also true thatteacher unions—or at least factions withinunions—have sometimes been reform advo-cates and governments have worked withthem for quality and efficiency improvingreforms, for example, in Minas Gerais(Grindle and Mason 2004).4 When explicitlyinvolved in the reform process, responsibleunions can play a positive role in defining thecontent of performance-oriented teacherservice reforms and facilitate their imple-mentation, as was the case in Chile over the1990s (Crouch 2004).5

Besides the frequent opposition frompublic sector unions, the most obvious cul-prits for lack of civil service reform are thecontinued prevalence of clientelism and thepoliticization of public administrations. Inmuch of Latin America, provision of publicjobs has been a major instrument both forachieving or maintaining national level sup-port and/or an instrument for the clientelis-tic exchange of local level favors. This prob-lem has been diagnosed in Bolivia, forexample, where the peculiarity of the politi-cal system creates permanent and strongincentives for political parties to divvy up


Haiti’s characterization as a failed state, orone that has been failing over decades,does not waive the rules of improving pub-lic service provision. However, it does maketheir application far more difficult.Thestate’s service provision role is minimal (forexample, private schools provide over 80percent of education), ministries managemuch of their budgets through “currentaccounts” for which there is no accountabil-ity, and public office has been used toreward partisan loyalists. Over the pastdecade, the shrinking group of donors stilloperating in the country lookedincreasingly to NGOs to implement theirprograms, or they managed their ownfunds, or if they channeled monies throughthe government, often short-circuited thehighly inefficient “normal” processes. Notsurprisingly, public sector human resourcequality, quantity, and distribution weretotally inconsistent with service provisionneeds. Most staff was located in Port-au-Prince; a United Nations Development Pro-gramme survey indicated that 68 percenthad only a primary or secondary education,and even offices with technically qualifiedpersonnel had too few of them to carry outtheir mandates.The near civil war leading toPresident Aristide’s departure was accom-panied by the exodus of many of the morequalified government staff, leaving an enor-mous gap at the upper and middle-management levels. Salaries and operatingbudgets, while never adequate, are now atan all-time low.

Assuming that the current transitionalgovernment and an eventual elected onecan maintain an adequate level of civilorder, aided by the presence of a UN secu-

rity force, they will be dependent on donorresources to finance anything beyond theiroperating budget and possibly even a fairshare of that. In this situation, the two high-est priorities are to regain control overexpenditures through the introduction ofan improved budgetary process and to findways to run or oversee service delivery andother basic government functions with thedepleted human resource base. Heredonors face a common dilemma in anextreme form—whether to emphasize exe-cution of their own programs throughwhatever means possible (such as by usingNGO executors, hiring the most qualifiedHaitians to backstop their efforts, or findingways to short-circuit less-efficient govern-ment procedures) or to sacrifice someadvances there in the interests of buildingpublic sector capacity. Should they decideto focus some efforts on capacity building,they will have to tailor them to the realitiesof existing organizational and humanresource capabilities. New procedures willhave to be simple, but effective, less relianton technology than in other national set-tings, and capable of use by the “average”employee, not just the upper level staff. Amix of implementation modalities will beinevitable, including the use of NGOs andthe private sector, but they should strive foran integrated rather than an either-orapproach. Salary subsidies or similararrangements to attract more qualified staffshould be used with caution as the fewbenefiting from them will not be sufficientto run the government, and the more basicproblem will be how to provide incentivesfor the ordinary state workers.

B O X 3 . 3 Haiti: Public Service Delivery in a Failed State


the public sector as a currency of politicalexchange to maintain minimum conditionsfor governing (World Bank 2000b). Bolivia’scase may be extreme in some sense, but it ishardly unique. In Mexico, with a moredeveloped pool of professional humanresources, politicization has reigned in theadministration of public sector personnel(Garibaldi n.d.). It is also seen in the onlyrecently changed tradition of use of teachingjobs as a means of patronage in Colombia(Angell et al. 2001). Public utilities are a fur-ther example of a patronage system of jobprovision and cheap, if low quality, services(Foster 2003). Overall, in a context wherejob granting remains a preferred mecha-nism to reward political clienteles, there is aserious challenge to design effective incen-tive systems that motivate front-line workersinvolved in service delivery.6 Furthermore,often union resistance to performance-enhancing reforms is a defensive moveagainst such practices.

Effective social service delivery dependscritically on the quality and dedication ofteachers, doctors, and nurses. Yet, organizedresistance by unions and informality sus-tained by patronage and clientelism severelylimit reformers’ options for using explicitincentive mechanisms for enhancing theirperformance orientations (for example,performance-based pay). In addition, inmost countries in the region, paying thelarge number of social sector workers ade-quately remains a significant challengegiven national fiscal realities. A long-termsolution seems to depend on somehowencouraging greater professionalization ofthe social sector workers (especially teach-ers). But in the short run, it may be possibleto find promising alternatives in greaterreliance on “social control” mechanisms,such as community vigilance of familyhealth workers in the Brazilian state ofCeará (Tendler 1997) or school directorelections as practiced in several Brazilianstates.

The consequences of the lack of civil ser-vice reform go beyond continued generalweaknesses in public administration. Theycan also undermine institutional innova-tions in specific sector reforms by deprivingthe public sector of technically competent

personnel inculcated in the ethos of rule-bound integrity (for example, teachers whoshow up at classrooms every day to teach)or of personnel management flexibility totake advantage of incentives for improvedorganizational performance. The latter typeof problem is well illustrated by the difficul-ties that public health sector reforms facedin seeking to improve the incentive frame-work for efficiency and responsiveness (seebox 3.4).

Overall, the evidence suggests that thehistorical weaknesses of Latin Americanstate bureaucracies still limit the effective-ness of conventional public sector delivery.State modernization reforms have upgradedgovernments’ institutional capacities con-siderably—at least in some areas of publicadministration systems—and it is likely thatsuch upgrading will have substantial effectsover the longer term. But very few countries


A main thrust of LAC health sector reformshas been to organizationally separate thethree key health system’s functions: regula-tion, financing, and service provision. In par-ticular, purchaser-provider split reformshave been introduced in many LAC coun-tries with the objective of improving thepublic health sector’s efficiency and respon-siveness incentive framework. Under inte-grated systems, the central authority hasthe simultaneous missions of maximizingthe impact of its resources on the popula-tion, of financing the system, and of ensur-ing financial sustainability of its ownproviders. Historically, that combination ofmissions created incentives for the state tofocus on ensuring a stable revenue flow forits own providers, rather than on examiningthe impact on the population. Chileprovides the earliest case of these reformswith the creation of the National HealthFund (FONASA) in 1981. After significantlegal, managerial, and financial reforms,there is almost complete separation of pur-chasing and provision in Chile’s publichealth sector. Colombia is the other para-digmatic case; in the mid-1990s, it initiatedan ambitious health insurance reform thatimplied a clear and nationwide separationof functions. Efforts at establishing publichospital autonomy were widespread, but

limited in scope and effects (World Bank2001b; World Bank 2002b; Over and Watan-abe 2003; World Bank 2003c).

For these reforms to be successful inenabling public sector performance changerequires significant flexibility in resourcemanagement and allocation by serviceproviders. In other words, reform implemen-tation requires that public providers beincreasingly able to adapt their service pro-duction functions and cost structures to thecontinuous evolution of price signals deter-mined by the new payment mechanisms.There is increasing evidence of the difficul-ties in implementing the purchaser-provider split (Baeza 1996). For example, theslow pace of the reforms in Costa Rica andChile—two countries with relatively effec-tive public administrations—is due to agreat extent to the management rigidity inall production factors at the provider level,particularly in the lack of flexibility in man-aging human resources. Indeed,international experience (England 1998)increasingly suggests that purchaser-provider split reforms are facing seriousconstraints due to the difficulties in reform-ing civil servant regulations in the publichealth sector and the resulting rigidity inmanaging production factors and modify-ing cost structures by public providers.

B O X 3 . 4 Lack of Civil Service Reform as a Health ReformBottleneck


have been able to move decisively towarddeveloping a results-oriented public admin-istration or achieve significant depoliticiza-tion of the state apparatus. We interpret thisas being due to a combination of the intrin-sically long-term nature of institutionalchange and the persistence of countervail-ing forces from groups, from patronage-oriented political groups to unions, thatbenefit from existing arrangements. Theimplication is, however, that service deliv-ery quality by state agencies remains a seri-ous bottleneck, particularly for those ser-vices (for example, education and health)for which personnel practices are crucialand in countries where resistance by inter-est groups is the strongest.

Service delivery functiondecentralization was attempted incountry after country withheterogeneous results Over the last two decades, most countries inthe region have gone through a process ofdecentralizing service delivery responsibili-ties. Practically all countries have under-taken some form of decentralization oftheir primary and secondary education sys-tem, which involved the transfer of deci-sion-making powers to actors within or,more likely, outside the ministry of educa-tion’s bureaucratic structure. In manycountries, the water industry was trans-formed from a single national providerunder the central government’s control, to ahighly decentralized structure with servicebeing delegated to hundreds of new munic-ipal utilities.7 Local and regional govern-ments were also charged with growingresponsibilities for public health facilitiesand programs. Decentralization also tookplace in the transport sector, with munici-palities and regional governments becom-ing increasingly involved in tertiary andsecondary road management. Over time,some countries also delegated responsibili-ties for social assistance programs to subna-tional governments.

These reassignments of responsibilitiesfollowed many paths and approaches,8 butin few, if any, cases was administrativedecentralization driven primarily by the

goal of improving the service delivery effi-ciency. More typically, decentralization ofservice delivery responsibilities was largelydetermined by the design and outcomes ofboth political and fiscal decentralizationprocesses. In all cases, however, the decen-tralization of service delivery responsibili-ties to local governments9 has resulted in areform of the compact between govern-ments and service providers whereby thelatter are increasingly under the supervisionof subnational governments, as well as newcompact relationships between the nationalgovernment that still sets overall sectoralpolicies and the subnational governmentscharged with service delivery.

Whether these changes result in better ser-vices depends crucially on local bureaucraciesbeing able to function more effectively thanthe national ones. The key is for local govern-ments to have the combination of resources,authority, and incentives needed to do so(World Bank 1999).This can also be seenfrom the perspective of the tradeoff betweenincreased information at lower governmentlevels and the influence of local power struc-tures and practices (Bardhan and Mookher-jee 2000a, 2000b). In the Latin American con-text, there is a strong history of the power oflocal strongmen and of the importance ofunequal local social structures. However,there is also a significant variety of experi-ences, including in terms of the democratiza-tion and decentralization responses. The evi-dence essentially points in the direction ofmixed results: decentralization’s impact onservice delivery varies significantly betweencountries and sectors or geographic areaswithin countries. Critical to explaining theresults are local political conditions and thecountry- and sector-specific policy frame-works under which decentralized servicedelivery operates.

Design issues. In many cases, the overallframeworks under which responsibilitieswere decentralized in most countries andfor most services have created overlaps andgaps, inadequate or insufficient authoritiesand resources, and few (enforceable) mech-anisms to promote local government per-formance orientation. Most countries hadmany difficulties in establishing well func-



tioning compacts between government lev-els to ensure the necessary incentives foreffective service delivery to the poor.10

In many cases, blurred responsibilitieshave both made the tasks of newly empow-ered local governments complex as well ascreated too many opportunities for individ-ual governments to essentially avoidbecoming accountable for results vis-à-vistheir citizens. In the case of health services,for example, in most cases decisions onhuman resources (for example, salaries andcivil service rules) remained centralized orin the hands of regional governmentsimposing restrictions on municipal man-agement autonomy (Bossert 2000). In thosecountries (such as Bolivia and Colombia)that distributed responsibilities for educa-tion services among three government lev-els, room was created for strategic behaviorfor each government level to seek to expandtheir power or budgets, and inconsistenciesin the assignment of responsibilities andresources made it extremely difficult forlocal actors to “manage” schools in practice.Similarly, the arrangement to have nationalregulatory agencies oversee the perfor-mance of subnational water utilities (forexample, in Bolivia, Colombia, and Peru)has not functioned very effectively due todiffering sector policy goals betweennational and subnational governments, andeven conflicting jurisdictions that under-mine the regulatory agency’s authority.11

Superimposing decentralization over othersector reforms is another source of inconsis-tency. For example, many countries now havelegal contradictions between policies pro-moting community-based approaches torural water and sanitation services anddecentralization laws. This is the case, forexample, in Honduras, where a recentlypassed water sector law provides legalrecognition and a mandate to communitywater committees to provide services, whilethe decentralization law gives municipalgovernments the same responsibility. As aresult, paradoxically, rural water and sanita-tion decentralization may actually result incentralization of decision-making power bytransferring legal responsibility and author-ity away from the hands of rural communi-ties to local governments.

Fiscal transfers remain the principalfinancing source for a large number of localgovernments, creating the potential forstrong accountability lines toward the cen-ter. In reality, though, the establishment ofclear, transparent compacts linked to servicedelivery outcomes is the exception ratherthan the rule. In Argentina, for example, theabsence of a clear “compact” between thenational level and the provinces (that is, lit-tle or no specification of targets and objec-tives and the use of general transfers thatare not tied to any particular efficiencyindicator) implied that decentralization’seffects on education quality and coveragevary a lot across provinces, depending onprovincial management (Galiani and Schar-grodsky 2002). This raises both efficiencyand equity concerns. For water and sanita-tion, notwithstanding extensive decentral-ization, central government often retains arole in financing services, whether throughbudget transfers or credit lines for majorinvestments. This financing relationshippotentially provides an additional mecha-nism for increasing the accountability oflocal service providers, since access toresources could potentially be conditionedon absolute performance thresholds orranking against utilities in other jurisdic-tions. However, it is striking how little usehas been made of this instrument in prac-tice, with the majority of municipalities inthe region receiving fiscal transfers withoutany accountability requirement against theresources spent.12

Nevertheless, there are good examples ofarrangements where strong intergovern-mental compacts coexist with large fiscaltransfers. In Chile, for example, the educa-tion delivery responsibility was transferredto municipalities through specific agree-ments that defined explicitly the responsi-bilities, rights, and obligations of the newproviders and linked central resources to thenumber of students attending class. As illus-trated by the case of Brazil’s FUNDEF, aninnovatory financing mechanism that col-lects resources from state and municipalgovernments in a single fund and redistrib-utes them according to the amount of stu-dents enrolled, thus ensuring spending min-imum levels in all states and municipalities,



when the compact between the center andsubnational governments is strengthened,there is great potential to generate improve-ments in educational achievement out-comes and enrollment, in particular in poorareas (World Bank 2003b).

Local responses. While the characteristicsof the overall fiscal, administrative, andpolitical frameworks under which decen-tralization takes place are clearly influentialon service delivery opportunities for localdecisions and actions, it is increasinglyapparent that localities respond differentlyto such frameworks, emphasizing the roleof local political systems and local socialorganization’s characteristics and strengths.

The experience across municipalities inBolivia is of interest here, particularly sincethey appear to be characterized by relativelyweak state capabilities and embeddedpatronage politics (World Bank 2000b).Popular participation was intended as anational reform to shift resources and con-trol to municipalities and legalize local levelcivil society groups, of which campesinos’syndicates were the most important inmany parts of the country. While thechanges brought benefits, on average,(Faguet 2000) the dynamics of changedepended on the variable local social con-text. In some parts of the country, notablyin some of the Cochabamba valleys, wherethere was a relatively strong social mobiliza-tion tradition (that is, greater agency capac-ities of the poor), the legal changes in thedirection of greater openness appear to bebringing initial benefits in the form of bet-ter public action. In the traditional sierracity of Sucre, by contrast, a traditionallymore closed political culture and a strongerelite may have caused the decentralizingreforms to primarily increase the localelite’s power (Gray Molina 2002; Calderonand Szmuckler 2004; Blackburn 2001).

But beyond these broad, albeit power-ful, structural forces, it is the combinationof many factors that appears to be behindsuccessful local service delivery reform.For example, Bogotá city engaged in acomprehensive reform program of thelocal education system, improving learn-ing and teaching outcomes. A variety of

factors contributed to this success: localelectoral competition as motivation topursue reform; the fact that educationreforms included both private and publicschools; close parental oversight; controlover human resources on the municipal-ity’s behalf; and a new incentive frame-work that the central government estab-lished (capitation grants instead ofcost-driven central transfers). All these fac-tors combined provided local politicianswith both the incentives and the ability tomove toward results-based management.In other words, changes along the variousaccountability relations (that is, not onlyor even primarily in the intergovernmentalcompact) made the observed results possi-ble—a recurrent theme in our analysis.

The experience of LAC countries sug-gests that effective decentralization oppor-tunities systematically differ between typesof services. In contrast to the education sec-tor’s mixed cross-country and in-countryevidence, there is surprisingly little evidencethat local elections have created any signifi-cant pressure to improve municipal waterservices. Only a minority of small andmedium-size cities in Latin America haveundergone major reforms of their waterutilities.13 It is only a matter of speculationas to why local level political incentivesimplicitly appear to be less in favor ofreforms in the water than in the educationsector. What is clear, though, is that asym-metric decentralization results by sector arerelated to the influence of size. In the watersector, a serious but unanticipated decen-tralization consequence has been the loss ofscale economies in service provision, as wellas the dispersion of scarce human resourceswith the capacity to manage and maintainwater systems. In commercial terms, theclient base for many smaller municipalitiesis entirely composed of low-income resi-dential customers, making it difficult toachieve financial viability. This type of scalefactor appears to be of much less impor-tance in the case of education, for example.Nevertheless, scale may also be a limitingfactor for some aspects of education serviceprovision, such as curriculum design orevaluation systems, which tend to remainthe central government’s responsibility.



More importantly, while small scale maynot be a serious bottleneck at any particularpoint in time, it may reduce flexibility toreact to changing conditions over time: forexample, by limiting the capacity of systemadministrators to change resource alloca-tion decisions in response to demographicshifts. Of course, these challenges do notnecessarily call for centralization. Rather,they emphasize the importance of arrange-ments for intermunicipal cooperation.

It is probably fair to say that the fruits ofdecentralization are still not fully harvested.In some cases (countries and sectors), someimportant changes (for example, in financ-ing arrangements to improve accountabilityfor results) are called for as a preconditionto create a better enabling environment forlocal delivery of services, and several coun-tries are already moving in that direction,Colombia being the most recent example.In other cases, more time may be needed toallow local governments and local politicsto play out as reformers hoped. In all cases,however, central governments are facedwith the need to provide incentives andimplement mechanisms to promote andfacilitate needed actions on the part of sub-national entities: central governments havea central role (monitoring results, providingincentives, and even intervening when ser-vice delivery fails) in making decentraliza-tion work for service delivery outcomes.

Establishing compacts with agenciesor actors outside the mainstreamstate apparatus has overallproduced good (short-run) resultsLAC governments have sought alternativeand often innovative ways to avoid the deepdifficulties of reforming bureaucracies(central and local) to improve servicedelivery outcomes. Many of these alterna-tive ways involve a combination of con-tracting out services to private firms andNGOs and relying on alternative bureau-cratic structures that bypass more tradi-tional hard-to-reform ones. Overall, theseefforts produced positive results in terms ofimproving service delivery outcomes.Below is a brief review of the assessmentsavailable for some experiences, followed by

a discussion of this approach’s limits onimproving service delivery.

Contracting out service delivery. Contract-ing out services to private operators,whether through long-term concessions oroutright asset sales, is most advanced in theinfrastructure sectors. Prior to 1980, utilitiesin most Latin American countries wereorganized as publicly owned and operatednational (or state level) monopolies. Politi-cians exerted their control over the sectorvia the appointment (and dismissal) of util-ity managers, and by providing subsidies tosupport unsustainable enterprises. Inreturn, utilities provided jobs, kept tariffsdown, and allocated investments and publicworks contracts on a political criteria basis.The consequences of this regime were spi-raling costs, low service quality, and a short-age of resources to fund badly needed ser-vice coverage expansions. From the early1990s onwards, there was a widespreadmovement to try and replace the “clientilis-tic” service provision model. The reformsstressed the need to provide a clear institu-tional separation between the roles of policymaker, regulator, and service provider. Ser-vice provision would be delegated to the pri-vate sector to ensure that services were runon commercial principles, and providerswere free to pursue service delivery objec-tives. In other words, the contracting outprovision to private firms implied that in thecase of utilities, many bottlenecks in thehealth sector from inadequate progress inimplementing civil service reforms (see box3.4) would not materialize.

The reforms implied a substantialincrease in private sector participation. Arecent survey of private sector participationin Latin American infrastructure during the1990s found that there had been more than200 concessions for water, electricity distrib-ution, and basic telephony services (Guasch2004). However, the scope for private partic-ipation proved to vary substantially acrosssectors; by the year 2000, over 90 percent oftelephone customers, 50 percent of electric-ity customers, and 15 percent of water cus-tomers in urban areas in Latin Americareceived their service from private operators.There is substantial evidence that private



sector participation has led to significantimprovements in the performance of utili-ties, particularly as regards connections,labor productivity, and service continuity.14

The increased coverage of electricity, water,and sewerage services primarily benefitedlower income groups, while the surge intelephone penetration primarily benefitedthe middle classes (McKenzie andMookherjee 2002). There is similar positiveevidence regarding water sector coverage.15

While coverage and efficiency effects weregenerally favorable, price effects were morevaried, often depending on initial condi-tions (De Ferranti et al. 2004).

An essential aspect of these reforms wasthe creation of a regulatory agency for utili-ties (with the key function of insulating ser-vice providers from political interference, sothat decisions can be taken on the basis oflong-term public interest). At present, LatinAmerica has more than 30 national levelregulatory agencies for the water, electricity,and telecommunications sectors. Regula-tors for subnational jurisdictions have alsosprung up in federal countries such asArgentina and Brazil, and for the water sec-tor in countries such as Ecuador and Hon-duras. In many cases, sectoral legislationincorporates several mechanisms designedto safeguard regulatory autonomy, in par-ticular by establishing clear criteria for reg-ulator selection and removal16 and makingregulators financially self-sufficient via sec-tor levies.

Similar, large-scale experiences of con-tracting out private concessions are lesscommon for social services. The Bogotáschool concessions program (which created18,000 new school places) is an importantexample of how private management ofpublic schools could work if used moreextensively. The municipal educationdepartment constructs new, state-of-the-artschools in low-income areas that are eachable to accommodate between 800 and1,200 students. A public procurementprocess offers the opportunity to managethe school; bidders (good-quality privateeducational institutions) are evaluated ontheir proposed management plans. Conces-sionaries are obliged to provide educationalservices to poor children and in return

receive remuneration from the departmenton a per-student basis. While the contractestablishes clear standards that must beupheld, the concessionaire has full auton-omy over school management and is evalu-ated solely on results. Concessionaries havealready produced striking managementimprovement results: they allocate, on aver-age, 55 percent of the per capita remunera-tion to human resources, well below the 90percent allocation by public schools.

