Routes to Better Service Delivery 51 the Centrality of the “Fitting Process” 57

Public Disclosure Authorized

Public Disclosure Authorized Citizens, Politicians, and Providers Public Disclosure Authorized

The Latin American Experience with Service Delivery Reform

Ariel Fiszbein, Editor Public Disclosure Authorized

Citizens, Politicians, and Providers

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

Ariel Fiszbein, Editor © 2005 The International Bank for Reconstruction and Development / The World Bank 1818 H Street, NW Washington, DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org E-mail: [email protected]

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 Bank for 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. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgement on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries.

Rights and Permissions The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development / The World Bank encourages dissemination of its 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 with complete 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 addressed to 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 0-8213-6089-2 978-8213-6089-7 e-ISBN 0-8213-6090-6

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

Preface vii Acknowledgments viii Abbreviations ix

1 Introduction 1 Service Delivery Progress 1 Latin American Political and Institutional Context 4 Notes 6

2 Service Delivery and Social Outcomes: A Story of Successes and Failures 7 Progress in Basic Service Coverage 7 Service Quality 11 Social Outcomes 12 Notes 14

3 LAC Service Delivery: Assessing Two Decades of Change 15 Reforming the Compact 15 Enhancing Client Power 29 Enhancing State Accountability to Citizens 36 Notes 45

4 Conclusions 49 Routes to Better Service Delivery 51 The Centrality of the “Fitting Process” 57

References 63

v vi CONTENTS

Boxes 1.1 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 3.8The Role of Community Participation and Cultural Adaptation in 35Service Delivery 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

Figures 1.1 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 Income Decile, 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

Table 4.1 Fitting Approaches to Country Conditions 59 Preface

Children regularly receiving health visits and education, the sick receiving proper and timely health care, safe water ﬂowing out of the tap, electricity reliably reaching homes and businesses—these apparently simple events are taken for granted in developed countries. In Latin America, despite two decades of social and infrastructure improvements, the poor and many of the middle class make do with low-quality services. Far too many of the poor receive no services. Improving service delivery to the poor is both a widespread political demand and central to the realization of the Millennium Development Goals (MDGs). This book interprets service delivery successes and failures in Latin America and provides guidance to policymakers and development practitioners on shaping public action to provide better-quality services for all. Its analysis builds on the accountability framework developed in the World Bank’s World Development Report 2004: Making Services Work for Poor People, which emphasizes the behavior of people—from teachers to administrators, politicians, and rich and poor citizens—within the chain of interactions from demand to actual service delivery. The report seeks to answer an essential question: If accountability relationships among citizens, policymakers, and service providers are key to effective service delivery and there have been both systemic reforms (expanding national and local democracy) and an array of speciﬁc experiments (privatization, increased choice), why is service delivery in Latin Amer- ica still so inequitable and often of low quality?

vii Acknowledgments

A World Bank staff team prepared this report. Led by Team Leader Ariel Fiszbein, the team included, 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 Santamaría (service delivery and social outcomes), Michael Walton (voice and empowerment), and David Warren (social funds). Contributing to the report were, Kathy Bain, Lisa Bhansali, Jack Fiedler, Gillette Hall, Linn Hammergren, Amy Kirkley, Jerry La Forgia, Heather Marie Layton, Edgardo Mosqueira, Harry Patrinos, William Reuben, Keta Ruiz, Mario Sangines, Jennifer Sara, Roby Senderowitsch, and Joseph Shapiro. The report beneﬁted from comments from, Joachim von Amsberg, Ana-María Arriagada, Shanta Devarajan, Ron Myers, Guillermo Perry, Fernando Rojas, Elena Serrano, Luis Servén, and Eduardo Velez. Tomas Casas was responsible for the production of the report.

viii Abbreviations

ADERASA Asociación de Entes Reguladores de Agua IFMS Guatemala Integrated Financial Management Potable y Saneamiento de las Américas System AIN-C Comprehensive Attention to Childhood in ISAPRES a private health insurer (Chile) the Community program (Honduras) IT information technology CCT conditional cash transfer LAC Latin America and the Caribbean region CEO chief executive ofﬁcer LAL Latin American Laboratory CEPAL Comisión Económica para América Latina y MDGs Millennium Development Goals el Caribe (Economic Commission for Latin NGO nongovernmental organization America and the Caribbean) OECD Organisation for Economic Co-operation CLAS local health administration committees and Development (Peru) PISA Program for International Student COFOPRI Comisión para la Formalización de la Assessment Propiedad Informal (Peru) PPP purchasing power parity EA executive agencies SIMCE standardized school scores EAA Executive Agencies Act TIMSS Trends in International Mathematics and EDUCO Educación con Participación de la Science Study Comunidad UNDP United Nations Development Programme FONASA National Health Fund (Chile) WDI World Development Indicators FONCODES Fondo de Compensación y Desarrollo Social WDR World Development Report (Peru)