More common is the practice of con-tracting out the delivery of social servicesto NGOs. In their efforts to improve effi-ciency and reduce costs, a growing numberof ministries of health in Central Americabegan to purchase primary health care ser-vices from NGOs—usually private, non-profit entities—to provide coverage indelimited geographic areas (Fiedler 2002).Through a formal, competitive process, ElSalvador, Honduras, Nicaragua, andPanama awarded contracts to NGOs toprovide coverage to remote (often indige-nous) populations with little or no accessto care, in return for a fixed, annual capita-tion payment. In most cases, the NGOsprovide care using a network of itinerantteams, usually complemented with com-munity-based paramedics. The NGOs aregenerally required to visit designated vil-lages or to provide care at specified deliverysites a minimum number of times per year.

These innovations started out as rela-tively small-scale experiments. In somecases—Nicaragua, for instance—they havenot been regarded as successful, and theirscale and financing were reduced. In othercases, both the clientele and authoritiesenthusiastically endorsed this approach,and its scope has gradually grown. InPanama, for example, NGO provider teamsnow cover about 95,000 people, and currentplans call for expanding the program tonearly 10 percent of the population in thenext few years. With the longest history ofcontracting out to NGOs among CentralAmerican countries and of purchasing agreater variety of services than its neigh-bors, Guatemala has contracts with approx-imately 90 NGOs that provide services tomore than three million people, about 30percent of the national population. Impact



evaluation studies show better results underthis approach than under the more tradi-tional public sector delivery model (La For-gia et al. 2004).

In terms of education, the largest case ofcontracting out is the Fe y Alegría schoolsthat operate in 14 Latin American countriesand cover approximately one million stu-dents. An evaluation of the program’s oper-ations in nine countries indicates thatschools integrated into Fe y Alegría outper-form traditional public schools, with lowerrepetition, dropout, and higher overallprogress rates in school and higher overallretention rates, on average (Patrinos 2002;Swope, Latorre, and Celedón 1998).

The national Fe y Alegría office in eachcountry oversees the education qualitygiven and offers training to its teachers andcenter directors, administers the educa-tional centers, and coordinates activities insuch a way that it becomes a local areadevelopment strategy. Fe y Alegría, forexample, developed many strategies toreduce grade repetition and school dropoutrates during the first years of schooling forchildren coming from the most disadvan-taged areas. Critical to Fe y Alegria’s successare its parent-centered strategies. These arebased on an open school environment thatpromotes effective parental involvement inthe learning process, community involve-ment in school management, and student-centered strategies that depend on an inter-active teaching method and recognize avariety of learning styles. The dynamic atplay in this case is strongly associated withthe altruistic and “empowerment”-orientedapproach that characterizes Fe y Alegría. Inother words, the effects of contracting outappear to be closely related to the contrac-tor’s nature.

Bypassing unreformed bureaucracies. Whenthe nature of the services or the political/his-torical conditions do not permit govern-ments to contract them out to either private operators or NGOs, the near impos-sibility of a governmentwide civil servicereform resurfaces as a major bottleneck to improving service delivery. In this context,a number of governments in the region have adopted, to different extents, targeted

approaches to public sector reform on anagency-by-agency basis. This approach hasshown promise of relatively quick success,including a handful of well-known examplesof impressive organizational turnarounds.Probably the best contemporary case is a setof semiautonomous agencies created in Peruin the 1990s. Instead of trying to reform thewhole government machinery, Peru createda series of “autonomous” agencies in keyareas related to sustenance of the marked-oriented reform and targeted poverty andsocial programs. These agencies included therevenue administration agencies, regulatoryagencies, and social program agencies[Comisión para la Formalización de laPropiedad Informal (COFOPRI) and Fondode Compensación y Desarrollo Social (FON-CODES)]. The agencies tended to be staffedwith competent professionals who were gen-erally better paid than civil servants in thecore ministries and operated with a degree ofautonomy from governmentwide rules gov-erning personnel and financial management.In some cases (for example, COFOPRI), theyalso experimented with various ways ofinvolving stakeholders in decision making asa means of improving accountability (seeCantuarias and Delgado 2004).

Although the approach of creating suchefficiency enclaves is more common in areassuch as revenue administration and regula-tion (see box 3.5), the model has also beenapplied widely for the provision of specificsocial infrastructure services through socialfunds. These funds were first launched as“social compensation” programs to softenthe impact of structural adjustment policieson the poor through the construction ofsmall-scale infrastructure projects targetedto poor communities in accordance withlocal priorities.

Perhaps the most significant innovationthat social funds have brought about hasbeen their role as means of investing inbasic infrastructure. Whereas traditionaldelivery lines—namely, the line min-istries—tended to be relatively inefficientand focused on infrastructure in capitalcities (perhaps due to political pressures, togreater visibility of problems, or simplybecause of economies of scale and ofagglomeration), social funds, on the other



hand, were able to channel investments intopoorer, hitherto unreached rural areas andto do so at lower costs and greater speedsthan the line ministries. An important rea-son for this efficiency is that the funds con-tract with third parties to implement theworks, which brings performance-basedincentives into play. Likewise, their stafftends to be recruited competitively, withemployment contracts more along the linesof those offered to private sector employees.There is considerable evidence that socialfunds have been effective in expanding ser-vice delivery. They have also benefited fromrelatively careful evaluation (Rawlings et al.2004) that often showed positive effects inoutcomes for the poor, with lower operatingcosts than for most comparative programs.

Social funds essentially created a new,shorter route of accountability between theproviders of small infrastructure invest-ment, on the one hand, and local communi-ties, on the other.17 The relationshipbetween the fund and the communities itserved was generally stronger than had been

the case with traditional government agen-cies or local governments, especially in cen-tralized regimes or in societies characterizedby the presence of local strongmen. Socialfunds’ new accountability arrangementshave allowed communities to express theirneeds (including making choices among avariety of sectors) and to stay informedregarding investment decisions and worksprogress. Rules often furthered thataccountability by establishing social controlmechanisms, as part of supervision—inother words, communities had to sign offbefore contractors were paid.

Several “bureaucracy avoiding” account-ability mechanisms operating in the case ofsocial funds can be found in conditional cashtransfer (CCT) programs. In the recent past,CCT programs18 have become a popularsocial assistance approach as they have beenshown to generate important gains regardingboth the provision of immediate assistanceand longer-term human capital develop-ment impacts (Rawlings and Rubio 2003).Through provision of cash grants directly topoor households, conditional on humandevelopment–related activities such as chil-dren’s school or clinic attendance, CCT pro-grams allow central governments to have adirect relationship with the target popula-tion, at least in terms of beneficiary selection,without the intermediation of local authori-ties or without having to use local serviceproviders to reach the target population todeliver social assistance.19 Indeed, the centralgovernment commonly administers CCTprograms directly, including identifying ben-eficiaries and verifying compliance anddelivery of cash transfers. Making thesefunctions operate in an efficient way (that is,having a well-designed and -maintainedobjective targeting system, making paymentsin time to large numbers of poor familiesdispersed throughout the country, and soon), depends crucially on the existence of arelatively small but technically solid andpolitically protected central bureaucracy—something that most line agencies in thesame country have failed to achieve.

Limits to alternative compacts. The com-mon denominator of these various innova-tions is the bypassing of well-established


After taking various administrative reformapproaches over the years, Jamaica intro-duced the creation of executive agencies(EA) as part of a broad public sector mod-ernization program.While previous initia-tives attempted some redesign as part ofcivil service reform, the system was still sub-ject to central controls over resource inputsand did not address performance manage-ment. EA, which are characterized by orga-nizational separation, offered Jamaica a wayto improve public sector functioning andenhance service delivery.

In creating the EA, Jamaica incorporatedessential features of new publicmanagement. It also made significant leg-islative and institutional changes to elimi-nate impediments arising from traditionalcentral control over allocation of financialand human resources in line departments.One such feature was to make the EA chiefexecutive officers (CEOs) directly responsi-ble to their portfolio ministers, with incen-tives to produce specific outputs (that is,rewards and sanctions were given to theCEOs for good and bad performance,

respectively). Delegating authority to theCEO to manage the financial and humanresources needed to produce results was amajor achievement of the reform.The CEOscould “hire, fire, and discipline” their employ-ees; enabling them to maintain control overthe allocation of most inputs.

The Executive Agencies Act (EAA) provi-sions that Parliament passed in 2002 incor-porate a strategic planning process toassess each agency during its restructuring,including its future organizational structure,staffing, and remuneration requirements.Since then, EA developed new performanceindicators and the agencies are now evalu-ated regularly. All agencies have preparedcitizens’ charters as agreements with theirclients on the agencies’ performance. Somehave incorporated customer satisfactioncomponents among their key indicatorsand have reported satisfactory levels in fol-low-up customer surveys.The EA improvedoperational efficiencies and was able tocharge reasonable user fees, in cases whereconsumers are “identifiable” and it isefficient and equitable to do so.

B O X 3 . 5 Executive Agencies in Jamaica


traditional state bureaucracies (central andlocal) in favor of compacts either with pri-vate for- and not-for-profit actors or with“elite” state agencies or units for which per-formance-based management is feasible (oreasier) to establish. There are many, mostlywell-documented successful examples ofsuch initiatives. At the same time, far frombeing a silver bullet to service deliveryreform, bypassing state bureaucracies alsoraises a series of broader (mostly longer-term) common concerns.

First, in most cases, success of theseapproaches depends on the consolidationof an “efficiency enclave”—for example, thesocial fund, the regulatory agency, or theCCT program management unit. Eventhough enclave agencies may be a good,second-best solution in some contexts, theyare hardly foolproof (Laking 2002). Of par-ticular concern is the potential lack of insti-tutional sustainability as often an enclaveagency’s short-term performance is sus-tained with exceptional political supportand government commitment (often by the presidents themselves) and, thus, arevulnerable to waning commitments thataccompany change of governments orshifts in the political environment. Moregenerally, the question as to how and whysuch enclaves are spared the typical clien-telistic force that weakens the quality ofpublic sector employment remains unan-swered. What type of political calculationled political leaders to provide specialadministrative procedures and practices toa selected few agencies? It is possible tospeculate that such approaches are moti-vated by an idiosyncratic need to ensure theeffective implementation of high prioritypolicies or initiatives (for example, estab-lishing credibility of a privatization reformby creating an independent regulatoryagency or creating an effective social pro-gram for social containment reasons). Insome cases, the donors’ role may be equallyimportant as a source of pressure and/orencouragement for such approaches.

Second, even when key agency sustain-ability is not an issue, there may be negativeconsequences of such islands of efficiencyoperating in the context of an unreformedpublic sector. The arms-length relations

between the government and these enclaveagencies and the latter’s exemptions frommany public sector regulations have some-times led to loss of effective central control,and, at times, corresponding opportunisticbehaviors. Some enclave agencies haveengaged in irresponsible or reckless finan-cial activities (for example, excessive bor-rowing or overspending) or in patronage-driven personnel practices.

Similarly, some CCT program criticsargue that, while they represent a creativeapproach to providing social assistance,they constitute an “end-run” around themore difficult task of reforming inefficientpublic education and health services. Thereare concerns that without greater attentionto the provision of quality services, CCTprogram conditionalities run the risk ofmandating the poor’s use of low quality ser-vices, tying them to ineffective serviceproviders.20 These supply-side concernshave led to calls for renewed attention to thebasic task of providing accessible, high-quality health and education services inpoor areas. A related concern involves thelocal government’s role. Although adminis-trative arrangements vary considerablyacross programs, CCT programs (especiallythose with highly centralized administra-tion such as Colombia and Mexico) havebeen accused of undermining local govern-ments’ effectiveness by bypassing theirauthority. This concern has particular reso-nance in countries where democraticallyelected local governments are in theirinfancy and where central governmentshave a long tradition of clientelism andpaternalism. Strong centralization also lim-its the program’s ability to address benefi-ciaries’ needs and build in local responsemechanisms for basic operational taskssuch as targeting, verifying compliance withconditions, or addressing beneficiaries’ con-cerns. It may also limit the program’s abilityto effectively coordinate with health andeducation service providers.21

These concerns have also led analysts toask under what conditions will the “out-sourcing” of functions create opportunitiesfor strengthening mainstream governmentagencies in charge of policy making. Forexample, contracting out health services to



NGOs in Central America implies newtypes of activities by ministries of healththat require the development of new capa-bilities, such as the ability to manage a com-petitive bidding process and to oversee,monitor, and evaluate contract compliance.Indeed, the development of these purchas-ing arrangements is seen as an opportunityto develop these skills via learning-by-doing. To date, perhaps with the exceptionof Guatemala, such changes have not mate-rialized (Fiedler 2002).

Similarly, many project documents andcharters from the mid-1990s suggested thatsocial funds would play a role only tem-porarily in the delivery of small infrastruc-ture services, until the line ministries werestrengthened and better able to performthese functions themselves—when it wasassumed that they would turn the functionsback to the line ministries. For a number ofreasons (including the funds’ administrativeefficiency which made them attractive togovernments and donors), in most cases thistransfer did not materialize. Governments inmany countries began using social funds asthe de facto and, at least in the case of Hon-duras, official principal agencies for small-scale social and economic infrastructure. Itremains an open question whether such spe-cialization is in the end effective in enablingsector ministries to concentrate their func-tions on policy-making and regulatoryaspects—a stated goal of the movement toinstitutionalize the role of social funds.22

Third, there is a broader issue on the dif-ficulty of balancing autonomy (for exam-ple, of social funds, CCT units, or regula-tory agencies) and ultimate accountability.The many successes throughout the regionof such approaches cannot obscure the realdanger of capture by clientelistic politiciansor by private suppliers and concessions;even enclaves are embedded in sociopoliti-cal systems. For example, analyses of themajor Mexican Social Fund scheme ofPronasol (or Solidaridad) document how itwas systematically used for clientelistic pur-poses (Diaz-Cayeros et al. 2002; Molinarand Weldon 1994). For the Peruvian SocialFund of FONCODES, while Paxson andSchady (2002) documented its success interms of pro-poor targeting, Schady (2000)

shows that President Fujimori specificallytargeted it using political criteria (targetingpotential swing areas near election time).Similarly, case study work by Tendler (2000)in the northeast of Brazil documents, in thecases analyzed, how “demand-driven” ini-tiatives were often captured, including onthe supply side, by private companies thatoffered communities particular productsthat they could deliver and then suggestedforming local committees to give the stampof participation.23

Problems of capture have also been iden-tified in the regulation of contracts with theprivate sector, which remains a major chal-lenge, notwithstanding substantial advancesin creating legal frameworks and regulatoryinstitutions. Key problems are the difficultyof ensuring an arm’s length relationshipbetween the regulator and the executivebranch and the challenge of balancinginvestor and consumer interests in societieswith a limited tradition of consumer organi-zation and participation (see the next sec-tion). Newly privatized companies becomepart of new distributional coalitions, wherelarge conglomerates play an important rolein contributing to a perception of concentra-tion of power and profits. A series of LatinoBarometro polls, for example, find that agrowing proportion of those questioned dis-agreed with the statement that privatizationhad been beneficial for their countries. Manycritics are concerned with how fairly privati-zation’s benefits have been distributedamong different stakeholders and, in partic-ular, the extent to which the state and privateinvestors may have benefited at the expenseof consumers and workers. This phenome-non can be attributed partly to biases in theoriginal design of privatization transactionsthat tended to emphasize fiscal concernsover social and economic considerations. Itprobably also reflects deficiencies in theongoing regulatory process, which has notalways been successful in transferring effi-ciency gains into lower tariffs for customers.Indeed, the challenge of developing an effec-tive counterweight in noncompetitive sec-tors has been difficult due to classical prob-lems of regulatory capture and the commoninterest of governments and privatized utili-ties in high profits.



As the previous review indicates, whenpublic bureaucracies don’t work well, short-circuiting them by contracting out servicesto private firms and NGOs or setting upparallel, more independent agencies incharge of service delivery is showing somegood results, at least in the short run. Thereare many indications, however, that in thelonger run both their sustainability and fulleffectiveness depend on whether the rest ofthe state apparatus improves its capacityand becomes more performance-oriented.

All in all, to initiate and sustain good,effective service delivery compacts (withinor outside the state), at a minimum,requires the basic commitment and supportof policy makers. As the examples reviewedin this section show, LAC countries havebeen successful in developing such com-pacts, particularly when they transferredmanagement responsibilities outside themainstream state bureaucracy. Nonetheless,many questions remain as to the politicalincentives at play to explain the genesis, sus-tainability, and capacity to scale up (or out)of such innovations. We speculate that theoverall success of compact reforms isrelated to broader efforts to enhance bothclient power and citizen voice.

Enhancing Client Power When clients have power over providers, ashorter, more direct form (that is, lessdependent on actions and performance ofother actors) of accountability for resultsmay arise. Over the last two decades, manyLatin American countries have exploreddifferent ways of enhancing client power. Insome cases, this entailed giving individualsthe power of choosing which organizationthey get their services from, with the expec-tation that this will create better (market-like) incentives to providers. In other cases,it involved giving clients a direct (collective)role in service management, with the expec-tation that such involvement should resultin stronger control over the actions offront-line providers and a correspondingpositive change in the quality of servicesprovided.

Unlike most of the changes discussed inthe previous section, these transformationshad service delivery improvement as a pri-

mary objective. As a result, many of theseexperiments have benefited from evalua-tions that seek to identify their success inachieving service delivery objectives. Wewill first review the experience with clientchoice and next with client participation inthe management of services.

In the limited areas in which it has been implemented, givingindividuals a choice over whichorganization they get services fromhas generated powerful pressures forresults on front-line providersThe introduction of choice as a mechanismto promote client power over serviceproviders has been a powerful one, albeit ofrestricted coverage. In the infrastructuresector, the 1990s reforms resulted in thegradual introduction of competition forelectricity only (in the case of a relativelysmall number of large industrial con-sumers) and for long distance and cellulartelephony. In the water sector, competitionhas been limited to bidding out concessioncontracts. However, even there, the extent ofcompetition has been quite limited due tothe relatively small number of multination-als active in the sector (Foster 2002). Rela-tively little attention has been paid to thecompetition that arises from small-scaleindependent providers of water services,such as water tankers and secondary waterretailers, particularly in underserved peri-urban areas (Solo 2003).24

In education services, two key experi-ences with choice are those of Colombiaand Chile (see box 3.6). In the first half ofthe 1990s, Colombia introduced a sec-ondary education targeted voucher pro-gram, which involved approximately one-fifth of the country’s municipalities andclose to 2,000 private schools. The mainobjective was to increase the number ofpoor students transitioning from primaryto secondary education. Chile, on the otherhand, represents a unique case in the sensethat a nationwide system of “quasi-vouch-ers” directed to all municipal schools andnonfee-charging private schools was estab-lished with the main objective of promotingcompetition among schools.



In the health sector, once again Chile andColombia introduced paradigmatic reformsthat put competition and choice at the cen-ter. In the 1980 context of a complete over-haul of its social security system, Chilereformed its social health insurance system,introducing private-public competition formandatory health insurance. The reformallowed all formal workers, independent oftheir income level, to choose between a pri-vate health insurer (ISAPRES) and the pub-lic national health fund (FONASA) formandatory health insurance coverage. Thereform, however, mandated that public sub-sidies would be given only to FONASA ben-eficiaries irrespective of the income and riskcharacteristics of those opting for ISAPRES.In 1994, Colombia also introduced a radicalsocial health insurance reform that, as in theChilean case, promoted competition among

insurers, but also introduced demand-sidepremium subsidization for the poor.Colombia also introduced an explicit benefitpackage and a risk/income equalizationfund, both important instruments to reducethe market segmentation incentives on arisk basis.

Overall, the results of competition havebeen very positive in all sectors where it hasbeen tried. In infrastructure services, intro-ducing competition in cellular telephonyreduced charges and catalyzed a major cov-erage expansion. The number of cellularlines in service grew 500-fold during the1990s (at a 57 percent annual averagegrowth rate), to the point where, by the year2000, cellular telephone penetration hadovertaken fixed line coverage in countries asdiverse as Bolivia, Chile, Guatemala, Mex-ico, and Venezuela. As a result of competi-


Chile’s higher education reforms of the 1980s,instituted under the military government, werethe most radical reforms of higher educationundertaken in Latin America.Those reformssought to increase learning and enrollmentwhile decreasing public spending on highereducation.These reforms successfully expandedschooling access, though their immediate effecton schooling quality has been unclear.

Competition between school types. InChile’s initiative for diversification, sepa-rate tiers were created for universities,Professional Institutions, and Centers forTechnical Education. In Chile’s deregula-tion effort, private institutions were cre-ated and allowed to operate with minimaloversight. In the 1980s, 40 private univer-sities, 78 private Professional Institutes,and 61 private Centers for Technical Edu-cation were created. In its decentralization

effort, Chile’s two oldest public universi-ties were subdivided into 16 new universi-ties, including pedagogical institutions.Overall, the purpose of these reforms wasto offer students more choices and toforce educational institutions to innovatein competing for students.

Competition for funds. Chile decreased pub-lic expenditure on higher education, butsought to increase funds from privatesources (that is, charging tuitions) tocompensate for the lower spending.Thegovernment also redesigned a smallportion of its spending in order to allo-cate some funds based on quality andneed, while addressing credit marketconstraints that could cause students tounderinvest in their own human capital.

Outcomes. Between 1983 and 2001, enroll-ment in higher education rose from

about 200,000 students in 1983 to about475,000 students in 2001. Also, the num-ber of Master’s students increased fromabout 1,500 students in 1982 to 8,000students in 2000; the number of doctoralstudents increased in similar proportion.Box Figure 1 shows that this increasewas more than double the meanincrease for Latin America, and slightlyabove the mean increase for OECDcountries.The immediate effect on qual-ity of schooling, however, was unclear.Partly due to the enrollment of manylower-achieving students, learning mayhave decreased in the short run. In 1999,to address concerns about quality ofeducation, national accreditation com-missions began setting program stan-dards and recognizing programs thatmet those standards.

B O X 3 . 6 Higher Education Competition: Chile’s Reforms in the 1980s

0

50

100

150

200

250

300

Mexico Brazil Argentina LA Average Uruguay NewZealand

Spain OECD Colombia Chile UnitedKingdom

Portugal

Increase in Tertiary Education Coverage, 1980–97 (percent)


tion, cellular call charges fell between 30and 50 percent in Bolivia, Peru, andVenezuela, largely as a result of the intro-duction of the “calling party pays” system.Similarly, long distance telephony liberal-ization drove down international callingcosts by 50 percent in Chile, Colombia,Mexico, and Peru, and by more than 70 per-cent in El Salvador and Guatemala. Even innoncompetitive sectors, the move towardcost-reflective charges (see box 3.7), com-bined with the establishment of customercomplaints channels, have begun to makeconsumers more demanding about thequality of service they receive.

The Colombia education voucher expe-rience was thoroughly evaluated (Angristand others 2002, 2004). Results indicate thatin the participating municipalities, the pro-gram provided an effective way to increaseeducational attainment and achievement(college-entrance test scores) for the poorstudents benefiting from the vouchers. Crit-ical were the sound criteria in the selectionof the participating municipalities andschools and the application of effective tar-geting mechanisms, with the use of perfor-mance incentives for participating schoolsand beneficiaries. Despite its positiveresults, the lack of the government’s com-mitment, the teacher union’s opposition,and administrative difficulties (disburse-ment delays and program monitoring bur-den) led to the program being discontinuedas a national initiative in 1997. Neverthe-less, some local governments continuedimplementing similar approaches withoutnational support. Beyond the idiosyncrasiesof the factors leading to the demise of thisparticular experience, the message appearsto be that, unless the ultimate beneficiariesof the increased choice have sufficient voice(that is, they are sufficiently empowered) todefend their gains from those whose inter-ests such schemes affect, sustainability maybe a serious bottleneck.

The Chile “education voucher” experi-ence generated a range of evaluation studiescentered on the relative performance ofpublic and private schools. Overall, with theintroduction of the per-student subsidy,enrollment increased considerably in theprivate voucher schools (from 14 percent in

1980 to 32 percent in 1990), which, control-ling for socioeconomic status, perform bet-ter than municipal schools (Carciofi,Cetrángolo, and Larrañaga 1996; McEwanand Carnoy 2000), pointing to positiveeffects of competition on the privateschools’ performance. However, it is gener-ally recognized (McEwan and Carnoy 1999,2000; Hsieh and Urquiola 2003) that a lackof incentives and municipal sector capacityto compete,25 a lack of clear information onrelative school performance, and some“cream-skimming” among private schoolshampered competition among private andmunicipal schools.

In the 1990s, further municipal auton-omy constraints (for example, the 1991teacher statute) and centrally driven inter-ventions in favor of the poorest/weakest


An integral element of the 1990s infrastruc-ture reform process was to ensure that util-ity tariffs reflected accurately the full eco-nomic costs of service provision.Cost-reflective tariffs are needed tosafeguard service providers’ financial sus-tainability and hence, politicalindependence.This principle had varyingimplications according to the sector andlocality, and it is difficult to generalize aboutthe experience. Indeed, a recent studyfound that out of 10 privatization casesstudied, prices rose in five and fell in theother five (McKenzie and Mookherjee 2002).

The same study shows that in mostcases, welfare losses arising from higherutility tariffs that the reform triggered aremore than compensated for by the welfaregains associated with expanding access toservices (McKenzie and Mookherjee 2002).The reason is that the budget shareallocated to the utilities is relatively low, andhence tariff increases in the cases studied(Argentina, Bolivia, and Nicaragua)accounted for less than 1 percent of house-hold budgets, while the gains from accesscan be greater than 10 percent of budgetsfor the poorest families, as households areable to abandon more expensivesubstitutes to utility services. In many coun-tries, the reform of service tariffs did notadequately evaluate the extent to whichcost-reflective tariffs would be affordable tothe poorest clients, or investigate appropri-ate compensating mechanisms.

Traditional cross-subsidy schemes areoften poorly designed, failing to target sub-sidy resources effectively to the poor. In LaPaz, Bolivia, for example, the average waterconsumption of a low-income household isaround 5 cubic meters per month, and theaverage consumption for a higher incomehousehold is around 25 cubic meters permonth, showing a clear differentiation.However, the water utility provides a subsi-dized rate on the first 30 cubic meters permonth, so that higher income householdsobtain the largest subsidy benefit. Even incountries like Chile and Colombia, whichhave developed more sophisticated individ-ual or geographical means for targeting sys-tems for water subsidies, significantdeficiencies remain. Gomez Lobo and Con-treras (2000) show that about 65 percent ofthese subsidy resources go to householdsin the top 70 percent of the income distri-bution, while 55 to 75 percent of the benefi-ciaries come from the top 70 percent of theincome distribution.

The accumulated evidence suggests thatthe transition toward cost-reflective usercharges needs to include a strategy to evalu-ate and safeguard affordability of a subsis-tence basket of utility services to the pooresthouseholds. In some circumstances, it maybe much more important from a social per-spective to focus scarce subsidy resources onreducing access costs for the unconnected,rather than reducing prices for those whoalready have access to the service.

B O X 3 . 7 User Fees


municipal schools complicated even morethe working of competition between munic-ipal and private schools.26 On the otherhand, dissemination of standardized schoolscores (SIMCE scores), increased adminis-trative and pedagogical municipal schoolautonomy; introduction of school cofinanc-ing and improved monitoring of studentattendance improved the voucher system’soperation. The studies centered on the1990s generally confirm that private subsi-dized schools perform better than munici-pal schools (Sapelli and Vial 2002) and haveimproved their performance due to strongincentives to compete (decision-makingautonomy, reliance on the voucher or parentcofinancing, and so on), therefore conclud-ing that competition and choice were help-ful in improving private schools’ quality andefficiency (Gallego 2002). However, compe-tition’s impact on municipal schools’ perfor-mance remains more controversial, withsome studies (Gallego 2002) finding thatimproved municipal performance waslargely due to competition, while others(Schiefelbein and Schiefelbein 2000; Coxand González 1998) assert that it was due tothe central compensatory programs.

Turning to the health sector, social healthinsurance reform in Chile resulted in a sig-nificant increase of both private healthinsurance coverage and private provision ofhealth services. At the same time, it alsoresulted in a severe segmentation of the riskpool, with the high-income and low-riskgroups concentrating in the private(ISAPRE) system and the low-income andhigh-risk concentrating in the public one(FONASA).27 The Colombian reform, onthe other hand, is expected to address a keyfactor that determined income market seg-mentation in the Chilean case through theintroduction of subsidized insurance premi-ums for the poor. Preliminary evaluations ofthe Colombian reform (Bossert et al. 2003)indicate an increase in service utilizationand coverage for the population at large.However, the evaluations do not distinguishclearly the specific effects on the poor, northe extent to which reforms improved thefinancial protection of households.

In general, the evaluation results fromthe experiences of expanded choice are

encouraging. Nevertheless, this should notobscure the fact that, as illustrated in thecases discussed above, the conditions forchoice to work (particularly for the poor)are not minor. On the demand side, theexistence of highly unequal economic andsocial conditions in most countries in theregion require active public policies to com-pensate for extensive informational asym-metries (for example, on the quality of edu-cation services as in Chile in the 1990s) anddifferences in purchasing power and needs(for example, by introducing income/riskequalization mechanisms for health insur-ance in Colombia). Without such compen-satory actions, increased social stratificationbecomes a significant danger. On the supplyside, the presence of independent providersoperating in a competitive environment is,naturally, key to the proper functioning of amodel based on choice. Both the existenceof structural bottlenecks to competition(for example, in the case of many utilities orin schooling in dispersed rural areas) andthe social desire to ensure minimum qualitystandards (for example, safe water or prede-termined learning goals) are satisfiedthroughout the country require, again, theneed for government intervention in servicedelivery, or most likely, in regulation.

Client participation in themanagement and operation ofservices has also shown good resultsfor services that allow a relativelysmall production scaleClient participation in service managementis the second way that client power oper-ates. The experiences with community-runschools, health clinics, and rural water sys-tems in a diverse group of countries providethe empirical basis to assess the power ofclient participation in service delivery.Overall, the evidence suggests that directclient involvement in the management ofservices can yield positive results when theservices involved allow a relatively smallproduction scale and rather simple man-agement challenges.

The experience of community-run schoolsin El Salvador and Nicaragua provides solidevidence on the benefits of a participation



approach to client power. Each school isrequired to establish an elected council—composed only of parents in El Salvadorand of parents, teachers, and the schooldirector in Nicaragua. The school must alsosign a special agreement with the Ministryof Education under which the latter decen-tralizes to the councils the management ofall funds, allowing them to manage allteachers, take care of operating and mainte-nance costs, and even make academic deci-sions. “Client power” is therefore expressedthrough active participation in school man-agement. In addition, in Nicaragua, clientpower also finds a way through the systemof national, departmental, and local educa-tion councils set up to ensure broad partici-pation at all levels.

The key aspect of this model is the controlthat parents have over teacher managementdecisions—exactly the aspect that has proventhe most controversial and hard to reform instate-managed school systems. In Nicaragua,school-based decisions on teacher manage-ment are shown to have a positive impact ontest scores (King and Ozler 1998). In El Sal-vador, enrollment in the rural areas increasedmassively following Educación con Partici-pación de la Comunidad (EDUCO), and ithas been shown that school-based decisionson teacher management have a positiveimpact on class attendance (Jiménez andSawada 1998).

The El Salvador and Nicaragua cases arethe most thoroughly evaluated experimentswith community-run schools, but there area growing number of such experiences inCentral America (for example, Guatemalaand Honduras) and elsewhere in the region.For example, the much referred to experi-ence of the Brazilian state of Minas Geraiswith community councils and electedschool principals (Guedes et al. 1997) is oneof many examples of a growing trendtoward school-based management experi-ences in Brazil (see Aglaê de MedeirosMachado 2002). Overall, the evidence israther strong in the sense that decentralizingresponsibilities to the school, particularly inthe teacher management area, makes it pos-sible to maximize the use of local informa-tion and accountability mechanisms. It islikely that, given the peculiarities of the edu-

cational process (proximity to the finalusers, continuity in the provision of the ser-vice), the scope for user participation will beparticularly important in that sector.

In contrast, the evidence on effectiveexperiments with community-run healthcenters is rather limited. One interestingexample is that of local health administra-tion committees (CLAS) in Peru. These areprivate, nonprofit community associationsbuilt around a health center or post (Cotlear2000). The CLAS prepare local health plansthat become the basis for a contract withhealth authorities. The contracts aredesigned to finance outputs rather thaninputs, and the CLAS are held accountablefor reaching specific targets (derived fromthe plans) and for the use of funds. TheCLAS are given flexibility in the use ofresources relative to state-run centers. Whilethere is no systematic evaluation of theapproach’s impact, the data do show animprovement in health indicators (for exam-ple, lower waiting time in CLAS centers).

There are a number of factors that seemto conspire against community participa-tion in health centers compared to that inschool management. First, in most casesusers of health services have a very intensive(when sick), but much less frequent, rela-tionship with providers than parents ofchildren have with schools. In that sense,solving the collective action probleminvolved in community-run services islikely to be more difficult in health. Second,informational asymmetries tend to be verylarge in the case of health services. As aresult, the opportunities for health profes-sionals capturing control under commu-nity-run schemes is likely to be higher thanin the case of schools. Overall, these factorsmay help explain the observed differencesin the popularity of the approach betweenthe two sectors.

Unlike the case of health centers, commu-nity committees (juntas de agua potable) andcooperatives throughout Latin America havesuccessfully managed rural water and sanita-tion services. Over the years, as coverage lev-els increased in rural communities, it becameimpractical and expensive for centralizedservice delivery agencies (usually located inthe Ministry of Health, Public Works, or



Urban Development) to take charge of oper-ating and maintaining water systems spreadacross the country. Overall, the community-management approach has proven to behighly successful for rural water services.

Approaches to community-based servicedelivery have evolved from water systemsconstructed by a national agency and subse-quently “handed-over” to communities formanagement, to more demand-responsiveapproaches where the community partici-pates in all phases of the project cycle andshares in the investment costs of projects.Chile provides a good example of the firstapproach (Alvestegui and Sara 2003), which,while generating impressive results in termsof coverage,28 may have led to insufficientcommunity attention to finding lower costsolutions and to investing in system rehabil-itation or expansion. Community participa-tion advantages are perhaps felt morestrongly in the latter approach, followed incountries such as Bolivia and Ecuador; thisapproach aims to ensure communitiesobtain a system for which they have thefinancial willingness and technical capacityto maintain it (World Bank 2002c).

The details of the service-specific autho-rizing environment—in particular howflexible and consistent are the rules andconstraints under which communitygroups need to operate and the nature andenforceability of accountability relations tothe authorities delegating powers—can bevery influential in the outcomes of theexperiences of community-run services.

In the case of community-run schools inCentral America, for example, both in ElSalvador and Nicaragua the recorded suc-cesses are, in part, linked to the existence ofa solid “compact” linking the school withthe Ministry, which seeks to ensure the ful-fillment of coverage, efficiency, and equitytargets. A weak (or weakly enforced) “com-pact,” on the other hand, implies that schoolautonomy and strong parent involvementin financial and administrative matterscould produce, for instance, cases whereteachers are unfairly fired or to cases of mis-handling of funds. There is, in fact, somepreliminary evidence from El Salvadorshowing that as the Ministry weakened itscontrol of schools, such cases begun to

appear. Similarly, the fact that ministries ofeducation in Guatemala and Honduras hadless capabilities to enforce “school com-pacts” than their counterparts in El Sal-vador and Nicaragua may be the principalreason for the difficulties experienced intheir community-run schools.29

The overall authorizing environment isalso crucial when elements of community-run services coexist with nonparticipatorymechanisms. As the experience with watersupply programs in rural areas (Gran et al.1997) showed in some cases, even thoughcommunities took educated decisions basedon the participatory program in question,since the construction of the systems con-tinued to be the responsibility of a “nonre-formed” state agency, there was no way theycould ensure that the works be conductedaccording to plan. In other words, demand-orientation must take place consistently inall the stages of the provision process(financing, production, delivery, and so on)otherwise the “weaker” link of the chainends up dominating the rest.

Beyond establishing the authorizing envi-ronment, central governments can developcomplementary mechanisms to support theoperation of community groups. In manyways, social funds are evolving in that direc-tion. The latest generation of funds isbecoming a mechanism that both empowerscommunities and local governments andchannels national preferences (such as apoverty targeting emphasis) in the form ofincentives to local actors. Community par-ticipation is now viewed as an end in and ofitself—almost on par with the goal ofincreasing access to basic infrastructure—and funds have incorporated not just newrules of the game (that is, in the operationsmanuals), but also specific training geared tostrengthen community capacities for plan-ning and implementation.30 Similarly, theexperience with rural water and sanitationprojects indicates the importance of provid-ing appropriate community outreach, orga-nization, and training programs to comple-ment hardware investments. The approachhas required centralized agencies to trans-form their roles from service providers tofacilitators. A new set of intermediaries, usu-ally specialized NGOs and small contracting



firms, have emerged to provide assistance tocommunities in formulating their demands,building capacity, and implementing pro-jects, while the central government providesthe normative framework, access to funding,and technical oversight.

Both the opportunity for and ultimateresults of client-participation schemesappear to be context-dependent: nationaland local political conditions, communityorganization traditions, and the strength offront-line providers (for example, teacherunions) matter in explaining outcomes. Forexample, both the El Salvador and Nicaraguaeducation experiences started after a civilwar and, therefore, were also part of a pacifi-cation and democratization attempt. Theywere initially implemented on a subnationalscale to be extended to different extents. Thisnot only reminds us of the idiosyncrasiesbehind these cases, but also that the politicalclimate may have been particularly welcom-ing to such initiatives given the need to con-solidate democracy to which such participa-tion was functional.31 In fact, the EDUCOexperience grew out of community effortsduring the civil war: it was a case where a rel-atively weak national state took advantage ofrelatively strong communities (rather than astrong state empowering weak communi-ties) to respond to growing social needs inthe postwar scenario.

Sustainability can be an issue when relyingtoo much on community-based managementmechanisms. Community-based schools, forexample, are often quite isolated from thestate’s overall institutional structure (includ-ing the absence of links with the intermediateand local government levels) and have devel-oped at the margin of the teacher unions; inthis context, they are particularly vulnerableto political change.32 Similarly, the CLASexperience in Peru indicates that while com-munities received them with great sympathy,they encountered significant resistance in astate bureaucracy that resented the impliedloss of control (Cotlear 2000).

Client power relationships are affectedby the incentives, motivation, and behaviorof front-line workers. These relationshipsare particularly embedded in the localsociocultural conditions, the behavior pat-terns, and collective expectations produced

by unequal group-based relations. This canbe of particular importance where there isgreater social distance between front-lineworkers and their “clients,” as with indige-nous students and nonindigenous teachers.The ways in which poor groups interactwith front-line providers will be a functionof both the larger social context and histo-ries of mobilization and engagement. At thesame time, when communities are offeredsufficient support, client power can con-tribute to changing the local balance ofpower, whether this is through increasedparental influence in EDUCO schools in ElSalvador, the introduction of culturallyappropriate practices to birthing and post-natal care in Quechua areas in Cusco, Peru,(see box 3.8, above), or the provision of the


Malnutrition in Honduras dropped rapidlythrough the 1980s, but progress stalled in1990.The Ministry of Health required publichealth centers to detect low growth (stunt-ing) among children, then to use stunting asa criterion for distributing health services.By 1994, the Ministry concluded that theprogram needed closer community ties tomaximize impact and designed the Com-prehensive Attention to Childhood in theCommunity program (AIN-C) for that pur-pose. AIN-C targets all children less thantwo years of age in poor communities.Vol-unteer monitors meet with mothersmonthly and plot the infant’s growthprogress on a simple bar chart. In AIN-C, it iscommunities that request the programbefore it begins operations, decide howmany monitors will operate, and select themonitors.The monitors give quarterlyreports to the community with focus oncommunal obstacles to good health, suchas contaminated water or a lack of sharedchildcare.They negotiate with mothersabout specific needed practices to improvea child’s nutrition and decide whether toreach mothers by house-to-house visits,neighborhood sessions, communitywideevents, or other means.

AIN-C has been shown to increasebreastfeeding to children under six monthsof age, improve provision of oral rehydrationsolution to children with diarrhea, expandthe giving of vaccines to infants, and spreadother preventative health measures. AIN-C

has about one-ninth the cost of the firstfacility-based program and has had far moreextensive reach than that program had.

Community participation also impliesthe potential to incorporate adaptation ofservice delivery to local culture. One clearexample comes from a maternal and childhealth project in the Paruro (Cusco)Province in Peru where an effort was madeto improve the number of women fromrural areas who attended the clinic for child-birth. Findings from earlier interventionshighlighted that women from remote ruralareas were unwilling to leave their childrenfor an extended period, found the cliniccold, and felt uncomfortable using formalWestern birthing procedures.The projectestablished a “waiting house” where womennearing childbirth could be lodged and fedin a house close to the clinic, installed extraheating in the birthing room, and providedan alternative bed to the standard Western“birthing table.”The result was a dramaticrise in medically attended births and asharp fall in postnatal mortality of bothbabies and mothers.The practices have alsobeen followed by many other health cen-ters in rural areas in Peru with the same dra-matic results.This is a clear demonstrationof how attention to cultural issues can pro-vide great advances in fighting poverty—inthis case on the health dimensions—and inthe general development effectiveness ofprojects and programs.

B O X 3 . 8 The Role of Community Participation andCultural Adaptation in Service Delivery


CCT grants to mothers, which combinedwith the election of mothers’ local represen-tatives to serve as conduits between benefi-ciary families and the CCT program, intro-duced changes in empowerment dynamicsthat are playing out in households andcommunities throughout the region (Adato2000; and Adato et al. 2000).

The evidence reviewed indicates thatclient participation can compensate for theweakness of long-route mechanisms. Ulti-mately, however, the effects and sustainabil-ity of such client power mechanisms aredependent on both the existence of capaci-ties (agency) among the poor and minimumstate capacities to initiate and regulate theirimplementation. When and if the condi-tions to initiate such approaches exist, a vir-tuous circle may be generated wherebyclient participation continues to generatecapacity-building opportunities for thesocial organizations of the poor and disad-vantaged and of individuals and families—factors that should contribute to making thestate more accountable overall to citizens.

Overall, the evidence strongly suggeststhat taking decisions away from publicbureaucracies (central or local, old or new)into the hands of individual or group ofbeneficiaries is proving to be a powerfulidea. The growing number of evaluationstudies available shows solid results regard-ing the benefits of initiatives oriented toenhance client power (through choice anddirect management control of front-lineservice providers), even though the spread(in terms of the number of services andnumber of people that benefit from them)of such approaches is still not too wide. Theevidence also suggests that client power asservice delivery mechanisms is also influ-enced by the overall strength of stateaccountability: weak overall state account-ability is likely to undermine service deliv-ery improvements that are achievedthrough enhanced client power.

Enhancing State Accountabilityto CitizensUltimately, the extent to which the state isaccountable to citizens (particularly poorones) constitutes the key factor in deter-

mining whether service delivery systemswork for people. To be sure, whether or notpolicy makers face the incentives to addressthe needs and preferences of citizens (thatis, whether they are accountable for results)is fundamental when it comes to carryingforward and implementing the reforms tothe compact between policy makers andproviders discussed above. And even if theclient-power alternatives discussed aboverequire less “hands-on” involvement bystate agents in the actual delivery of ser-vices, whether and how such reforms areimplemented and supported over time isclearly influenced, once again, by the natureand intensity of the accountability relation-ship linking policy makers and differentgroups of citizens. Said differently, politics,political institutions, and citizen voice con-stitute fundamental determinants ofwhether services work for people—particu-larly poor ones.

In this section we review the evidence onhow some fundamental political and socialtransformations in Latin American coun-tries over the last two decades have affectedstate accountability for service delivery.Indeed, the sharp expansion of electoraldemocracy and constitutional rule (and theassociated expansion in freedom of expres-sion and association) is one of the moststriking transformations that took place inLatin American over the last two decades.How these changes affected state account-ability is the subject of this section.

Despite the widespreaddemocratization process at thenational and local levels, manybottlenecks remain to the fulloperation of mechanisms of stateaccountability Universal suffrage came late to Latin Amer-ica compared with the United States orEurope, excluding poorer groups for longer,via literacy, wealth, or gender (De Ferranti etal. 2004). But suffrage has now been univer-sal for several decades and certainly so forthe latest democratic transitions. Arguably,LAC is better placed than the other develop-ing regions to take advantage of the so-called long route to accountability—citizens



pressuring politicians, and politicians andpolicy makers pressuring providers—because, on average, democracy is betterestablished in the region. Democracy wouldbe expected to be of particular importancefor service delivery in Latin America in lightof the ineffectiveness of most of the region’sauthoritarian regimes to deliver broad-based services. For a number of reasons,East Asian–style developmental states thatwere effective both at promoting develop-ment and in including most of the popula-tion in the growth process were not anotable feature of Latin American develop-ment. This can be attributed to a range offactors, including the lack of external threatsthat were forces for state-building in EastAsia (and earlier in Europe) and greatersocial cleavages in Latin America that madenational institution-building more difficultand reduced the responsiveness of elite-dominated governments to poorer groups.There were partial exceptions,33 but eventhese cases fall far short of developmentaction by authoritarian states in East Asia,from predemocratic Korea to Vietnam inthe recent period.

The ineffectiveness of the authoritarianregimes has naturally heightened generalexpectations of democratization to bringabout major improvements in inclusivegovernance. However, an overview of therecord suggests that the democracy wave ofthe 1980s and 1990s has not yet ushered in asignificant transformation of service deliv-ery. To be sure (as discussed above), therehas been a marked expansion of social ser-vices in the last two decades that reflects atleast partly the response of governments tocitizen demands. This pattern is confirmedby the growing empirical literature seekingto establish causal relationships betweenpolitical regimes and economic and socialoutcomes (see box 3.9, below). But electoraldemocracy alone has been less effective insituations where what was required toimprove service delivery to the poorinvolved tackling vested interests (for exam-ple, in reforming higher education financ-ing or reforming “truncated” social insur-ance systems),34 reducing state “capture” bypowerful firms, organizations, or individu-als, or resolving distributional conflicts. The

political accountability mechanisms pre-vailing in Latin American democracies donot appear to be strong enough to make theachievement of effective service delivery toall citizens a central goal of governmentactions.

Indeed, there is a seeming consensusamong political scientists that the demo-cratic regimes that have been established inLatin America present serious institutionalbottlenecks, particularly in relation to thedevelopment of adequate and effectivemechanisms of accountability (for example,O’Donnell 1993, 1994, 1996; Shifter 1997;Schedler, Diamond, and Plattner 1999;Smulovitz and Peruzzotti 2000). LatinAmerican democracies have been referred toas “delegative” rather than “representative”(O’Donnell 1994), “thin” (Evans 2002),“illiberal” (Zakaria 1997), and “insufficient”(Méndez 2000), and characterized by “citi-zenship of low intensity” (O’Donnell 2001).

In part, those bottlenecks are genericand not specific to Latin America. Repre-sentative democracy requires the existence


Building on a previous set of studies look-ing at the relationship between democracyand economic growth, there is a growingliterature that analyzes empirically thequestion of whether democracies valuemore social services than authoritarianregimes do. Using data on coverage of keyeducation and health services and democ-racy indicators (for example, the degree ofcompetitiveness of political participation tothe degree of constraints on the chief exec-utive), Lake and Baum (2001) find thatincreases in the democracy level are posi-tively related to levels of public servicesboth in cross-section and time series analy-ses.They also find that the effects of regimechanges (that is, democratization) on thelevel of public service provision were bothrapid and disproportionate. Baum and Lake(2003) find that democracy is associatedwith higher life expectancy and educationlevels, which explain higher economicgrowth rates.

Using data from countries in Africa, Asia,and Latin America, Brown (1999) finds thatdemocracy (measured both as a dichoto-mous and a continuous variable) has astrong influence on primary school enroll-

ment rates.With data for Latin Americancountries in the 1973–97 period, Kaufmanand Segura-Ubiergo (2001) find no effect of“domestic politics” (whether a country isdemocratic or not and whethergovernment’s support base is stronglylinked to unions) on aggregate socialspending levels.They do find, however, thatdemocratic governments spend more onhealth and education, while “popularlyelected” governments protect expenditureson pensions and other transfers that tendto benefit their support base. Using datafrom 17 Latin American countries between1980 and 1992, Brown and Hunter (1999)find that in times of economic crisis, democ-racies tend to protect social spending rela-tive to nondemocracies.

In one of the few country level analysesavailable, Brown (2002) explores educationspending trends in Brazil and links them toelectoral competition patterns. He finds thatincreasing electoral competition in Brazilhas been associated with an increasingshare of the federal budget allocated toeducation and a shift from university to pri-mary education spending.

B O X 3 . 9 Does Democracy Lead to More Social Spending?


of institutional mechanisms that bridge thegap between citizens and their political rep-resentatives—accountability mechanismsthat ensure that public officials are answer-able for their actions. Elections constitutethe primary form of political accountabilityeven though they are recognized as ratherrestrictive and limited, among other rea-sons, given that citizens only get one shot topunish or reward politicians for numerousdecisions and actions (Przeworski et al.1999; Keefer and Khemani 2003). Nationalelections are a rather crude mechanism forcitizens to enforce policy makers’ responsi-bilities for the delivery of specific services.

But there are reasons specific to LatinAmerica that may help explain why the lat-est wave of democratization has not yetdelivered on the promise of radicallyenhanced political accountabilities. Thefunctioning of democracy in Latin Ameri-can countries appears to be deeply influ-enced by the depth and resilience of clien-telistic structures, whereby politiciansgenerally draw support from relatively nar-row, geographically defined constituenciesto which they provide tangible materialgoods (that is, patronage) in exchange forelectoral support. An additional feature ofLatin American political history is a patternof “corporatism,” in which governmentsbrought working class or peasant move-ments into national, institutionalizedpatronage-based systems, as in the case ofMexico with the PRI, Argentina with Peron-ism, and to a certain extent, in Bolivia withthe MNR and Brazil under Getulio Vargas(De Ferranti et al. 2004).

In many cases, clientelism and corpo-ratism had a direct negative impact on pub-lic sector service delivery systems,35 makingdifficult the implementation of reforms instate bureaucracies or the transfer of powerto alternative service providers (private sec-tor firms, NGOs, or communities). At thesame time, clientelism and corporatismhave also worked against the developmentof programmatic (idea-based) parties thatwas an important part of the political evo-lution of most European countries and theUnited States (Shefter 1994). One conse-quence has been that most Latin Americancountries have faced democratic consolida-

tion and market-oriented reforms in the1980s and 1990s with relatively weak insti-tutions for the articulation, aggregation,and negotiation of the interests of poorergroups and inadequate mechanisms for res-olution of distributional conflicts. This mayhave left the new democracies too exposedto capture by all sorts of special interests.

A related characteristic of new LatinAmerican democracies is an observedweakness in systems of checks and balancesthat the executive power is subject to. Thepolitical science literature speaks of weak“horizontal” (O’Donnell 1994, 1998) or“intra-state” accountability (Mainwaring2003), associated with the inadequate roleplayed by legislatures, judiciaries, and statecontrol organisms in most Latin Americancountries. The idea (see figure 3.1) is thatstate accountability takes place not onlythrough elections (“vertical accountabil-ity”), but also through the checks and bal-ances among different organs of the state.

Legislatures in most Latin Americancountries show some worrisome signs ofweak capacities to act as effective counter-balances to strong executives, for example,in terms of budget formulation. Rare are theinstances where Congress tries to hold theexecutive branch accountable for the perfor-mance of particular policy programs and forthe budget. Instead of working through theformal budget process on the basis of theconstitutionally vested power to oversee theexecutive branch, Congress often resorts topassing legislation with serious budgetaryimplications without submitting them torigorous fiscal analysis. This then creates asituation whereby a country effectively hasdual budget processes, one formally gov-erned by the budget law, but informallymanaged with executive discretion, and theother developed through substantive legisla-tion with budgetary impacts that onlybecome apparent ex post.36

Prevailing electoral and party systemsoften do not contribute to creating pressureto change this behavior pattern. In manycountries, the less-than-democratic processof selecting congressional candidates withinparties (particularly when elections are runbased on closed party lists, as practiced inBolivia and Venezuela) leads to weak voter



representation (Shugart et al. 2000). At theother extreme, weak party control of candi-dates (as in the case of Colombia and Brazil,for example) leads to personalistic repre-sentation that, in many cases, contributes to“pork-barrel” congressional politics. As aresult, efforts to modernize legislatures haveshown so far limited impact (Mendez2000). The implication is double. Legisla-tures remain weak to oversee actions by theexecutive branch. They also remain a weakactor in terms of promoting legal reformsoriented to improving incentives for moreeffective service delivery systems.

The judiciary is another key actor in thepresidential system of checks and balancesadopted in Latin America. Still a rarity inthe region is an effective judiciary that bal-ances independence from undue politicalinterference with external accountability forits decisions. Courts, by design, are the leastaccountable to external actors for theiractions, and thus the least proactive amongthe three branches. Their role is to encour-age rule-based behavior by resolving con-crete cases with a consistent application of

the legal norms, but they cannot exercisethis power except through a direct request(by a party with standing). Even when casesare heard and ruled on, their ability to forcecompliance with their decisions tends to belimited, especially when directed againstgovernment itself (Henderson et al. 2004).In the best of situations they are most effec-tive in providing relief to individuals, pun-ishing rule violators, and determining thelegality of actions and the constitutionalityof laws. Since resorting to judicial means iscostly, this “reactive” characteristic of thecourts has tended to benefit those who canafford judicial recourses, that is, the rela-tively well-off. Since the rulings typicallyapply to individual cases one-by-one, evenfavorable rulings for upholding citizenrights to public services, for example, tendnot to benefit the majority poor who havenot participated in the particular courtcases. Courts can hold political leadersaccountable for breaches of the law, but it israre that they can have much effect in forc-ing them to do what they will not or cannotundertake.


Supreme AuditInstitutions

Human Rights/Public Ombudsman

Attorney General/Public Prosecutor

Horizontal Accountabilty

Vertical Accountabilty

Elections

Media

Citizens Civil Society

Executive

Legislative Judiciary

Figure 3.1 Accountability Mechanisms in Presidential Democratic Systems

Source: Payne et al. 2002.


Against this general backdrop, judicialindependence remains the subject ofintense criticism and controversy in theregion (Galindo 2003). More activist courtsthat have taken an atypically aggressive rolein reviewing government policies anddemanding provision of second and thirdgeneration rights to aggrieved citizens havebeen accused of unwarranted meddling inexecutive affairs and even a sort of judicialpopulism (Wilson and Handeberg 1998). Ina few countries where the judiciaries havehigh degrees of institutional independencefrom the executive (for example, Brazil,Colombia), judicial decisions have some-times struck down important policy mea-sures with effects that may have been wel-fare-enhancing for the general population(for example, by curtailing legally sanc-tioned privileges and by imposing unaf-fordable fiscal costs on the government,both current and future). Courts sticking toa more conservative definition of theirfunctions, and often with less political inde-pendence, have been charged with anantipoor bias and an overly formalisticapproach to their work (Buchanan 2001;Garro 1999; Pinheiro 1999; O’Donnell2001). Both types have been accused of cor-ruption, self-interested rulings, and extra-ordinary inefficiency. From the well-docu-mented case of Peru in the late 1990s to theongoing frequent accusations of Venezuelatoday, the judiciaries that lack indepen-dence from the executive branch are seen asthreats to the integrity of democratic insti-tutions and the rule of law. Methods used todeprive the judiciary of institutional inde-pendence vary from outright politicalpurges of judges (for example, Bolivia wentthrough 16 massive judiciary purgesbetween 1936 and 1982), appointment oftemporary judges whom the executive canremove arbitrarily as in Peru and Venezuelaat the time of this writing, or blatant use ofpartisan criteria for appointing and remov-ing judges as in Honduras, where untilrecently, the Supreme Court and the major-ity of the bench was renewed at each gov-ernment change.

In many countries, establishing a mod-icum of judicial independence is a priorityfor assuring a minimum rule of law, though

having independence is hardly sufficient tohave well-functioning courts. Yet despitereform efforts, such as court modernizationprograms, independence of the courtsremains questionable in most Latin Ameri-can countries. Many countries in the regionstill experience limited capacity to provideequal access to justice to all citizens(Buchanan 2001; Garro 1999; Pinheiro1999; O’Donnell 2001). Historical legaciesof arbitrary behavior by the state and theability of powerful elites to operate above oroutside the law contribute to negative per-ceptions of the justice system (Domingoand Sieder 2001). As a result, courts tendnot to play an active role as enforcers ofstate accountability vis-à-vis citizens, par-ticularly poor ones (Correa Sutil 1999;Pasara 2002; Dodson and Jackson 2000).

In the last 20 years, legal and constitu-tional reforms across Latin America havegenerated quasi-autonomous institutionsthat monitor the exercise of public author-ity. The creation of an ombudsman office(“Defensor del Pueblo”) has probably beenone of the most successful efforts to com-pensate for weak judiciaries. In 1985,Guatemala was the first Latin Americancountry to create an ombudsman’s office,and in the 1990s, many countries followedsuit. 37 These offices investigate complaintsover human rights, environmental protec-tion, freedom of the press, and elections.Although institutional design varies fromcountry to country, this office’s decisions orrecommendations are never binding. Inmost countries, the legislative branchappoints the ombudsman, though theombudsman operates as an independententity. But, in most countries, ombudsmanoffices have very limited (legal and other-wise) capacity to enforce their findings(Shugart et al. 2000) and their effectivenessdepends mostly on political dynamics andnot on legal status.38

Audit institutions (such as comptroller’sand inspector general offices) oversee bud-geting and expenditure of public funds.Recent reforms have often involved changesfrom simply publicizing budget figures toevaluating efficiency or cost-effectiveness.Public auditors have also gained increasedinstitutional autonomy. Prosecutorial or



investigative agencies (often the MinisterioPúblico or Procuraduría General) fulfill sev-eral roles, but the most important fordemocratic accountability has been theirincreasing role overseeing prominent gov-ernment corruption investigations. In bothcases, there are indications of weak capaci-ties in these institutions (Payne et al. 2002).

Decentralization has moved issues downto the local level, adding a new layer ofpolitical relations that influences opportu-nities and capacities for citizens to influencepolicy makers. Seemingly dramatic transi-tions to new political and social equilibriahave occurred in some cities or municipali-ties, leading some observers to see morepotential for rapid change at local levels(Campbell 2003). When this happens, bet-ter service delivery systems typicallyresult.39 However, it is still unclear howwidespread the benefits of local leveldemocratization, as exemplified in thesecases, will be throughout the region. As thecase of Bolivia previously discussed (seeabove) indicates, even a sharp nationalreform as Popular Participation may havelimited impact on political influence of thepoor, depending on local context. The over-sight and participatory instruments itintroduced—that is, the vigilance commit-tees and the constructive vote of censure—sometimes became attractive targets forpolitical parties and interest groups, therebyweakening accountability and transparency.In those localities where the traditional par-ties remained strong (and civic associa-tional activities weak), vigilance committeestended to function less effectively eitherbecause the latter were co-opted by the tra-ditional elites or because the municipalgovernments and the vigilance committeesbecame arenas for open conflict. The con-structive vote of censure, instead of beingused as an ultimate means of holding may-ors electorally accountable for their perfor-mance, was often abused as a tool of parti-san conflict.

The “insufficiency of democracy” forservice delivery for the poor is partly areflection of the intrinsic complexity ofshaping institutions for effective services—an issue that is on the agenda even in richsocieties with relatively efficient states. But

this challenge is exacerbated by Latin Amer-ica’s sociopolitical conditions that have lim-ited the scope for effective and equitablepublic actions. Widespread elections andpolitical decentralization are importantradical changes that have taken place inLatin America over the last two decades. Inparticular, they have created opportunitiesfor creativity, innovation, and experimenta-tion in country after country. But patron-age, elite rule, and group-based exclusionappear to be resilient. While elections canbe established overnight, rule of law andrepresentative institutions require changesin political culture and habits that areharder to change. Ultimately, whether andhow the service delivery institutions will betransformed as a result of the transitionsfrom authoritarianism to democracydepends on the strength of mechanisms forvertical (electoral) and horizontal (withinstate) as well as social accountability.

A virtuous circle may be in themaking, whereby socialaccountability mechanisms enablemore effective state actions inservice deliveryThe mixed record of the new democracies intransforming the historical patterns ofunequal and ineffective service delivery sys-tems may be one of the reasons behind thegrowing uneasiness with political institu-tions identified in opinion polls throughoutthe region [United Nations DevelopmentProgramme (PNUD) 2004]. The “discon-tent” with the observed weak state account-ability mechanisms has led, in some cases, toconflict and political instability, which inextreme circumstances may reduce even fur-ther the state’s ability to guarantee mini-mum service standards even in basic ser-vices. At the same time, there is someevidence that the actions of organized citi-zens (for example, monitoring and exposingwrongdoing by the state) can trigger positivechanges in government actions. In otherwords, the tension created by weak stateaccountability does not necessarily meanthere is a zero (or worse, negative) sum bal-ance of power between state and citizens.Rather, when civic actions trigger reactions



of the more conventional vertical, and par-ticularly intrastate, accountability mecha-nisms, a synergic process may take place.

Partly enabled by the democratizationprocess, the last two decades have seen risingactivity by civil society actors outside thestate, from NGOs to identity-based move-ments (notably amongst indigenousgroups). Classical social actors (those exist-ing before the new wave of democratiza-tions) appear to have lost some of theirsocial significance and tended to corpora-tize, while new ones have appeared but are

still rather unstable (Garretón 2002). Forexample, human rights organizations andcivic networks that monitor elections to pre-vent fraud or police actions to prevent abuseand violence have emerged as importantactors in countries such as Argentina, Brazil,Chile, Colombia, Mexico, and Peru (Peruz-zotti and Smulovitz 2001). The developmentof the public interest law movement(McClymont and Golub 2000) is anothergood example of how the emergence of civilsociety organizations—in this case, thosethat demand due process and “proper” pro-cedures—is affecting accountability mecha-nisms. A free(er) and more inquisitive press(Peruzzotti and Smulovitz 2001) and inde-pendent think tanks are additional facts inpostauthoritarian Latin America.

This more active civil society has pro-vided the basis for what has come to beknown as social accountability (Smulovitzand Peruzzotti 2000): the actions by a mul-tiple array of citizens’ associations andmovements and the media to monitor gov-ernment actions and expose wrongdoing.The exercise of this type of accountabilityhas typically followed three complementarystrategies (Peruzzotti and Smulovitz 2001):legal approaches (involving the courts),40

social mobilization, and mediatic—the lat-ter two as a means of calling attention topublic wrongdoing. In all cases, theseactions seek to activate the operation ofintrastate accountability mechanismsdirectly (that is, by decisions of the relevantstate control agencies, including the legalsystem) or indirectly by pressure operatingthrough the political system. In otherwords, civic action in response to state fail-ures appears to be directly oriented toencouraging responses by formal stateaccountability mechanisms.

Of particular interest and relevance forservice delivery systems are a growing num-ber of cases of civil society organizationsinvolved in participatory monitoring andevaluation of service delivery (for example,through the use of report cards as exempli-fied in box 3.10, above), public policies, andbudgets as well as of the transparency in theperformance of public servants—a goodsample of which were surveyed in World


A report card is a survey soliciting citizenfeedback on the performance of public ser-vices. Cards can provide feedback from userson: availability, reliability, and coverage ofservices; user satisfaction with service qual-ity; responsiveness of service providers toproblems; prevalence of hidden costs likecorruption; and user willingness to pay.Report cards enable citizens to signalneeded reforms to public agencies andpoliticians and may exact accountabilitythrough media coverage and civil societyadvocacy of service quality. Although theirimpact has not been rigorously evaluated,they show potential for increasing account-ability and improving service quality.

Report cards were first implemented inBangalore, India, as a response to growingdemand on essential public services—demand that service agencies could noteffectively meet. Beginning in 1993, Banga-lore gathered citizen feedback through acitywide survey.That feedback pushedproviders to improve the quality and avail-ability of public services. It also empoweredcitizens to directly evaluate civil servantsand hence altered the relationship betweencitizens and local government.

LAC Report Cards. Peru was the first LatinAmerican country to implement reportcards to assess national nutrition,education, health, and employmentprograms. Bogotá’s Cómo Vamos initiativesoon followed in Colombia. In 1997, leadersfrom Bogotá’s El Tiempo newspaper, theCorona Foundation, and Chamber of Com-merce decided to found a reporting systemon public services.The logic was simple: ifcitizens paid taxes to support local services,why should citizens not play a direct role in

evaluating those services? The resulting sys-tem, Bogotá Como Vamos, has collected sur-vey data from citizens on the coverage andquality of health, education, security, andother services.The data have been collectedboth from an annual survey of citizens andcity administrators, and from focus groupsincluding citizens from all socioeconomicstrata. El Tiempo and the local city televisionstation disseminate survey results—eachmedium reaches about 3 million people—and Cómo Vamos sends a quarterly bulletinto 3,000 citizen organizations, libraries,research centers, and universities.The proj-ect costs about $75,000 annually—half foroffice and staff costs and the remainder forobtaining and disseminating data.The pro-gram’s main visible effects have been onpublicizing information.

Colombia’s secretary of education nowpresents performance data collected byCómo Vamos on the secretary’s public Website. Cómo Vamos also published a citizens’guide for the 2000 mayoral election with asummary on trends and achievements.

Citizen report cards are being plannedin Argentina, Bolivia, Ecuador, Honduras,and Mexico. A related tool is communityscorecards, which focus on monitoring per-formance of services at the facility or com-munity level—the level of the school, healthpost, treatment plant, or other service facil-ity.The scorecard process includes an inter-face meeting between service providersand community members that allows forimmediate feedback and hence has thepotential to be a strong instrument forempowerment. Scorecards will beimplemented in 12 municipalities inNicaragua, among other locations.

B O X 3 . 1 0 ¿Cómo Vamos? Report Cards, Scorecards, andCitizen Monitoring of Service Quality


Bank (2004).41 This has sometimes beenassociated with efforts by national or localgovernments to open up government deci-sion making to citizen and civic organiza-tions in response to major governabilitycrises (see box 3.11 on Peru). In some coun-tries, regions, and cities, efforts to open upgovernment decisions to citizens and civicorganizations—unheard of until relativelyrecently—are emerging in very differentsettings.

Overall, these various cases reinforce thepoint that the success and sustainability ofexperiences of social accountability are heav-ily dependent on the development of syner-gistic relationships between autonomous

and capable organizations of civil societyand formally established state institutions(see World Bank 2003b). In other words, civilsociety pressures, demands, and actions mat-ter the most when they enable and empowerthe state’s own accountability mechanisms(Fox 2000).

The latter point is well illustrated by theexperience and difficulties encountered bycivil society in its interaction with the regu-latory and control agencies set up by thestate in the case of infrastructure services[see Asociación de Entes Reguladores deAgua Potable y Saneamiento de las Améri-cas (ADERASA) forthcoming; Collado2003]. Under the reforms implemented


The problem. In 2002, 34 percent of Peruvianslisted corruption as the biggest national prob-lem, making it third after unemployment andpoverty. Most Peruvians expressed distrust ofpolitical parties (81 percent), Congress (73 per-cent), and the police (66 percent). Furthermore,Peruvians described receiving poor service frommunicipalities (19 percent), the police (33 per-cent), and judges (42 percent).

Peru’s Response: Building Accountability.In response to the lack of confidence in demo-cratic institutions, Peru instituted a variety ofmechanisms seeking to improve governmentoperations and government accountability.

• National Roundtable for Poverty Allevia-tion: In January 2001, the transitional gov-ernment of President Valentin Paniaguainstitutionalized a roundtable of intellec-tuals, government agencies, local govern-ments, representatives of civil society, reli-gious groups, and internationalrepresentatives to develop a consensusabout social policies as well as to provideoversight to their implementation.

• Local poverty alleviation roundtables. TheNational Roundtable helped organizedistrict- and provincial-level roundtablesof public officials and civil society repre-sentatives to cooperatively design localsocial programs and strategic develop-ment plans. By 2004, over 1,000 roundta-bles were in operation.

• Participatory budgeting: Since theNovember 2002 regional and municipalelections, all subnational governmentshave been required by law to work with

citizens and NGOs to design agencybudgets as part of Peru’s decentraliza-tion process.This process has involvedcitizens in the prioritization and alloca-tion of local investments and the designof local development projects thatmunicipal budgets finance. Regulationsand a manual on participatory budgetformulation were issued, and initiativesto improve quality of participation areunder way.

• The National Agreement: Upon arriving inpower in July 2001, newly elected Presi-dent Alejandro Toledo appealed to allpolitical parties represented in theNational Congress and civil society todesign a “National Agreement”of long-term policies.The agreement, consistingof 30 state policies in a variety of areas,was signed in September 2002; a matrixof indicators was also developed. In May2004, a decision was made to establish amonitoring system to track social policyimplementation.The system involvesboth government and civil society actors.

• Freedom of Public Information Law. In2003, Peru passed a Law on Trans-parency and Access to Information,allowing citizens and public officials toaccess information on governmentoperations at all levels.There is a clearneed to improve the government effi-ciency in responding to an exponentialincrease in demand for information fromcitizens, political parties, and civil societyorganizations.

• Transparent financial Accounting for pub-lic expenditures: Peru required nationaland subnational governments to recordexpenditures through a transparentfinancial management system (SIAF)that regularly compares public spendingto budgeting and physical goals. Forsocial programs, the information hasdetails down to the district level.Thesystem is transparent and open to all cit-izens through the Internet, so that any-one can monitor spending. In 2003, thesite was made user-friendly to encour-age its use by civil society through aportal called “Consulta Amigable.”

• Report cards on social programs: Peru’sNational Household Surveyincorporated questions that captureinformation about users’ satisfactionwith health and food programs. Repre-sentative regional level data gatheredthrough the survey will be available tothe public in 2005.The information willbe used by the national monitoring sys-tem, which tracks the quality of publicexpenditures.

These social accountability mechanismssought to improve the quality of public services,and at the same time, to give legitimacy to ademocratic government by involving citizens inprogram design, implementation, and monitor-ing. However, given that the reforms have onlyrecently occurred, the extent of their effects willnot be clear for some time.

Source: Felicio and John-Abraham 2004.

B O X 3 . 1 1 Opening Up to Social Accountability: Peru, 2001–04


during the 1990s, the regulatory agency isintended to act as an arbitrator between theconflicting interests of consumers and ser-vice providers. However, for this model tofunction effectively, it is necessary for con-sumers to have a spokesperson that canarticulate their interests, and counterbal-ance the industry perspective, which is usu-ally very strongly represented by the serviceproviders. Without this countervailinginfluence, there may be a significant risk ofregulatory capture by industry interests. Inother words, without social accountability,formal “horizontal accountability” mecha-nisms are unlikely to be fully effective, atleast not in a systematic manner.

In Latin America, there was considerablegrowth in the number of consumer organi-zations during the 1990s. However, in manycountries their presence is still incipient,and in general, they suffer from a shortageof funding, which in turn limits their abilityto access the human resources they need toengage in complex regulatory debates. Mostcountries in the region continue to lack thelegal framework that would give regulatorsa clear basis for interacting with civil soci-ety.42 There are very few examples where theregulator has established a permanentchannel of interaction with civil society43

and in those cases in which such interactionis formalized, ensuring the independenceand sustained credibility of both consumerrepresentatives and regulators remains achallenge. Following the U.S. model, publichearings to consult on major regulatorydecisions are already widely used in a num-ber of countries including Argentina, Brazil,Colombia, Costa Rica, and Peru. In coun-tries such as Bolivia and Panama, moreinformal public consultation meetings areheld. However, consumer associations oftencomplain that regulators do not provideaccessible information sufficiently far aheadto allow them to prepare for these events.44

Moreover, public hearings are sometimesperceived as a formality that does not ulti-mately affect the decisions that are taken.Furthermore, there is little evidence that,when held, such hearings may have led toinquiries or actions by, for example, the rel-evant congressional committee—a keyforce for accountability in the U.S. model.

A further concern is the frequent prox-imity between consumer associations andpolitical parties, which may potentiallyundermine their ability to act as indepen-dent representatives of consumer interests.In a number of countries, the creation of anombudsman (“Defensor del Pueblo”), as anindependent publicly funded spokespersonfor consumer interests, has helped to com-pensate for some of these problems. A lackof tradition in consumer activism remains,nonetheless, a bottleneck particularly forpoorer and subordinate groups. For exam-ple, a case study of the Bolivian experiencewith complaint mechanisms after privatiza-tion found that there were at least threetimes more complaints per 1,000 customersoriginating from the relatively higherincome city of La Paz than from its poorertwin city of El Alto, both served by the sameutility companies (Foster and Irusta 2003).

Overall, the evidence reviewed presentsus with a “glass is half-full, glass is half-empty” situation. Clearly, the politicaltransformations of the last two decadeshave positioned countries in Latin Americamuch better to enhance the citizen voice’srole and with it the strength of the “longroute” to greater accountability. At the sametime, a long history of weak states and socialand economic inequalities presents itselfas a heavy burden in most countries. Thereare signs suggesting that stronger socialaccountability mechanisms—like the onesdescribed above—create pressures thatcould trigger the operation of horizontal(intrastate) accountability mechanismsand, more broadly, pressures for reformingthe state. If sustained, these changes couldfurther enable civil society (and particularlythe poor) to actively and productivelyengage in influencing policies and out-comes. As suggested by Fox (2000), this typeof virtuous circle requires rejecting the stillwidely held assumption that state and soci-ety are necessarily engaged in a zero-sumbalance of power. This points us in thedirection of complementary efforts toenhance various accountability mecha-nisms: long- and short-route approachescomplementing and reinforcing each other.The next chapter explores this idea in moredetail.



Notes1. For a succinct historical account of public sec-

tor reforms in Latin America, see Spink (1999).

2. Better financial management systems, however,have proven rather powerless in controllingquestionable expenditure management whenpowerful officials are determined to circum-vent rules of good financial management. Suchwas the case in Peru in the 1990s, when theexecutive branch circumvented its effects oninternal control by resorting to various infor-mal means ranging from highly centralizedcash management to the decision to leave cer-tain “political” expenditures outside SIAF cov-erage’s scope and the use of presidentialdecrees to modify budgets with minimal con-gressional oversight. See World Bank 2001a.

3. The extent of union power depends on both theeconomic and political markets. In areas withless competition—education more than health,old-style public monopolies, and national oilcompanies—workers have more leverage.

4. In the case of the Minas Gerais reform ana-lyzed by Grindle and Mason (2004), there werea number of factors leading to a more positiverole of teacher unions than in most of the othercases she studied. In particular, the union wassupportive of the competitive selection ofschool principals and of introducing broadercommunity participation, as it counteractedprevious abusive employment practices byclientelistic local politicians.

5. Among the reasons cited for the positive rolethe Chilean teacher’s union played is its rela-tively high degree of professional and technicalcompetence, related to its reliance on its pro-fessional reputation for its legitimacy. Com-pared to most teacher unions in the region thattend to concentrate their focus on bread-and-butter issues of remuneration and work condi-tions, the Chilean teacher’s union openly andsystematically engages in serious and consid-ered debates on education policy and is thus arespected voice in Chile’s education policy net-work (Crouch 2004).

6. It should be recognized that conventional“command and control” service deliveryapproaches do work in the case of specific ser-vices, whereby informational problems arelimited and social demands for immediateresults strong. A good example has been polioeradication in LAC (Olive et al. 1997).

7. The only exceptions to this general pattern areChile, which decentralized water to the

regional rather than municipal level, and anumber of Central American countries(Panama and Nicaragua), which still retain sin-gle national providers. In Brazil, the decentral-ization was from the state to the municipallevel and is still only partial.

8. The decentralization of education services, forexample, presents a variety of experiences. Afirst group of countries (Argentina, Brazil,Chile, and Mexico) placed a subnational levelof government at the center of the decentral-ization process. A second group of countries(most notably Bolivia and Colombia) haveoperated, de facto, with a system of sharedresponsibilities between local, regional, andnational governments. More specifically, in thecase of both countries, the management ofhuman resources was attributed primarily tothe regional level and the management ofschool infrastructure primarily to the munici-pal level, with the responsibility for the plan-ning, pedagogical, and curricular aspects of theeducational process shared mostly across thecentral/regional/local level. A third group ofcountries (principally in Central America)have decentralized responsibilities to theschool level, adopting an approach of school-based management with little influence of sub-national governments.

9. We use the term “local government” looselyhere to refer to a broad range of subnationalstate actors.

10. In a number of countries including Argentina,Bolivia, Brazil, Colombia, Ecuador, Peru, andVenezuela, conflictive relations and politicalrivalries between the national and subnationalgovernments have reduced the scope for effec-tive intergovernmental coordination. Theessentially political nature of these relationshas also made it difficult for these countries toresolve the perennial problem of verticalimbalance in resource and responsibilityassignments across government levels, as re-assignment of resources and responsibilitiesoften implied zero-sum games between com-peting political actors. See, for example, Abru-cio (1999) for a thesis on “competitive andpredatory federalism” in Brazil and, similarly,Tommasi et al. (2001) for the case of Argentina.

11. In many cases, only a small minority of munic-ipalities have taken any measures to modernizetheir water utilities, whether by private sectorparticipation or by corporatizing the publicprovider. This has led to situations where thenational regulatory agency applies modernincentive-based regulation to utilities that are



still managed along traditional “clientilistic”lines and have no real incentive to meet regula-tory targets. In extreme cases, there may evenbe jurisdiction conflicts. For example, Peru’snational regulator found that when it autho-rized municipal water companies to increasetheir tariffs towards cost-reflective levels, in 30percent of cases the boards of these companiesvoted to keep their tariffs below the regulatoryprice cap for political reasons.

12. Brazil and Mexico have federal banks (Caixaand Banobras), which to some extent condi-tion access to credit against financial indica-tors. Ecuador is currently developing a mecha-nism that would allocate central governmentcapital transfers across municipalities based ona formula that captures the level of financialand institutional performance of the utilities,taking poverty levels into account. A similarscheme has been proposed for Bolivia.

13. In countries where decentralization has takenplace, there are a number of cases where localmayors have played a pivotal role in water sec-tor reform processes, such as Cartagena inColombia. In countries that have thus farretained a centralized service provider, such asHonduras and the Dominican Republic, therehas been an increasing tendency for individualmunicipalities to opt out in the hope of beingable to provide a better service at the local level.An interesting example is the city of PuertoCortes on the Atlantic Coast of Honduras; fol-lowing a hurricane that interrupted the city’swater supply, the mayor succeeded in negotiat-ing a temporary transfer of operation andmaintenance responsibilities from the nationalwater utility and reached an agreement withthe population that tariffs could be raised aslong as services improved. Thereafter, a fulltransfer of responsibilities to the municipalitywas achieved by means of a congressionaldecree, and the utility was converted into a lim-ited company with mixed municipal and work-force ownership.

14. Comparing the decade leading up to reform tothe decade that followed reform, the expansionrate of service connections increased by a factorof two for water and sewerage services, a factorof four for fixed-line telephony, and a factor of27 for cellular telephony, while it declinedslightly for electricity. A similar comparison forlabor productivity shows that the improvementrate of this index increased by almost threefoldfor electricity, by more than fivefold in thefixed-line telephony sector, and by more thantenfold in the cellular telephony sector; how-

ever, for the water sector, the improvement ismore modest. There have also been significantimprovements in the continuity of service, par-ticularly in the electricity sector.

15. McKenzie and Mookherjee (2002) find that theincrease in access to water following privatesector participation in La Paz and El Alto inBolivia was significantly higher than the previ-ous historical trend for those cities and to thetrend observed in nonprivatized cities over thesame period. Moreover, a recent study fromArgentina showed that (controlling for otherrelevant factors) municipalities that had priva-tized their water services had significantlylower infant mortality rates than those thatretained the service in public hands, the mainexplanation being improved service quality(Galiani et al. 2002).

16. Foster (2002) found that the typical period oftenure of water regulators was two to fouryears, compared with five- to six-year legalterms. This kind of evidence illustrates that it isdifficult to legislate against the prevailing polit-ical culture and that regulators may comeunder pressure to “resign voluntarily” longbefore their legal term has expired, particularlyfollowing changes of administration. Never-theless, while two- to four-year terms may fallshort of expectations, they still represent agreater degree of stability than is the case forsector ministers or public utility managers,who typically have a higher turnover rate.

17. It should be emphasized that responsiveness to“communities” is not always, or not necessarily,the same as responsiveness to the articulatedpreferences of poor groups in the population.“Communities” are heterogeneous and embed-ded in unequal social and political structures.There is undoubtedly substantial variation inthe ways in which different social funds engagewith (different groups in) local communities.

18. While Bolsa Escola in Brazil was probably thefirst in the region, Oportunidades (previouslyPROGRESA) in Mexico has become the mostfamous example (with similar schemes inpoorer countries, including Honduras andNicaragua).

19. The basic alteration of traditional accountabil-ity relationships described above is comple-mented by a series of innovations that CCTprograms have adopted to address many of theshortcomings of traditional social assistanceprograms: addressing both current and futurepoverty, more sophisticated systems for target-ing the poor, providing cash, fostering syner-



gies in human development, and using evalua-tions strategically.

20. This is related to concerns about the geographi-cal selection criteria that most CCT programsapplied, which can leave out poor areas withlimited health and education supply capacity orlimited financial infrastructure. For a discus-sion of some of the targeting concerns in thespecific case of Oportunidades-PROGRESA,see Adato (2000) and Adato, de la Brière, Min-dek, and Quisumbing (2000).

21. To address many of these concerns, CCT pro-grams have set up channels through groups ofmothers’ representatives that act as informa-tion conduits between the centrally adminis-tered program and the local beneficiaries. Thissystem may have led to women’s empowermentand provided a communication channel withlocal communities, but it operates indepen-dently of elected local officials.

22. For a review of the institutional developmentimpact of social funds, see Carvalho et al.(2002).

23. Tendler also argues that there are some struc-tural features of many social funds that makethem particular conducive to patronage—thedelivery of “obras,” the queues, or lists of appli-cants to be chosen from, and the capacity toturn on the money in preelection periods.

24. These providers often provide the only realisticshort-term supply alternative for low-incomecustomers, yet they often operate in a climateof informality. This limits the possibility ofregulating these activities, for example, toensure that safe drinking water is provided. Italso complicates the formation of potentiallyhelpful partnerships between formal utilitiesand independent providers, whereby, forexample, the utility provides a bulk supply ofpotable water that independent providers thendistribute to peri-urban areas.

25. Namely, the existence of soft budget constraintsat the municipal level, restrictions on municipaldecision making, and the lack of autonomy andaccountability of municipal schools.

26. See Di Gropello (2002) for the negative impactof the teacher statute and Sapelli and Vial(2002) for the negative impact of central inter-ventions on competition through the implieddifferences in per-capita budget between thetwo types of schools.

27. There is still significant controversy about boththe sources and long-term impact of this seg-

mentation on the financial protection of, andservice use by, the poor (Massad 1996; Baezaand Cabezas 1998; Bitran et al. 2000; Vega et al.2003). It should be noted that, particularly dur-ing the 1990s, increased targeted public financ-ing for FONASA resulted in significant, pro-gressive anti-poverty effects of public healthexpenditures.

28. Over 20 years, water services have been pro-vided to 1.3 million rural communities, there-by reaching full coverage of the rural concen-trated population.

29. In Honduras, for example, insufficient capacityin the Ministry of Education contributed tosubstantial delays in salary payments toautonomous schools and teachers being moreabsent for training needs (Di Gropello andMarshall 2004).

30. Development of this emphasis was affected (notwithout endogeneity) by many factors—including the rise of democratization through-out the region; the involvement of “communityparticipation” specialists to help improve meth-ods in social funds, but who had a tendency tosee community participation as a good in andof itself; and rising emphasis in the interna-tional development community on notionssuch as social capital, empowerment, and so on.

31. In the case of Nicaragua, decentralization to theschool level was also functional to the process of“de-Sandinization” that the Chamorro admin-istration initiated. See Grindle and Mason(2004) and Kaufman and Nelson 2004.

32. As, for example, in the case of Guatemala withPRONADE, which was put under severe strainwhen there was a change in government.

33. Cuba has had exceptional results in delivery ofquality education and health services, underauthoritarian auspices (World Bank 2003a)—although possibly at extremely high costs inother dimensions of development. Chile underPinochet, while presiding over distributionalworsening along some dimensions, did haveareas of relatively effective public action for thepoor (Sen and Drèze 1989). Both were buildingon previous histories of relatively effectivesocial development.

34. In Brazil, for example, the democratic transi-tion involved as much as anything a failure toresolve distributional conflicts—with the 1988constitution reflecting such a multiplicity ofpressures. Expansion of debt and an unusualcapacity to raise taxes (by Latin American stan-dards) instead supported a mix of measures



that sought to provide something to all groups,including expanded education for the poor.But this was done without tackling the multi-ple areas of opportunity-hoarding or vestedinterest in the society. Highly subsidized col-lege education has traditionally been one ofthese elite preserves in Brazil, and the numbersindicate a sharp rise in per student spending.

35. Recent work that tries to model clientelism’seffects on policy outcomes under differentassumptions all point to the same conclusion:clientelism results in less than optimal levels ofpublic goods production (Keefer 2002; Robin-son and Verdier 2002; Ames 2003).

36. An example of budget management with exec-utive discretion is Peru in the 1990s. Between1994 and 2001, the Executive issued 748urgency decrees, 27 percent of which directlyamended the budget and 41 percent had a cleareffect on the budget or public finances (WorldBank 2001a). An example of substantive legis-lation affecting budget allocation and even fis-cal sustainability was the passage of a teacherstatute in Honduras in the late 1990s; thisgranted generous salary increases and otheremployment benefits without correspondingmeasures to enhance teacher performance.

37. Ombudsman offices were established inArgentina, Bolivia, Brazil, Chile, Colombia,Costa Rica, Ecuador, El Salvador, Guatemala,Honduras, Mexico, Panama, Peru, andVenezuela (Smulovitz and Peruzzotti 2000).

38. In Peru, for example, Alberto Fujimori’s gov-ernment blocked the Peruvian legislature fromacting on the Ombudsman’s recommendationto prosecute actions of the armed forces andintelligence organizations.

39. For example, Porto Alegre in Brazil is a case ofan apparent transition in which the participa-tory budgeting “instrument” played a role. Thedesign of the participatory budgeting process—by which delegates are allocated to the budgetcouncil in proportion to the participation levelof associations in open forums—has stimulatedassociational activity. It has also had the effectof promoting horizontal coalition building, incontrast to the past when neighborhoodsprocessed demands vertically through politicalpatrons with the effect of severely compromis-

ing their associational autonomy and weaken-ing civil society.

40. Both the Brazilian (1988) and Colombian(1991) constitutions introduced legal figures(Acciones de tutela and Acao Direta de Inconsti-tutionalidade) that allow for citizens’ legalactions in response to failures to enforce rights.

41. Some innovations have targeted the typicallyhard-to-reach legislatures (Mendez 2000). Agood example is the case of Poder Ciudadano inArgentina, with the creation of databases withthe voting records of members of Congress,which has been taken up by civil society orga-nizations in Colombia, Dominican Republic,Ecuador, Guatemala, Mexico, and Panama.

42. Relatively little attention was given to con-sumer issues in the original regulatory frame-works. An important reason for this is that,with the exception of Brazil, utility reformpredates the establishment of broader con-sumer defense legislation. In a number ofcountries (such as Argentina, Colombia, CostaRica, and Peru), consumer defense legislationwas introduced a few years after utility reform,obliging regulators to review their currentpractices in this area, as well as providing themwith new legal instruments to facilitate civilsociety participation.

43. An interesting exception is the case of ETOSS,the water regulator in Buenos Aires, which in1999 established a Consumers Commissionthat was made up of one representative fromeach consumer association active in the waterand sanitation sector. ETOSS funds the com-mission and the commission has the opportu-nity to review and comment on all of the majordecisions that the Board of Directors makes. Ina few isolated cases, such as the state regulatorin Rio Grande do Sul in Brazil, the provincialregulator in Cordoba, Argentina, and (for abrief period) the telecom regulator in Peru,consumer associations have been invited tojoin the board of directors of the regulatoryagencies.

44. In response to these criticisms, Costa Rica hasintroduced a consumer advisor in the regulatoryagency staff; the advisor’s role is to help con-sumer groups analyze and understand the infor-mation issued ahead of each public audience.



4

Conclusions

49

Over the last two decades Latin Americancountries have seen some radical transfor-mations in the state’s organization and struc-ture as well as in the relationship between thestate, citizens, and markets. First, the regionhas moved from being a set of countrieswhere democracy was the exception, to onein which it is the rule. Second, LAC govern-ments have become significantly moredecentralized in political, administrative,and fiscal terms. Third, there was a process ofredefining the state’s role that involvedincreased roles of private service providersand a movement toward enhancing the roleof competition in many areas of economicactivity, including the provision of key ser-vices. Fourth, there has been rising activityby civil society actors, both in services deliv-ery and as actors seeking different policies.

These transformations created opportu-nities for change in the accountabilitymechanisms that link policy makers, citi-zens, users of services, and providers. Figure4.1 graphically illustrates some of thesechanges. The accountability relationshipbetween citizens and policy makers hasbeen altered by the generalization ofnational elections (and by the variousmechanisms of intrastate controls impliedby constitutional rule) and the strengthen-ing of civic rights; it has also been altered bythe introduction of local elections (and theincreased role of local governments enabledby administrative decentralization) and thegrowth and increased activity of civil soci-ety organizations acting as watchdogs andintermediaries between citizens and politi-cians. The evidence suggests that expanded

c h a p t e r

Figure 4.1 A Network of Influence and Accountability Mechanisms

LOCALGOVERNMENTS

NATIONALPOLICY MAKERS

INTERMEDIARYORGANIZATIONS

REGULATORYAGENCIES

Advocacywatchdogs

Localelections

Civil societydevelopment

Administrativedecentralization

Nationalelections

Public sectormodernization

Contracting out/privatization

Competition andcommunity-based management

PROVIDERSPublic & PrivateCITIZENS


political voice by citizens was an importantfactor in expanding service coverage undergovernment auspices that took place overthe last two decades, particularly for socialservices.

The accountability relationship betweenpolicy makers and providers has beenaffected both by contracting out servicespreviously delivered by state bureaucraciesand by changes in the way those bureaucra-cies operate, brought about by public sectormodernization efforts. The accountabilityrelationship between providers and usershas been transformed, in many cases, as aresult of the introduction of competitionand community management of services.The evidence reviewed in chapter 3 showsthat, when successfully implemented,changes in accountability relations betweenpolicy makers, providers, and clients dohave an impact on the quantity and qualityof services delivered. However, the evidencealso shows that a particular service deliveryapproach may have very different results

across countries and localities dependingon political and social context.

Using the WDR language (see box 4.1,below), we can say that Latin American coun-tries have “experimented” with both long andshort routes to accountability. Together,democratization and state reform effortshave created more opportunities for the longroute (citizens exercising control overproviders through elected politicians or pol-icy makers) to operate well in practice. Butmany weaknesses remain in the long route,particularly in terms of the citizen–policymaker relationship, which in turn weakenspolicy makers’ incentives to hold governmentand nongovernment providers accountablefor services delivered. Overall, short routemechanisms (clients making providersaccountable for results through the exerciseof choice or through direct management ofservices) have, on balance, been successful inimproving the delivery of many services,albeit with significant variation dependingon the local sociopolitical context. Moreover,


How can the World Bank and others in the inter-national community contribute to improvingservices for the poor? This work has identifiedthree main areas for innovation:

1. What to do? Reform lending and analyticalwork

Mainstream institutional analysis as a keycomponent of the project cycle. Beyondtraditional economic and technicalanalyses, a direct implication of ouranalysis is that “donors” would be well-served by a better understanding of thepolitical economy characterizingaccountability relationships betweenpolicy makers, provider organizations,front-line providers, and citizens. Projectwork would be more relevant and effec-tive if donors were to systematically con-sider the incentives and individual coun-try and sector contexts that affectservice delivery. Seemingly “neutral” or“technical” analysis will often fail to takeaccount of the institutionaldeterminants of success or failure.

More flexible lending instruments and opera-tional practices are a precondition tofocusing on outcomes and institutional

change rather than on inputs.Tradition-ally, donors often set up parallelprocesses to ensure effective outcomeson isolated projects, but coordinatedefforts to strengthen (not weaken) exist-ing service delivery systems are key toengaging clients in results-oriented ser-vice delivery reforms.

2. Tools that can help? Monitoring and evaluation

Promote impact evaluations. Promotingstructured learning from rigorousimpact evaluations of key interventionsis critical to generating knowledge ondevelopment effectiveness—especiallywhere this is integrated intomechanisms for a variety of forms ofdebate on the results, from communitiesto technical seminars to parliaments orcongresses. Getting the basic material tofeed such debates requires more invest-ment in both evaluation design andimplementation, notably because of thepublic good aspects of theseevaluations, which remainundersupplied without subsidization

Establish performance management systemsand key performance indicators linked

to longer-term strategic goals; monitorand evaluate progress toward reachingthose goals and use the information toinform program and policy decisions.

3. How to do it? Engage clients

Strengthen accountability and client “voice”through access to information. Publicaccess to transparent information ongovernment performance and,conversely, supporting government’saccess to public opinion are key tostrengthening the accountability of pol-icy makers and service providers to beneficiaries.

Process-oriented, responsive clientpartnerships. There are no quick andunique fixes to service deliveryproblems.Technical advice alone is nolonger sufficient; instead, donors canincrease their effectiveness by adoptinga partner role based on regular, open,sensitive, and in-depth policy dialoguewith clients.This process-oriented (asopposed to a task-oriented) approach islikely to become more appropriate asinstitutional reform issues become moreimportant in moving forward the reformagenda.

B O X 4 . 1 Lessons for Donors and International Organizations


the weaknesses encountered along the “longroute” may impose often serious bottle-necks even to apparently successful innova-tions empowering service users, particularlypoor ones.

Looking forward, further progress isunlikely to happen if governments keeprelying only on increasing resources allo-cated to the provision of services and moreeffective delivery systems are not developed.This is the case particularly when fiscal con-strains limit the capacity to respond by sim-ply adding more resources and for servicesfor which ensuring quality and relevance isessential. It is in this context that the reportstresses the importance of assessing theextent to which service delivery systems aretoday better positioned to address the manyremaining challenges facing Latin Americancountries, particularly in terms of servingthe poor. In this final chapter, we summa-rize our assessment of how changes in keyaccountability relations that have occurredin Latin America since the early 1980saffected service delivery, and discuss theimplications from this assessment forfuture reform efforts. In some cases thisinvolves entering a more speculative terrain.

Routes to Better ServiceDeliveryOverall, public bureaucracies remain a bot-tleneck for effective service delivery in mostcountries. As shown in chapter 2 of thereport, in most LAC countries governmentshave increased the amount of resourcesallocated to expanding coverage of basicservices. Conventional, centralized publicsector service delivery (involving pure pub-lic financing, delivered through public sec-tor organizations and subject to standardpublic employment practices) has played arole in the substantial expansion in servicecoverage in many sectors and countries. Butthe weak quality of public LAC bureaucra-cies has not shown much improvement.This is despite the significant progress instate modernization reform that upgradedgovernments’ institutional capacities con-siderably, in some areas. It is likely that suchupgrading will have substantial effects overthe longer term in the quality of services

Conclusions 51

that the state delivers. But very few coun-tries in the region have been able to movedecisively toward developing a results-ori-ented public administration or achieve sig-nificant depoliticization of the state appara-tus. Some “islands of efficiency” (such assome social funds) have been created in thepublic sector with consequent positive ser-vice delivery results. But their ability to sus-tain those results is often limited withoutreforms in the operation of “mainstream”state agencies. Decentralization to subna-tional governments has changed structures,but in most cases without radically chang-ing this conclusion—at least not so far.Many local governments have indeedgrabbed the opportunities created bydecentralization and introduced a variety ofinnovative service delivery approaches. Butoverlaps and gaps in responsibilities, inade-quate or insufficient authorities andresources, few (enforceable) mechanisms topromote performance orientation, and thepersistence of clientelistic relationshipsboth at local and central levels (as well asbetween government levels) remain bottle-necks across the region.

As documented in chapter 3, the manyexperiments Latin American countries haveattempted with alternatives to conventionalpublic sector delivery have, overall, pro-duced positive results, though with a widevariation, depending especially on thesociopolitical and institutional context.Contracting out to the private and NGOsectors, promoting competition and choice,and enabling consumer participation in themanagement of services have been tried—with some notable successes—in varioussectors throughout the region. While theempirical evidence is hardly conclusive, it isfair to say that (i) the balance of evidencebetween contracting out and service deliv-ery in-house in the state is mostly in favor ofthe first, under current conditions in mostcountries; (ii) the experience shows thatthere is considerable scope for expandingprivate sector participation in service provi-sion, provided this occurs in a sound regula-tory environment; (iii) giving individualschoice over who they get their services fromhas generated, in the limited areas where ithas been implemented, powerful pressures


for results on front-line providers; and (iv)direct client involvement in the manage-ment of services can yield positive resultswhen the services involved allow a relativelysmall production scale and rather simplemanagement challenges.

There is often a temptation to try to“pick winners” and point our evaluative fin-ger in the direction of one service deliveryapproach—a silver bullet. Specifically, thedifficulties in reforming state institutionsand the relative success in the various alter-native delivery models reviewed above maylead some to argue that the solution to LatinAmerica’s many remaining service deliverygaps resides solely in a scaling up of the“client power” and “contracting out or pri-vatization” experiences. While it is easy tobecome enthusiastic with the results ofindividual positive experiments, the evi-dence is too strong regarding the impor-tance of context to simply argue for scalingup or replicating “best practices.” First, it isoften hard to replicate the specific condi-tions that enabled the initiation and effec-tive implementation of the innovativeefforts reviewed in chapter 3 (be they PRO-GRESA/Oportunidades in Mexico, waterprivatization in La Paz and El Alto inBolivia, or EDUCO schools in El Salvador).How such reforms will be shaped andimplemented will depend crucially on thesocial, political, and public sector context atnational and local levels. Put another way,the reasons for success often lie in the par-ticular ways that groups interact in eachsociety, rather than the specifics of technicaldesign, however important that may be.Second, in the long run most such innova-tions would, in the end, be ineffectual with-out reforms in the operation of “main-stream” state agencies that are more deeplybeholden to the fundamentals of interac-tions between citizens, policy makers, andservice delivery agencies (public or private).In situations where the state’s capacity toestablish and enforce regulations, to priori-tize and coordinate policy actions acrossministries, sectors, and government levels,and to establish and sustain compensatorymechanisms is weak, there are dangers tothese routes—a lack of sustainability, newforms of capture, and social stratification,

among others. In the end, it is unrealisticand unlikely that LAC countries can getaway without both stronger state account-ability and complementary action to createeffective core public bureaucracies, evenwhen deciding to move decisively in thedirection of delegating major responsibili-ties for delivery to nonstate actors.

The political transformations of the lasttwo decades have positioned Latin Ameri-can countries much better to enhance stateaccountability. At the same time, a long his-tory of weak states and social and economicinequalities presents itself as a heavy burdenin most countries. While elections can beestablished overnight, rule of law, an effec-tive system of checks and balances, and rep-resentative institutions require changes inpolitical culture and habits that are harder tochange. This will no doubt require more cit-izen influence on government decisions andactions, including from poor and excludedcitizens. Stronger voice (as a preconditionfor more accountable politicians) is needednot just to reform the state bureaucracy, butalso to ensure that contracting out, choice,and participation do benefit the poor. Thisdoes not mean that until and unless stateaccountability is fully ensured—that is, untilLatin American countries become fullymature and consolidated democracies—efforts to reform service delivery are boundto fail. On the contrary, there are many waysin which short- and long-term goals can bemade complementary. To give one example,community participation approaches to ser-vice delivery (as in community-run schoolsin El Salvador) cannot only have goodshort-term results but, depending on howthey are structured, can also help build indi-vidual, community, and state capacities overthe longer run.

In the last analysis, using the WDR ter-minology, the short route (clients exertingdirect control over providers) should beconceived as a complement rather than asubstitute for the long route to accountabil-ity (citizens exerting control over providersthrough politicians), a potential shortcutthat will only be consolidated if linked tomore fundamental changes. Moreover, ourreading of the evidence makes us skepticalof an approach whereby reformers put all



their eggs in one transformational basket.Rather, we are inclined to support thenotion of “strategic incrementalism”(World Bank 2003a): continued efforts todeepen societywide conditions for account-ability (deepen democracy, citizen partici-pation, and state reform), while at the sametime pursuing opportunities for specificservice delivery reforms and program inno-vations that push for choice, client involve-ment in decision making, and perfor-mance-based contracts both in and outsidethe public sector. In other words, pragmaticincremental reforms while unlikely to fullyaddress service delivery problems, can con-tribute both to short-term results and tocreating conditions for deeper and morefavorable change over time. In this context,complementarities will arise to the extentthat such innovations are designed in waysthat contribute to strengthening both stateand individuals’ capacities as citizens andconsumers. Seen from this perspective, theexperiences reviewed in this report providevaluable lessons for the design of alternativeservice delivery approaches.

Public sector management reforms:seeking stronger links with concreteand observable service deliveryimprovementsThe evidence suggests that reforming statebureaucracies and developing a perfor-mance orientation are naturally difficultendeavors, from which quick results shouldnot be expected. The primary reasons forstate reform failures are political and less sotechnical. Addressing the problems associ-ated with weak traditions of bureaucraticautonomy and the highly politicized man-agement of government offices involvestouching on vested interests. But when thepolitical system does not force accountabil-ity mechanisms to function effectively, tech-nocratic approaches to state reforms areunlikely to produce lasting results. It is hardto aspire to a Weberian bureaucracy incharge of service delivery when stateaccountability mechanisms are not fullydeveloped and functional. This represents amajor challenge for Latin America’s youngdemocracies that still experience too many

barriers in the citizen and policy makerrelationship. As a result, in most countriesthe case is weak for ambitious new publicmanagement–type reform measures beingthe solution to poor service delivery sys-tems over the short to medium term.

This assessment, however, should not beconstrued as an argument against publicadministration modernization efforts.Rather, the implication is that Latin Ameri-can countries should indeed continueinvesting in general state capacities, but withstronger links to specific in-sector reformsand linked to service delivery results. Specif-ically, we hypothesize that the likelihoodthat investments in improving the quality ofadministrative and financial systems willhave significant and/or lasting impact isgreater if and when they are linked toreforms in specific service delivery systems.Indeed, management reforms tied to betterservice delivery outcomes can create newconstituencies and the consequent politicalsupport without which their sustainabilitywould be questioned. This hypothesis is cor-roborated by the relative success of adminis-trative reforms tied to substantive policyreform agendas. Examples include financialmanagement reforms, such as the introduc-tion of computerized financial managementinformation systems as a potent instrumentfor expenditure control and thus broaderfiscal adjustment efforts in the region, or theestablishment of regulatory agencies as partof a broader privatization policy in severalcountries in the region. Similar examplesrelated to service delivery as such are morelimited, but the same logic would apply.

In some circumstances, the strengthen-ing of an “enclave” in the administrationmay be required to achieve a specific servicedelivery goal. For example, this could be thecase when political or social conditionsrequire quick results on the ground thatmainstream agencies would be hard pressedto achieve. Experience, though, suggeststhat the creation of islands of efficiency mayresult in the perpetuation of parallelbureaucracies with consequent efficiencylosses. The response cannot be simplyrejecting altogether the approach. Instead,what is required is facing the dilemmabetween short-run results and long-term

Conclusions 53


institutional development, and designingupfront clear guidelines and strong incen-tives for transitions from islands of effi-ciency to the mainstreaming of administra-tive reforms.

Service delivery decentralization:creating better conditions forperformance-orientation at allgovernment levelsDecentralization is no panacea. It can be assusceptible to (local) elite capture and cor-ruption as national service delivery. But it is apolitical fact of life, and for many services, ithas the potential for both more effectivelyresponding to local demands and fosteringlocal structures of more equal and effectiveaccountability. Moreover, under the rightconditions, decentralization can provide acontext for innovation and experimentation.At the core of successful decentralization is acombination of an improved compactbetween government levels and strengthenedlocal accountability to all citizens. Dependingon the country and sector, this may requiresimpler and more transparent use of inter-governmental transfers, the enforcement ofhard budget constraints for local govern-ments, clear definitions of responsibilities(including through legal and/or administra-tive instruments), strengthening the capacityof central government monitoring andauditing, and developing information andevaluation systems. It may also require spe-cific intervention instruments (by nationalauthorities or federal bodies) when servicedelivery failures threaten the well-being ofcitizens. Last but not least, it requires atten-tion to the other side of empowerment: thecapacity of poorer groups to participate inlocal decision making.

Characteristics of different services makethem more or less prone to local delivery,particularly depending on the presence ofeconomies of scale. In fact, it is typicallyspecific aspects or functions in broad ser-vices that involve such economies of scale.To give just one example, while day-to-dayoperations of schools can be managed in avery decentralized fashion, aspects such ascurriculum design or student learningassessment systems are more effectively

handled in a more centralized manner. Inmany countries, there is great heterogeneityin the size of subnational entities (for exam-ple, municipal governments ruling overlarge cities and very small localities withdispersed populations), which are in a verydifferent position to deliver the variousaspects of particular services. Consideringthe limited experience with decentraliza-tion programs that follow asymmetric rulesamong subnational entities with the samelegal status, addressing these heterogeneitiesmay require the active monitoring andengagement of central authorities.

The latter comment highlights anotherimportant lesson: making decentralizationwork is not a question of getting the blue-print right and then letting things happen.A more proactive role of central govern-ments and NGOs to improve informationflows between subnational entities andamong citizens could contribute tostrengthening the agency of the poor at thelocal level—a key factor for decentralizationto materialize in welfare improvements forthe poor. Central governments (and a rangeof nongovernmental actors) can alsoenhance the benefits from decentralizedinnovation by facilitating cross-jurisdic-tional learning and exchange of experi-ences. At the same time, central govern-ments may often need to be proactive tohelp changing behaviors amongst localactors, for example, through the use offinancial and other incentives, such as con-ditional matching grants, or even directengagement in recruitment, in sharinginformation, and in support for mobiliza-tion. In the end, however, more time may beneeded for decentralization’s benefits tomaterialize in a more generalized fashion.

Delivery by nonstate providers:ensuring adequate regulatory andcompensatory mechanisms are inplace to increase service deliveryeffectiveness and equityThere is solid evidence showing the positiveimpact of reforms oriented to increasingthe participation of nonstate providers (pri-vate for-profit and not-for-profit organiza-tions) in service delivery, when the condi-



tions are right. Privatization and contract-ing out reforms can improve efficiency andquality of service delivery, particularlywhen clients’ choice is enhanced. At thesame time, privatization has sometimes ledto increased concentrations of market (andperhaps political) power in conglomeratesand has sometimes been associated withcorruption. Not all forms of concessionshave proved to be sustainable. Some coun-tries have experienced difficulties in manag-ing contracts over time. Increased socialstratification may be an undesirable out-come of increased competition and choicewhen informational asymmetries are signif-icant and differences large in purchasingpower among consumers. The answer is notto throw out the baby with the bath water,as suggested in some critical analyses of pri-vate sector participation in service delivery,but rather to reap the lessons of experienceto improve the way that such arrangementsare designed.

First, it is essential to ensure that thebenefits of private sector participation areequitably distributed between all the keystakeholders, namely, current customers,potential customers, employees, investors,and the state. This is likely to require a com-bination of regulations and subsidies, par-ticularly to ensure affordability among low-income households. Second, greater effortsmust be made to improve the efficacy ofregulatory agencies that are key to provid-ing a stable operating environment for con-tractors and enforcing contract provisions(including price and quality of services) tobenefit all consumers. Third, there needs tobe greater transparency throughout theentire process, from the transaction design,to the allocation of any resulting proceeds,to the subsequent regulatory decisions thataffect the evolution of the contract. Thiswill involve providing the public with betterand more accessible information, establish-ing formal channels for citizen participa-tion, and supporting the capacity of citizen-based groups to interpret the policies andpractices of privatized firms that oftenrequire technical analysis. Stronger voice bycitizens benefiting from choice and compe-tition may be an important factor in pro-tecting such reforms from interest groups

seeking to reverse them (as occurred in theColombia education voucher case) andensuring their sustainability.

There are likely to be important differ-ences between services in the extent (andnature) of potentially useful public interven-tions. The more that a service is like a classiccommodity, the more likely are choice andcompetition models to work. This is mosttrue of cellular telephony, but becomes ofrising difficulty in areas where market poweris intrinsic (for example, local water), thereare major informational asymmetries(health, perhaps education), and significantprobabilities that social, income, andprovider structures will lead to greater strati-fication (health and education). In theseareas choice can play a role, but it is nopanacea: it is more likely to be a complementthan substitute for state effectiveness—whether in public sector service delivery oreffective regulation—and for stronger citizenvoice, whether influencing policies or effec-tively representing consumer interests.

Community-based service deliveryapproaches: seeking designs that strengthen capacities oforganizations of the poor whileensuring accountability for resultsOverall, when market mechanisms are hardto develop—for example, due to a lack ofcompetition—community-run services canbe a powerful approach to generate results(even in the short run) for services that donot necessarily involve large scales (andthus both requiring sophisticated technicalskills and involving significant collectiveaction challenges) and for which there arenot substantial asymmetries of informationbetween clients and providers. But theirlong-term impact is likely to be the greatestwhen they contribute to changing the localbalance of power in favor of the poor anddisadvantaged—simply giving more influ-ence or resources to local communitiesmay, in itself, risk just pushing problems ofelite capture, inefficiency, or patronage to amore local level. In the end, the key iswhether and how they help empower thepoor through both creating capacities forchoice, aspiration, and organization (that is,

Conclusions 55


building agency) among the poor as well aschanging the opportunity structure wherethey live to increase the probability of theirvoice and actions making a difference. Howthis is done will depend on social and polit-ical context. In particular, to the extent thatthe experience of running services leads tothe strengthening of associational auton-omy among poor communities, even a nar-row exercise of client power can contributeto an enhanced ability of citizens to holdgovernments (local and central) account-able for a broader set of actions and poli-cies. At the same time, the government’srole in the promotion of community-basedapproaches may need to be more proactivethan simply “transferring resources” togroups of organized citizens. A solid “com-pact” between the authorizing state author-ity and the empowered community groupswill typically be required to ensure the ful-fillment of coverage, efficiency, and equitytargets, as illustrated in the case of EDUCO,for example. This brings up again the com-plementarities between state modernizationand client power approaches.

Social accountability: seekingopportunities for synergisticrelations to strengthen the capacityand accountability of stateorganizationsEnhancing the accountability of policymakers to citizens requires more than regu-lar and clean elections. It is through the vir-tuous interaction of electoral (national andlocal), intrastate (legislatures, judiciaries,and control agencies), and social account-ability mechanisms that Latin Americancountries are most likely to reverse theinherited pattern of clientelism, elitism, andcorruption characteristic of most of itsstates. In particular, given the difficult pathto the strengthening of rule of law in coun-tries with such a long history of weak publicinstitutions, stronger social accountabilitymechanisms may help trigger a virtuouscircle by creating the pressures needed tomobilize intrastate checks and balancesand, eventually, create the opportunities forstronger and more inclusive political repre-sentation mechanisms. In other words,

effective social accountability is not aboutbypassing the formal channels of stateaccountability (elections, parliamentaryactivity, independent courts, and so on)but, rather, about supporting them.

In that sense, investments in socialaccountability mechanisms at the nationaland local levels, focused on both overallstate performance and specific service deliv-ery issues, can potentially have an extremelyhigh payoff. It is in this context that thegrowing number of cases of civil societyorganizations involved in participatorymonitoring and evaluation of service deliv-ery (for example, through the use of reportcards), public policies, and budgets as wellas of the transparency in the performanceof public servants constitute a powerfulinstrument for change.

In a region characterized by a long tradi-tion of exclusionary public policies, theactual enforcement and application of rightsappears as a key challenge for improvingstate accountability, particularly vis-à-vis thepoor. The expansion of political rights asso-ciated with the new wave of democratizationoffers a unique opportunity to expand civiland social rights—particularly of the poor,the excluded, the discriminated against—anessential step to counteract the current pat-tern of policies against poverty and inequal-ity being captured and distorted byingrained clientelism and paternalism prac-tices. But the active and explicit enforcementof rights could be an even more directaccountability instrument when it is linkedto specific services. An example is the case ofthe recent efforts to establish guaranteedhealth insurance benefits in several LatinAmerican countries. Although all countriesinclude the right of citizens to good healthand/or access to health services as a constitu-tional mandate, until the mid-1990s, effec-tive instruments for the poor to demandcompliance by the state or other actors didnot accompany such mandates. First Colom-bia in 1994, and since then Bolivia (1997),Chile (2001), Mexico (2002), and Argentina(2003), have introduced legislation and sec-tor reforms to guarantee a package of ser-vices to all citizens. Such reforms also includelegal mechanisms for beneficiaries to com-plain and demand state compliance. Of



course, whether this approach is enforceabledepends crucially on how fiscally and opera-tionally reasonable the mandates really are.

The Centrality of the “FittingProcess”Our review of the Latin American experi-ence supports the view that there cannot bea “one-size-fits-all approach” to servicedelivery: differences in types of services,characteristics of the population, and thenature of the political system make anyattempt to define the perfect “size” futile(World Bank 2003a).

Seeking the “size that fits” is at the core ofany service delivery reform effort. Character-istics of different services—such as the degreeof discretion in decision making and theintensity of transactions involved in thedelivery process, the existence of economiesof scale, and so on—make them better orworse suited for alternative delivery models(Pritchett and Woolcock 2002; World Bank2003a). For example, as discussed above, it ismuch easier to create the conditions for awell-functioning competitive market forsome services than for others (for example,cellular telephones compared to rural watersystems). Similarly, some services (for exam-ple, vaccination) are much better suited formassive, standardized approaches than oth-ers that require much more customization(for example, the pedagogical aspects ofschooling). An important aspect of the “fit-ting” involves disentangling the nature of dif-ferent aspects or components of each service(for example, how easy it is to monitorresults) to determine the most appropriateapproach.

But service characteristics are just oneelement in the equation. Probably moreimportant is the country context. Ofcourse, countries differ along many differ-ent dimensions and any systematization ofsuch differences runs the risk of oversimpli-fying complex realities. Nevertheless, basedon the review of experiences presented inchapter 3, it is possible to identify four keydimensions that help characterize a countrycontext and provide a useful orientation tothe range of feasible reform options thatpolicy makers face in seeking to alter the

way specific services are being delivered toimprove performance.

1. The first dimension is the state’s overallstrength and formality. We have arguedthat there is a general tendency acrossthe region for states to be weak in thesense of doing a poor job delivering keypublic goods and quality basic economicand social services. Naturally, there aredifferences between countries. Stateweakness is manifested in various waysbut, for service delivery, the nature of thestate bureaucracy (ranging from “dys-functional” to “Weberian”) and thedegree to which the rule of law is estab-lished (ranging from informality to dif-ferent degrees of formality) appears tobe critical (see, for example, Burki andPerry 1998). Chile, for example, is acountry with a relatively well-function-ing bureaucracy; significant parts ofBrazilian’s bureaucracy work effectivelydespite high levels of clientelism in parts.Poorer countries such as Bolivia orGuatemala have much weaker and lessrule-bound bureaucracies. The furtheraway from having a “Weberian” bureau-cracy a country (or a sector in a country)is, the less likely sophisticated publicadministration reforms seeking to estab-lish performance orientation in the stateare to succeed. Similarly, in countrieswhere the rule of law is not well estab-lished, the enforcement of contractingout schemes is likely to be weak, and thussuch approaches will be more vulnerableto abuse.

2. The second dimension is the prevailingtype of political representation. We haveargued that Latin American politics havehistorically been characterized by clien-telism and corporatism, which have adirect negative impact on public sectorservice delivery systems and make it dif-ficult to develop alternative deliverymethods that may fit better changingrealities and needs on the ground. Onceagain, there are differences among coun-tries in the extent to which patronageand authoritarianism continue being thepattern by which politicians and citizensrelate to each other, as well as in the

Conclusions 57


degree to which programmatic (idea-based) parties have replaced corporatistor populist movements. Efforts to reformintragovernmental compacts (for exam-ple, by establishing performance-basedreward systems in the state) are likely tobe captured by insiders in the presence ofstrong corporatist movements. Similarly,in highly clientelistic environments,schemes involving community-managedservices may be highly vulnerable tomanipulation and capture by localpatrons. Many countries are in transi-tion, especially at local levels: Colombiais traditionally highly clientelistic, but hasexperienced major changes in Bogotá, forexample.

3. The degree of social and economicinequality is the third dimension. It is wellknown that Latin American countriesexperience high levels of income inequal-ity relative to other parts of the world. It isalso increasingly recognized that severalcountries in the region experience sharp,group-based social inequalities (particu-larly linked to race and ethnicity). Highlevels of social and economic inequalitytypically are reflected in large gaps inaccess to quality services and thusincrease the cost of meeting the goal ofensuring minimum service levels for all.Social inequality and stratification arealso likely to weaken the government’sability to generate necessary consensusfor service reforms that often involvestradeoffs in the short run. Managinginequality is particularly challengingwhere it is associated with group-baseddifferences and social exclusion—as withindigenous groups in the Andes and partsof Mexico and Central America, and Afrogroups in Brazil, Colombia, and Vene-zuela, and many in the Caribbean.Besides, societies stratified along ethnicand racial lines tend to be vulnerable topopulist appeals of politicians who appealto specific societal groups by providingthem with selective private goods ratherthan public goods that benefit society as awhole (Keefer and Khemani 2003). In acontext of high social and economicinequality, an emphasis on choice andcompetition may yield further stratifica-

tion unless the state is capable of enforc-ing strict quality regulations and/or sub-sidizing access to quality services by tradi-tionally excluded groups.

The previous three dimensions arestrongly interlinked in what has beencharacterized as the syndrome of “weakand unequal institutions” (De Ferranti etal. 2004), which affects a majority ofLAC countries. These “structural” fac-tors play out in various ways dependingon historical trends.

4. Thus, a fourth dimension is the point ofdeparture regarding how services arebeing delivered. The history of how spe-cific services have been delivered mattersto the extent that it creates constituenciesfor the status quo as well as ideologicalreference points for society’s conceptionof what are appropriate roles for differentactors and thus influences the costs of atransition to a different approach. In par-ticular, the historical pattern of stateinvolvement in the delivery of specificservices differs between countries (forexample, countries with a strong tradi-tion of publicly provided basic educationversus others where private schools havehistorically represented a large share of the system) and with it the strengthand attitudes of key stakeholders (includ-ing, of course, citizens as consumers of those services) to alternative reformapproaches. Overall, path dependenceplays a big role in defining feasible servicedelivery reforms: forces for inertia tendto be high.

Of course, it is the combination of these(and other) factors that together definecountry context and help determine whichtype of service delivery approach is mostlikely to succeed in improving results. Asan illustration of how context matters, intable 4.1 we provide a simple set of criteriaderived from our assessment of the LatinAmerican experience. These are meant asillustrations of how to analyze countrycontext in assessing prospects for specificservice delivery reform approaches.

In a country where state bureaucraciesoperate broadly along Weberian lines andthe rule of law is well established (most



likely, countries where patronage has weak-ened significantly as the primary form ofpolitical relationship), the choice betweendeepening reforms of public bureaucra-cies—seeking to make them more efficientand performance-oriented—and expand-ing alternative service delivery modelsinvolving nonstate actors will be highlydependent on the departure point. In sec-tors where there is a strong tradition of statepresence, the costs of radically shiftingapproaches may be too high. In those casesa cost-benefit analysis is likely to support astrong push for results-based management(as is currently being tried in the case ofChile). This may need to be complementedby actions to open up the state to feedbackand influence from clients or citizens, if nosuch tradition exists. For those sectors inwhich there is a tradition of private (for- ornot-for-profit) participation in servicedelivery, a key factor to consider in selectingalternative delivery mechanisms is thedegree of social and economic inequality.Choice and competition are likely to oper-ate better either under conditions of lowinequality or where there is the potential fordeveloping means to ensure that poorergroups are not excluded from the wideningpossibilities. Under conditions of highinequality, stronger state involvement isrequired and, thus, contracting out usingstrong compensatory schemes may be a bet-ter alternative (for example, targeting pub-licly subsidized concessions to low-incomeareas as in the Bogotá school concessionsprogram).

In the other extreme, in countries withdysfunctional state bureaucracies andwhere rule of law is limited, the set ofchoices available to policy makers is muchmore limited. In such settings, reformingpublic sector delivery systems is a long-term endeavor that will face major techni-cal and political constraints. Ambitiousefforts to modernize administrative sys-tems are likely to go wasted in the face ofweak civil service, and attempts to reformpersonnel practices are likely to generatestrong resistance from those benefitingfrom the prevailing use of government jobsas instruments of political patronage. Thus,in these settings, administrative reforms arelikely to be a long-term affair that will notbring major short-run service delivery ben-efits. It will often be necessary to comple-ment such reforms with parallel efforts toenhance service delivery by nonstate actors.Limited rule of law is likely to be a seriousconstraint to the enforcement of contracts,unless contractors have very strong intrin-sic motivations to deliver quality services toclients (as in the case of some faith-basedNGOs, such as the noted example of Fe yAlegría schools). The choice between anapproach that builds on community-man-aged services and one based on competi-tion and choice as alternatives to publicsector provision depends on social andpolitical context. In these countries, theability of governments is likely to be low todesign and implement compensatory mea-sures to address informational asymme-tries and differences in purchasing power.

Conclusions 59

Table 4.1 Fitting Approaches to Country Conditions

State characteristics

Poin

t of d

epar

ture

Strong tradition of state presence in service delivery

Tradition of diversified servicedelivery

Weberian bureaucracy andacceptable rule of law

Fruitful to try results-basedmanagement. Need forcomplementary actions to promotesocial accountability (anddecentralization) in cases ofauthoritarian tradition and weakcitizen participation.

Low inequality: competition andchoice.

High inequality: contracting out withstrong compensatory design.

Acceptable rule of law but weakbureaucracy

Small-scale experimentation withresults-based management (forexample, an enclave) within overalllong-term efforts at increasingpressures (demand) for reform at alarger scale. Decentralization is anoption for large countries.

Contracting out a good option aslong as opposition from vestedinterests (political patrons andcorporative interests) allows it.

Weak bureaucracy and limited ruleof law

Very modest administrative reformslinked directly to client powerinitiatives (see below).Decentralization problematic asnational standards will be hard toenforce.

Client power may be the only optionin the short run. If inequality is high,community management may be less costly than choice/competitionunless clientelism is pervasive.


Thus, under conditions of high inequality,community management schemes may bea better alternative (as, for example, in thecase of EDUCO in El Salvador). On theother hand, when local level clientelism ispervasive, the danger that such schemesend up being captured by local patrons issignificant, and thus competition amongprivate providers may be the best option.Where the rule of law is weak, there are alsogreater risks of privatization being cap-tured or corrupted—although the problemis that it is often precisely in these cases thatpublic delivery is also weakest and mostcaptured.

Naturally, there is a whole range of casesin between these two extremes. In table 4.1we have identified a particular configura-tion characterized by an acceptable rule oflaw but weak bureaucracy—a configurationthat may approximate the case of a largenumber of Latin American countries wherestrong traditions of clientelism and/orstrong corporatist movements still have alarge negative effect on the quality of statebureaucracies, even though rule of law hasbeen strengthened over the last twodecades. Unlike the case of countries with“Weberian” bureaucracies, the opportuni-ties are much reduced for large-scalereforms seeking to improve service deliveryby the state through the introduction ofperformance-based management, whichsuggests that smaller scale experimentationmay be more appropriate. It is in these casesthat the tension between short-term fixesand longer-term reforms may be thestrongest. Unlike the case of countrieswhere the rule of law is not yet established,there are likely to be opportunities to create“enclaves” or “islands of efficiency” thatbypass the overall weak state bureaucracies(as illustrated by the experiences of varioussocial funds). At the same time, as discussedabove, unless those experiments have somebuilt-in mechanisms through which experi-ments are, somehow, mainstreamed, theshort-term impacts of innovations may belost over time. The fact that the rule of lawis more established suggests that there maybe room to experiment with decentraliza-tion of service delivery responsibilities—possibly another way of bypassing some

hard-to-reform state bureaucracies. Therule of law is also a supportive factor forcontracting out service provision with non-state actors. This may be the most appropri-ate approach as long as opposition fromvested interests is not strong enough toblock it. This suggests that contracting outis likely to work better in countries in whichpatronage, if not completely eliminated, isin the process of being weakened.

In practice, though, the policy makerchoice is constrained and conditioned by amultiplicity of factors, which we would finddifficult to predict exante. Finding the “rightapproach” is both a political and a techno-cratic process. This is true both at the macroor national level, where there is growing evi-dence of the political complexity of servicedelivery reforms (Grindle and Mason 2004;Kaufman and Nelson 2004), and at themicro delivery point, where specific designchoices interact with provider and clientbehaviors in often-complex ways. As Robin-son (2003) has argued, the issue is much lessthe specific design choices or the new magicbullet (that quickly loses its shine) than theprocesses that generate policy adjustmentsto changing outcomes in response to soci-etal and political pressures. This may takeplace at the national level or at local levels—under the conditions of substantive politicaland administrative decentralization thatexist in Latin America. Paraphrasing Fox(2000, 2001), we can say “the challenge canbe understood as a kind of political FengShui—the art of placing things in balancedrelationship to one another.”

That is why the process by which servicedelivery reforms are defined, designed, andimplemented is so crucially important. Inother words, identifying the “size that fits”country and sector circumstances is likelyto require the participation and involve-ment of different voices in society. AsPritchett and Woolcock (2002) suggest,unlike some macroeconomic policies, ser-vice delivery reforms can hardly be man-aged by “ten smart people.” This may verylikely involve more open and deliberativeforms of governance than what is presentlythe case in most Latin American countries,with the consequent requirements in termsof institutions for consensus building and



conflict management. Of course, strategicincrementalism applies also in this area:there is scope for many partial changes andpayoffs in different parts of the accountabil-ity structure that will often involve specific,partial forms of opening and deliberation,and that will support the broader shift ininstitutions.

But to be effective and avoid becoming asynonym for “chitchatting” debate, consen-sus building and deliberation need to be fedby rigorous analysis based on solid empir-ics. This reinforces the importance ofrobust and systematic evaluations, few ofwhich are readily available, that feed thepolicy-making process in a timely fashion.In other words, strengthening the linksbetween policy evaluation, debate, design,and implementation is fundamental toachieve a better process. That is, a learningfrom experience approach is needed. In thisapproach, policy design is the outcome of adialogue among interested parties, andimplementation is organized in ways thatenable the elaboration of regular and solidpolicy evaluations, which frame and sup-port further dialogue and debate.

The practice of “informed dialogue”(Reimers and McGinn 1997), wherebyresearch-based knowledge is used to shapepolicy, requires a strong sensitivity toprocess and the engagement of all relevantstakeholders. Implementing such anapproach would undoubtedly involve capa-bilities on the part of governmental andnongovernmental actors that, in most cases,do not yet exist. Fox (2001), for example,identifies a key role that civil society organi-zations can play in their advocacy work, byproducing policy and program evaluations.He emphasizes both the capacity and orga-nizational challenges of such an initiativeand the large expected payoffs that wouldresult from it. To the extent that such invest-ments do materialize, the synergistic civilsociety or state relations discussed earliercould well develop in this area as well. Thus,a key way to support service deliveryreforms is to strengthen the capacities of allrelevant actors to participate effectively inpolicy evaluation and dialogue.

Furthermore, particularly given theincreasing range of state and nongovern-

mental actors involved in service delivery,one of the biggest challenges ahead ofLatin American countries is to find simpleways by which the many advantages of“learning by doing” can be leveragedthrough mechanisms that facilitate andencourage “learning from others,” withoutthe artificiality implicit in the traditionalcalls for “replication.”

Where would the leadership for such anapproach come from? In the very graphicterms used in a recent review of the litera-ture on state accountability in Latin Amer-ica, “somebody has to kick the status quofrom its point of equilibrium” (Schedler1999, 347). Who? Improving state account-ability will most likely require reform fromabove (politicians), below (citizens), inside

Conclusions 61

A daunting challenge facing most countriesin the region is to develop an institutionalarrangement whereby results of technicalanalyses are effectively used to inform pol-icy, budgetary, and managerial decisions.Establishing a process where evaluationresults naturally feed into the process ofbudget negotiations is only one piece ofthe puzzle. An even greater challenge is tofoster effective demand for good analysis asa basis for government decision making.

Consistent with its reputation for goodgovernance and competent public adminis-tration, since 1994 Chile has graduallydeveloped a robust system of performancemeasurement and, increasingly,management. Although the Chileanapproach is decentralized (that is, no singleentity is responsible for developing andmaintaining a national system as in Colom-bia), the processes are relatively well coordi-nated by the Interministerial Committee atthe cabinet level.The Ministry of Finance’sDirección de Presupuestos (Budget Direc-torate) also plays a key role in coordinatingvarious decentralized initiatives and usesevaluation results as a basis for defining thebudget framework (Marcel 1997; Armijo2003). Performance is evaluated at the levelof policies and expenditures, organizations,and individual public servants.The govern-ment has begun to link performance evalu-ations to organizational incentives.

Moving to a comprehensive results-based management system will be difficultin most countries. A more realistic objectiveis to build areas of evaluation, debate, and

results-based approaches among particularsets of programs that foster demands for agreater role for evaluation in other parts ofthe government.

The LAC experience with the rapidintroduction of conditional cash transferprograms also illustrates the powerful rolethat can be played by the strategic use ofsound impact evaluations (Rawlings andRubio 2003). Unlike many other social ser-vice delivery reforms, impact evaluationswere included as an integral part of thedevelopment and application of some CCTprograms.These evaluations have servednot only a technical purpose in informingprogram expansion and modification deci-sions, but also a political purpose thatallows policy makers to protect effectiveprograms during political transitions, asPROGRESA/Oportunidades in Mexico exem-plified.This latter case was one in which arigorous evaluation was rare in the country,but the high profile of the PROGRESA/Opor-tunidades evaluations—and theirimportance in underpinning the programcontinuation and extension despite thechange in government—has led to a risingdemand for evaluations in other parts ofthe government and in Congress.The diffi-culties that CCT program administratorsovercame in ensuring that soundevaluations were conducted are also part ofthis important lesson, underscoring theneed to secure a solid commitment frompolicy makers from the program designstage forward to maintain the integrity ofthe program and evaluation designs.

B O X 4 . 2 The Role of Evaluation


(state bureaucrats), and outside (interna-tional actors). The recent Latin Americanexperience provides sufficient evidence thatsuch leadership exists within and outside thestate. Often it may have become ineffectiveby single-mindedly seeking to push a magicbullet solution to service delivery. Shouldn’ta “learning from experience approach” as

advocated above help turn it into a moreeffective source of change? We humbly thinkso. We see processes of systematic explo-ration and evaluation, integrated withindemocratic debate, as key both for the “tech-nical” design issues and for the broaderprocess of empowering all citizens and mak-ing decision makers more accountable.



63

Abrucio, F. 1999. “Reforma do Estado e Contexto FederativoBrasileiro.” Pesquisa 12. Konrad-Adenauer-Stiftung.

Adato, M. 2000. “The Impact of PROGRESA on Community SocialRelationships.” International Food Policy Research Institute,Washington, DC, September.

Adato, M., D. de la Brière, D. Mindek, and A. Quisumbing. 2000.“Final Report: The Impact of PROGRESA on Women’s Statusand Intrahousehold Relations.” International Food PolicyResearch Institute, Washington, DC, July.

ADERASA (Asociación de Entes Reguladores de Agua ySaneamiento de las Américas). Forthcoming. “Las Tarifas deAgua Potable y Alcantarillado en America Latina.” Grupo de Tar-ifas y Subsidios, (Working group on rates and subsidies).

Aglaê de Medeiros Machado, M. 2002. “School Autonomy in Brazil:General Overview and Selected States.” Photocopy, World Bank,Brasilia, Brazil, May.

Alvestegui, A., and J. Sara. 2003. “Estudio de Sustentabilidad de losServicios de Agua Potable Rural en Chile.” Photocopy, WorldBank, Santiago, Chile.

Ames, B. 2003. “The State, Civil Society and Inequality.” Photocopy,University of Pittsburgh, Pittsburgh, Pennsylvania.

Angell, A., P. Lowden, and R. Thorp. 2001. Decentralizing Develop-ment: The Political Economy of Institutional Change in Colombiaand Chile. New York: Oxford University Press.

Angrist, J., E. Bettinger, E. Bloom, E. King, and M. Kremer. 2002.“Vouchers for Private Schooling in Colombia: Evidence from aRandomized Natural Experiment.” American Economic Review92 (5).

———. 2004. “Vouchers for Private Schooling in Colombia: Evi-dence from a Randomized Natural Experiment.” Photocopy,World Bank, Washington, DC.

Armijo, M. 2003. “La Evaluación de la Gestión Pública en Chile.” InEvaluación de Resultados para una Gestión Pública Moderna yDemocrática: Experiencias Latinoamericanas, ed. N. Cunill Grauand S. Ospina Bozzi. Caracas: CLAD (Centro Latinoamericanode Administración para el Desarrollo).

Atkinson, P., and P. van den Noord. 2001. “Managing Public Expen-diture: Some Emerging Policy Issues and a Framework forAnalysis.” Economics Department Working Papers 285, OECD,Paris.

Baeza, C. 1996. “Recursos Humanos: La Agenda Pendiente en laModernización del Sector Público de Salud en América Latina.”Revista del Programa de Administración en Salud (PIAS) 12, Uni-versidad de Chile, November.

Baeza, C., and M. Cabezas. 1998. “La Separación de la Función deCompra y Provisión en la Modernización del Sector Salud.” SerieDocumentos para el Diálogo en Salud 45. Santiago, Chile: CLAISS(Centro Latinoamericano de Investigación de los Sistemas deSalud).

Bardhan, P., and D. Mookherjee. 2000a. “Decentralizing Anti-Poverty Program Delivery in Developing Countries.” WorkingPaper, University of California, Berkeley.

———. 2000b. “Capture and Governance at Local and NationalLevels.” American Economic Review 90 (2): 135–9.

Baum, M. A., and D. A. Lake. 2003. “The Political Economy ofGrowth: Democracy and Human Capital.” American Journal ofPolitical Science 47 (2): 333–47. April.

Bitran, R., J. Munoz, et al. 2000. “Equity in the Financing of SocialSecurity for Health in Chile.” Health Policy 50 (3): 171–196.

Blackburn, J. 2001. “Popular Participation in a Prebendal Society:An Actor Oriented Analysis of Participatory Municipal Planningin Sucre, Bolivia.” PhD diss., unpublished, Institute of Develop-ment Studies, Sussex University.

Bossert, T. J. 2000. “Decentralization of Health Systems in LatinAmerica: A Comparative Study of Chile, Colombia and Bolivia.”LAC Health Sector Reform Initiative. Harvard School of PublicHealth 29 June.

Bossert, T. J., O. Larrañaga, et al. 2003.“Decentralization and Equityof Resource Allocation: Evidence from Colombia and Chile.”Bulletin of the World Health Organization 81 (2): 95–100.

Brown, D. S. 1999. “Reading, Writing and Regime Type: Democ-racy’s Impact on Primary School Enrollment.” Political ResearchQuarterly 52 (4): 681–707. December.

———. 2002. “Democracy, Authoritarianism and EducationFinance in Brazil.” Journal of Latin American Studies 34: 115–141.

Brown, D. S., and W. Hunter. 1999. “Democracy and Social Spend-ing in Latin America, 1980–1992.” American Political ScienceReview 93 (4): 779–90. December.

Buchanan, J. 2001. “Judicial Reform in the Americas.” FOCAL.November. Ontario, Canada.

References

RefCitizens.qxd 2/16/05 12:40 PM Page 63

Burki, S. J., and G. Perry. 1998. “Beyond the Washington Consensus:Institutions Matter.” World Bank Viewpoints. Washington, DC.

Calderon, F., and A. Szmukler. 2004. “Political Culture and Devel-opment.” In Culture and Public Action, ed. V. Rao and M. Walton.Stanford: Stanford University Press.

Campbell, T. 2003. The Quiet Revolution: Decentralization and theRise of Political Participation in L.A. Cities. Pittsburgh: Universityof Pittsburgh Press.

Cantuarias, F., and M. Delgado. 2004. “Peru—Urban Land TitlingProgram.” Paper presented at the Scaling-Up Poverty ReductionConference, Shanghai, China, May.

Carciofi, R., O. Cetrangolo, and O. Larrañaga. 1996. “Desafíos de laDescentralización de la Educación y la Salud en Argentina yChile.” ECLAC, Santiago, Chile.

Carvalho, S., G. Perkins, and H. White. 2002. “Social Funds, Sus-tainability and Institutional Development Impacts: Findingsfrom an OED Review.” Journal of International Development 14:611–25.

CEPAL [Comisión Económica para América Latina y el Caribe (Eco-nomic Commission for Latin America and the Caribbean)]. 2001.“Panorama Social de América Latina.” CEPAL, Santiago, Chile.

Chambers, R., D. Narayan, M. Shah, and P. Petesch. 2000. Crying outfor Change: Voices of the Poor. New York: Oxford University Press.

Collado, C. 2003. “Control Social: Participación de los Consumi-dores en la Regulación y Control de los Servicios Públicos Domi-ciliarios.” Seminario sobre la Participación de la Sociedad Civilen la Regulación de los Servicios Públicos, Seminar on Civil Soci-ety Participation in the Regulation of Public Services, Lima,Peru, August.

Correa Sutil, J. 1999. “Judicial Reforms in Latin America: GoodNews for the Underprivileged?” In The (Un)Rule of Law and theUnderprivileged in Latin America, ed. J. Mendez, G. O’Donnell,and P.S. Pinheiro. Notre Dame, Indiana: University of NotreDame Press.

Cotlear, D. 2000. “Peru: Reforming Health Care for the Poor.”LCSHD, Working Paper Series 57, World Bank, Washington, DC,March.

Cox, C., and P. González. 1998. “Educación: de Programas de Mejo-ramiento a Reforma.” In Construyendo Opciones: PropuestasEconómicas y Sociales para el Cambio de Siglo, ed. Cortázar andVial. Santiago, Chile: CIEPLAN (Corporación de Estudios paraAmérica Latina).

Crouch, L. 2004. “Political Economy, Incentives, and Teachers’Unions: Case Studies in Chile and Peru.” Photocopy, WorldBank, Washington, DC.

De Ferranti, D., G. Perry, H. G. Ferreira, et al. 2004. “Inequality inLatin America: Breaking with History?” World Bank, Washing-ton, DC.

Diaz-Cayeros, A., B. Magaloni, and B. Weingast. 2002. “Federalismand Democratisation in Mexico.” Paper presented at the 2002meeting of the American Political Science Association, Washing-ton, DC.

Di Gropello, E. 2002. “An Assessment of the Impact of Decentral-ization on the Quality of Education in Chile.” Economist’s Forum.World Bank, Washington, DC.

Di Gropello, E., and J. Marshall. 2004. “Teacher Effort and StudentAchievement in Rural Honduras.” Photocopy, World Bank,Washington DC.

Dodson, J., and D. Jackson. 2000. “Democratic Transitions in Cen-tral America: Justice Reform and the Civic Culture.” Paper pre-sented at the Institutions, Accountability, and Democratic Gov-ernance in Latin America conference, the Helen Kellogg Institutefor International Studies, University of Notre Dame, May 8–9.

Domingo, P., and R. Sieder. 2001. “Conclusions.” In Rule of Law inLatin America: The International Promotion of Judicial Reform,ed. P. Domingo and R. Sieder. London: Institute of Latin Ameri-can Studies.

Dorotinsky, W., and Y. Matsuda. 2001. “Reforma de la gestiónfinanciera en América Latina: una perspectiva institucional.”Reforma y Democracia 23: 141–66.

DNP (Departamento Nacional de Planeación). 2002. “Exposiciónde motivos de la reforma a la Ley 60 de 1993: Sector Educación ySector Salud.” Archivos de Economía 173.

Egaña, R. 2003. “Profesionalización de la función pública en Chile:Contexto, avances y desafíos.” In Retos de la Profesionalización de la Función Pública, ed. D. Arellano, R. Egaña, O. Oszlak, and R. Pacheco. Caracas: CLAD.

England, R. 1998. “Health Sector Reform: Human ResourcesIssues—Underlining the Importance of Human Resource Issuesfor Commonwealth Countries.” Institute for Health Sector Devel-opment 14. London.

Estache, A., V. Foster, and Q. Wodon. 2002. “Accounting for Povertyin Infrastructure Reform: Learning from L.A.’s Experience.”Development Studies Series, World Bank, Washington, DC.

Evans, P. 2002. “Beyond Institutional Monocropping: Institutions,Capabilities and Deliverative Development.” Photocopy, Univer-sity of California, Berkeley, California, January.

Fay. 2000. “Financing the Future: Infrastructure Needs in LatinAmerica 2000–2005.” World Bank Working Paper No. 2545,World Bank, Washington, DC.

Felicio, M. and I. John-Abraham. 2004. "Perú: Toward a System ofSocial Accountability," En Breve No. 39, World Bank, Washing-ton, DC.

Fiedler, J. 2002. “Resource Allocation and Purchasing in LatinAmerica and Caribbean: A Regional Review.” HNP, DiscussionPaper, ed. A. S. Preker, World Bank, Washington, DC.

Foster, V. 2002. “Ten Years of Water Service Reform in Latin Amer-ica: Towards an Anglo-French Model.” In Reinventing Water andWastewater Systems, ed. P. Seidensat, D. Haarmeyer, and S. Hakim.New York: John Wiley & Sons Inc.

———. 2003. “Notes on Infrastructure in Latin America.” Photo-copy, World Bank, Washington, DC.

Foster, V., and O. Irusta. 2003. “Does Infrastructure Reform workfor the Poor? A Case Study on the Cities of La Paz and El Alto inBolivia.” World Bank Policy Research, Working Paper 3177,World Bank, December.

Fox, J. 2000. “Civil Society and Political Accountability: Proposi-tions for Discussion.” Paper presented at the Institutions,Accountability, and Democratic Governance in Latin Americaconference, the Helen Kellogg Institute for International Studies,University of Notre Dame, May 8–9.



———. 2001. “Vertically Integrated Policy Monitoring: A Tool forCivil Society Policy Advocacy.” Nonprofit and Voluntary SectorQuarterly 30 (3): 616–27. September.

Galiani, S., and E. Schargrodsky. 2002. “Evaluating the Impact ofSchool Decentralization on Educational Quality.” Economía 2(2), Journal of the Latin American and Caribbean EconomicAssociation.

Galiani, S., P. Gertler, and E. Schargrodsky. 2002. “Water for Life:The Impact of the Privatization of Water Services on Child Mor-tality.” Photocopy, University of San Andrés, Buenos Aires,Argentina.

Galindo, P. 2003. “Indicadores Subjetivos: Estudios, Calificacionesde Riesgo y Encuestas de Percepción Pública sobre los Sistemasde Justicia.” Centro de Estudios de Justicia de las Américas, Sis-temas Judiciales, Resultados Recientes para las Américas 3 (6):4–35. Santiago, Chile.

Gallego, F. 2002. “Competencia y Resultados Educativos: Teoría yEvidencia para Chile.” Cuadernos de Economía 39 (118): 309–52.

Garibaldi, J. A. N.d. “Instituciones Informales y Desempeño delSector Público en México.” Photocopy, FLACSO and InstitutoTecnológico Autónomo de México.

Garretón, M. A. 2002. “La Transformación de la Acción Colectivaen América Latina.” Revista de la CEPAL 76. April.

Garro, A. M. 1999. “Access to Justice for the Poor in Latin America,”In The (Un)Rule of Law and the Underprivileged in Latin America,ed. J. Mendez, G. O’Donnell, and P.S. Pinheiro. Notre Dame,Indiana: University of Notre Dame Press.

Gomez Lobo, A., and D. Contreras. 2000. “Subsidy Policies for theUtility Industries: A Comparison of the Chilean and ColombianWater Subsidy Schemes.” Photocopy, University of Chile,Department of Economics, Santiago, Chile.

Government of Jamaica. 2003.“Government at Your Service: PublicSector Modernisation Vision and Strategy 2002–2012.” MinistryPaper 56, Kingston, Jamaica. January.

Gray Molina, G. 2002. “The Offsprings of 1952: Poverty Exclusionand the Promise of Popular Participation.” Paper presented at theconference on The Bolivian Revolution at 50: ComparativeViews on Social, Economic and Political Change, Harvard Uni-versity. Cambridge, MA.

Gray Molina, G., E. Pérez de Rada, and E. Yeñez. 1999. “LaEconomíca Política de Reformas Institucionales en Bolívia.”Working Paper R-350, Inter-American Development Bank(IDB), Office of the Chief Economist, Washington, DC.

Griffiths, M., and J. S. McGuire. 2004. “A new dimension for healthreform: Atención integral a la niñez en la comunidad en Hon-duras.” Photocopy, World Bank, Washington DC.

Grindle, M. S., and E. S. Mason. 2004. Despite the Odds: ContentiousPolitics and Education Reform. Princeton: Princeton UniversityPress.

Guasch, J. L. 2004. “Granting and Renegotiating InfrastructureConcessions: Doing It Right.” WBI Development Studies. April.

Guedes, A., T. Lobo, R. Walker, and A. L. Amaral. 1997. “GestiónDescentralizada de la Educación en el Estado de Minas Gerais,Brasil.” Working Paper 11, Grupo de Desarrollo Humano,América Latina y el Caribe, World Bank, Washington, DC, May.

Hall, G., and H. Patrinos. 2004. Forthcoming. “Indigenous People,Poverty and Human Development in Latin America,1994–2004.” Washington, DC: World Bank.

Henderson, K., A. Shah, S. Elena, and V. Autheman. 2004. “RegionalBest Practices: Enforcement of Judgments: Lessons from LatinAmerica.” IFES, Rule of Law Paper Series, Washington, DC,April.

Hicks, N., and J. Todd. 2003. “A Brief Examination of the Relation-ship between GDP/capita and Performance on Selected SocialIndicators in the Latin America and Caribbean Region.” Photo-copy, World Bank, Washington, DC.

Holm-Nielsen, L., K. Thorn, and J. Prawda. 2004.“Chile: Decades ofEducational Reform Deliver.” World Bank En Breve 44.

Hsieh, C. T., and M. Urquiola. 2003. “When Schools Compete, Howdo They Compete? An Assessment of Chile’s Nationwide SchoolVoucher Program.” Working Paper Series (U.S.); No. 10008:1-(43), National Bureau of Economic Research, September.

Jiménez, E., and Y. Sawada. 1998. “Do Community ManagedSchools Work? An Evaluation of El Salvador’s EDUCO Pro-gram.” World Bank Economic Review 13 (3): 415–41.

Kaufman, R. R., and J. Nelson. 2004. Crucial Needs, Weak Incentives:Social Sector Reform, Democratization, and Globalization in LatinAmerica. Baltimore: John Hopkins University Press.

Kaufman, R. R., and A. Segura-Ubiergo. 2001. “Globalization,Domestic Politics and Social Spending in Latin America: A TimeSeries Cross-Section Analysis, 1973–97.” World Politics 53 (4):553–87. July.

Keefer, P. 2002. “Clientelism, Credibility and Democracy.” Photo-copy, World Bank, Washington, DC.

Keefer, P., and S. Khemani. 2003. “Democracy, Public Expendituresand The Poor.” Policy Research Working Paper 3164, WorldBank, Washington, DC, November.

King, E., P. Orazem, and D. Wohlgemouth. 1999.“Central Mandatesand Local Incentives: The Colombian Education Voucher Pro-gram.” World Bank Economic Review 13 (3): 467–91.

King, E., and B. Ozler. 1998. “What’s Decentralization Got to Dowith Learning? The Case of Nicaragua’s School AutonomyReform.” Working Paper Series on Impact Evaluation of Educa-tion Reforms 9, World Bank, Washington, DC.

King, E., L. Rawlings, M. Gutierrez, C. Pardo, and C. Torres. 1997.“Colombia’s Targeted Education Voucher Program: Features, Cov-erage and Participation.” Working Paper Series on Impact Evalua-tion of Education Reforms 3, World Bank, Washington, DC.

La Forgia, G., P. Mintz, and C. Cerezo. 2004. “Is the Perfect theEnemy of the Good? A Case Study of Large-Scale Contracting forBasic Health Services in Rural Guatemala.” Photocopy, WorldBank, Washington, DC, March.

Lake, D. A., and M. A. Baum. 2001. “The Invisible Hand of Democ-racy: Political Control and The Provision of Public Services.”Comparative Political Studies 34 (6): 587–621. August.

Laking, R. 2002. “Agencies: Their Benefits and Risks.” Photocopy,Victoria University of Wellington, New Zealand.

Maceira, D., and M. V. Murillo. 2001.“Social Sector Reform in LatinAmerica and the Role of Unions.” Working Paper #456, IDBResearch Department, Washington, DC.

References 65


Mainwaring, S. 2003. “Introduction: Democratic Accountability inLatin America.” In Democratic Accountability in Latin America, ed.S. Mainwaring and C. Welna. New York: Oxford University Press.

Marcel, M. 1997. “Modernización del Estado e Indicadores deDesempeño en el Sector Público: La Experiencia Chilena.”Departamento de Ingeniería Industrial, University of Chile, San-tiago, Chile.

Massad, C. 1996. “Health Care Reform in Chile.” Revista Médica deChile 124 (7): 880–6.

McClymont, M., and S. Golub, eds. 2000. Many Roads to Justice: TheLaw-Related Work of Ford Foundation Grantees Around the World.New York: The Ford Foundation.

McEwan, P., and M. Carnoy. 1999. “The Impact of Competition onPublic School Quality: Longitudinal Evidence from Chile'sVoucher System.” Photocopy, Stanford University, Palo Alto.

———. 2000. “The Effectiveness and Efficiency of Private Schoolsin Chile's Voucher System.” Education Evaluation and PolicyAnalysis, vol. 22, No. 3, 213–39.

McKenzie, D., and D. Mookherjee. 2002. “Distributive Impact ofPrivatization in Latin America: An Overview of Evidence fromFour Countries.” Photocopy, Departments of Economics Stan-ford University and Boston University, Palo Alto and Boston.

Méndez, J. 2000. “Legislatures, Judiciaries and Innovations in Hori-zontal Accountability.” Paper presented at the Institutions,Accountability, and Democratic Governance in Latin Americaconference, the Helen Kellogg Institute for International Studies,University of Notre Dame, May 8–9.

Ministerio de Salud. 2000. “Sistematización de la Estrategia Integralen Salud Materno Infantil en la Provincia de Paruro.” Photocopy,Ministerio de Salud, Dirección Regional de Salud Cusco, UBASS.

Molinar, J., and J. Weldon. 1994. “Electoral Determinants and Con-sequences of National Solidarity.” In Transforming State SocietyRelations in México, ed. W. Cornelius, A. C. Craig, and J. Fox. SanDiego: University of California at San Diego.

O’Donnell, G. 1993. “On the State, Democratization and SomeConceptual Problems: A Latin American View with Glances atSome Postcommunist Countries.” World Development 21 (88):1355–69.

———. 1994. “Delegative Democracy.” Journal of Democracy 5 (1):55–69. January.

———. 1996. “Illusions about Consolidation,” Journal of Democ-racy 7 (2): 34–51. April.

———. 1998.“Accountability Horizontal.” AgorA 8: 5–34. Summer.

———. 2001. “Reflections on Contemporary South AmericanDemocracies.” Journal of Latin American Studies 33: 599–609.

OECD (Organisation for Economic Co-operation and Develop-ment). 2003. “Reviews of National Policies for Education inChile.” OECD, Paris. October.

Olive, J. M., J. B. Risi, Jr., et al. 1997. “National Immunization Days:Experience in Latin America.” The Journal of Infectious Diseases175 (suppl 1): S189–93.

Over, M., and N. Watanabe. 2003. “Evaluating the Impact of Orga-nizational Reforms in Hospitals.” In Innovations in Health ServiceDelivery: The Corporatization of Public Hospitals, ed. A. Prekerand A. Harding, 105–65. Washington, DC: World Bank.

Patrinos, H. 2002. “Contracting Out in Education.” Photocopy,World Bank, Washington, DC.

Pasara, L. 2002. “Justicia y Ciudadanía Realmente Existentes.”Política y Gobierno IX (2). September.

Paxson, C., and N. Schady. 2002. “The Allocation and Impact ofSocial Funds: Spending on School Infrastructure in Peru.” WorldBank Economic Review 16 (2): 297–319.

Payne, J. M., D. Zovatto G., F. Carrillo Flórez, and A. AllamandZaval. 2002. Democracies in Development: Politics and Reform inLatin America. Washington, DC: Inter-American DevelopmentBank and International Institute for Democracy and ElectoralAssistance.

Peruzzotti, E., and C. Smulovitz. 2001. “Civil Society, the Media andInternet as Tools for Creating Accountability to Poor and Disad-vantaged Groups.” Occasional Paper 2002/13, UNDP (UnitedNations Development Programme), Human DevelopmentReport Office. November.

Pinheiro, P. S. 1999. “Introduction in Mendez.” In The (Un)Rule ofLaw and the Underprivileged in Latin America, ed. J. Mendez,G. O’Donnell, and P. S. Pinheiro. Notre Dame, Indiana: Univer-sity of Notre Dame Press.

Pritchett, L., and M. Woolcock. 2002. “Solutions When The Solu-tion Is the Problem: Arraying the Disarray in Development.”Working Paper 10, Center for Global Development. September.

Przeworski, A., S. Stokes, and B. Mannin. 1999. “Introduction.” InDemocracy, Accountability and Representation, ed. Przeworksi etal. Cambridge: Cambridge University Press.

Rawlings, L. B., and G. M. Rubio. 2003. “Evaluating the Impact ofConditional Cash Transfer Programs.” Policy Research WorkingPaper 3110, World Bank, Washington, DC, August.

Rawlings, L. B., L. Sherburne-Benz, and J. Van Domelen. 2004. Eval-uating Social Funds: A Cross-Country Analysis of CommunityInterventions. Washington, DC: World Bank.

Reimers and McGinn. 1997. Informed Dialogue: Using Research toInform Education Policy around the World. Westport: PraegerPublishers.

Rinne, J. J. 2001. “Redesigning the State in Latin America: Pundits,Policymakers, and Organized Labor in Argentina and Brazil.”PhD diss. Department of Politics, Princeton University.

Robinson, J. 2003. “Politician-Proof Policy?” Background paper tothe 2004 WDR. April 12.

Robinson, J., and T. Verdier. 2002. “The Political Economy of Clien-telism.” Working Paper 3205, CEPR (Centre for Economic PolicyResearch), London.

Salamon, L., H. Anheier, R. List, S. Toepler, S. Sokolowski, and Asso-ciates. 1999. “Global Civil Society: Dimensions of the NonprofitSector.” The Johns Hopkins Center for Civil Society Studies,Baltimore.

Santamaría, M. 2003. “Income Inequality, Skills and Trade: Evi-dence from Colombia During the 80s and 90s.” Photocopy,World Bank, Washington, DC.

Sapelli, C., and B. Vial. 2002. “The Performance of Private and Pub-lic Schools in the Chilean Voucher System.” Cuadernos de Econo-mia 39 (118): 423–54.



Schady, N. 2000. “The Political Economy of Expenditures by thePeruvian Social Fund (FONCODES).” American Political ScienceReview 94 (2): 289–304.

Schedler, A. 1999. “Restraining the State: Conflicts and Agents ofAccountability.” In The Self-Restraining State, ed. A. Schedler,L. Diamond, and M. Plattner. Boulder: Lynne Rienner PublishersInc.

Schedler, A., L. Diamond, and M. Plattner. 1999. “Introduction.” InThe Self-Restraining State, ed. A. Schedler, L. Diamond, and M. Plattner. Boulder: Lynne Rienner Publishers Inc.

Schiefelbein, E., and P. Schiefelbein. 2000. “Three DecentralizationStrategies in Two Decades: Chile 1981–2000,” Journal of Educa-tional Administration 38 (5).

Sen, A. K., and J. Drèze. 1989. Hunger and Public Action. Oxford:Clarendon Press.

Shefter. M. 1994. Political Parties and the State: The American His-torical Experience. Princeton: Princeton University Press.

Shepherd, G. 2003. “Civil Service Reform in Developing Countries:Why Is It Going Badly?” Paper presented at the 11th Interna-tional Anti-Corruption Conference, Seoul, Republic of Korea.

Shifter, M. 1997. “Tensions and Trade-Offs in Latin America.” Jour-nal of Democracy 8 (2). April.

Shugart, M. S., E. Moreno, and B. F. Crisp. 2000. “The Accountabil-ity Deficit in Latin America.” Paper presented at the Institutions,Accountability, and Democratic Governance in Latin Americaconference, the Helen Kellogg Institute for International Studies,University of Notre Dame, May 8–9.

Smulovitz, C., and E. Peruzzotti. 2000. “Societal Accountability: TheOther Side of Control.” Journal of Democracy 11 (4). October.

Smulovitz, C., M. Walton, and P. Petesch. 2003. “Notes on Evaluat-ing Empowerment.” Photocopy, World Bank, Latin America andCaribbean Region, Washington, DC, September.

Solo, T. M. 2003. “Independent Water Entrepreneurs in LatinAmerica. The Other Private Sector in Water Services.” Photo-copy, Private Sector Development and Infrastructure, Energy,and Water Department, and Department of Finance, WorldBank, Washington, DC.

Spink, P. 1999.“Possibilities and Political Imperatives: Seventy Yearsof Administrative Reform in Latin America.” In Reforming the State: Managerial Public Administration in Latin America, ed.L. C. Bresser Pereira and P. Spink. Boulder: Lynne Rienner Publishers.

Swope, J., M. Latorre, and F. Celedón. 1998. “Estudio EficienciaInterna de las Escuelas Formales de Fe y Alegría.” Santiago: CIDE(Centro de Investigación y Desarrollo de la Educación).

Tendler, J. 1997. Good Government in the Tropics. Baltimore andLondon: The Johns Hopkins University Press.

———. 2000. “The Rise of Social Funds: What Are They a ModelOf?” Photocopy, MIT, Boston.

Tommasi, M., S. Saiegh, and P. Sanguinetti. 2001. “Fiscal Federalismin Argentina: Politics, Policies and Institutional Reform.” Econo-mia 1 (2): 157–201.

Vega, J., P. Bedregal, et al. 2003. “Gender Inequity in the Access toHealth Care in Chile.” Revista Médica de Chile 131 (6): 669–78.

Wilson, B. M., and R. Handeberg. 1998. “Opening Pandora’s Box:The Unanticipated Political Consequences of Costa Rican LegalReform.” Paper presented at Midwest Political Science Associa-tion, Chicago, April 23–25.

World Bank. 1999. Beyond the Center: Decentralizing the State.Washington, DC: World Bank.

———. 2000a. World Development Report: Attacking Poverty. NewYork and Washington, DC: Oxford University Press, IBRD,World Bank.

———. 2000b. “Bolivia from Patronage to a Professional State:Institutional and Governance Review.” Report 20115-BO, WorldBank, Washington, DC.

———. 2001a. “Peru. Institutional & Governance Review.” Report22637-PE Latin America and The Caribbean, Poverty Reductionand Economic Management Unit, World Bank, Washington, DC,August.

———. 2001b. “Argentina: An Assessment of Institutional Capac-ity for Social Sector Reform in Argentina.” Report 21557-AR: 81,Latin America and The Caribbean: Argentina, Chile, Paraguayand Uruguay Country Management Unit, World Bank, Washing-ton, DC, September.

———. 2002a. Empowerment and Poverty Reduction: A Sourcebook.Washington, DC: World Bank.

———. 2002b. “Uruguay—Health Sector Development ProjectImplementation Completion Report.” Report 38, World Bank,Washington, DC.

———. 2002c. “Rural Water Supply and Sanitation Toolkit forMultisector Projects.” Photocopy, World Bank, Washington, DC.

———. 2003a. Making Services Work for Poor People. World Devel-opment Report 2004. Washington, DC: World Bank.

———. 2003b. “Brazil Municipal Education: Resources, Incentivesand Results.” Photocopy, World Bank, Washington, DC.

———. 2003c. “Venezuela—Health Service Reform Project Imple-mentation Completion Report.” Report 54, World Bank, Wash-ington, DC.

———. 2003d. “State-Society Synergy for Accountability.” Photo-copy, Civil Society Team, Latin America and the CaribbeanRegion, World Bank, Washington, DC, April.

———. 2004. “Voice, Eyes and Ears: Social Accountability in LatinAmerica.” Civil Society Team, Latin America and the CaribbeanRegion, World Bank, Washington, DC.

World Development Indicators (WDI). World Bank Database.http://devdata.worldbank.org/dataonline.

Zakaria, F. 1997. “The Rise of Illiberal Democracy.” Foreign Affairs76 (6): 22–43.

References 67


Documents

Citizens, Politicians, and Providerssiteresources.worldbank.org/BRAZILINPOREXTN/...1.1 A Stylized Presentation of Accountability Relationships 3 2.1 Secondary Education Coverage for