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

Introduction

Children going to school and learning, par- between groups of people. But these rela- ents taking their children for regular health tionships are embedded in formal and visits, the sick receiving proper and timely informal institutional structures. Influence chapter health care, safe water flowing out of the tap, patterns on service delivery can work reliable electricity supplies—these appar- through two broad pathways: first, what is ently simple events are taken for granted in termed the “long route” from citizens to developed countries. However, in most of politicians to service delivery agencies to 1 Latin America (as in most developing coun- service providers; and second, the “short tries), only the privileged have reliable and route” of direct interactions between indi- quality access to such services, and they viduals, communities, and front-line work- often achieve this by using high-cost private ers. These considerations apply to some solutions. The poor and many of the middle degree to all public action areas—from classes suffer low quality services, and, in macroeconomic policy to policing. But they the poorer countries and regions, often no are particularly important to service deliv- service at all. How come? While technical ery areas that involve intensive transactions difficulties may still be a binding constraint and discretionary behavior of providers— for some sophisticated services, they are features typical of the full panoply of basic clearly not a bottleneck for the most essen- services that are central to the lives of citi- tial ones; it is not primarily a technical zens, from schooling to water supply. problem when children systematically fail In this introduction we outline service to learn to read and write or a community delivery progress and introduce the prism does not receive reliable supplies of safe through which we believe service delivery drinking water. should be interpreted, particularly in the This report seeks to answer this ques- Latin American political and institutional tion. In so doing, it aims to provide guid- context. We then provide an interpretation ance as well to policy makers and develop- of how this context can help understand the ment practitioners on how to shape public current state of service delivery and sketch action to get better quality services for all. the public action implications. The ideas This challenge is both a widespread political and results are developed at greater length demand throughout the region and central in the body of the report. to the realization of the Millennium Devel- opment Goals (MDGs). This report’s major Service Delivery Progress theme is that to understand questions of Latin America’s political and economic his- service access and quality, we need to exam- tory is associated with a tradition of ine people’s behaviors, from teachers to unequal service delivery, with the poor typ- administrators, politicians, and rich and ically suffering from a severe lack of decent poor citizens. The big issue is whether those quality services. Conditions have been bet- responsible for designing and delivering ter in the relatively richer societies, notably services are accountable to the citizens who in the Southern Cone and to a certain are demanding the services—and also pay- extent in Mexico. They are typically worse ing the taxes and fees that finance services. in poorer and more socially fractured soci- This is fundamentally about relationships eties, such as Bolivia and Guatemala. They

1 2 Citizens, Politicians, and Providers: The Latin American Experience with Service Delivery Reform

are worst in Haiti, the region’s poorest and Latin America (Hall and Patrinos, forth- institutionally weakest society. In particular, coming) and between Afro-Latinos and the 1980s debt crisis hurt progress in public others. provisioning in most countries in the With respect to quality, the education region. sector has specific evidence of low test However, there has been significant scores at least relative to countries of the progress in the past decade or so; chapter 2 Organisation for Economic Co-operation reviews this progress. Progress is most clearly and Development (OECD). There are wide- seen in service coverage, notably in educa- spread reports of dissatisfaction with ser- tion, health, and basic infrastructure water vice delivery, especially from the poor, and electricity services. These gains were notably from country studies in the region associated with expanded public spending for on the Voices of the Poor (Chambers et al. social services. 2000). CEPAL [Comisión Económica para To interpret the Latin American experi- América Latina y el Caribe (Economic Com- ence, this report builds on the framework mission for Latin America and the developed in the 2004 World Development Caribbean)] (2001) estimates that during the Report (WDR). The WDR sought answers 1990s, total per capita social spending for observed service delivery failures world- increased by about 50 percent, if pensions are wide. It developed a framework that included, and 30 percent if they are not. By emphasizes people’s behavior within the contrast, during the same period, public chain of interactions from demands of citi- infrastructure services expenditures fell from zens to actual service delivery—whether 1.2 percent of GDP to 0.6 percent of GDP, this is by public, private, or nongovernmen- declines that were offset by surging private tal organizations (NGOs). For example, for investment in the middle of the decade, but children to learn how to write, there needs not since. to be some type of school, the children need In the poorer countries, an important to be present, teachers need to show up and agenda of increased coverage remains. To be paid, and the teachers need to have the varying degrees, all countries suffer prob- skills, resources, and motivation required to lems of both inequality and quality. With genuinely teach. Behind this there are a respect to inequality, in Mexico, the average whole set of interactions from policy and person in the bottom quintile of the income budgetary decisions, through the perfor- distribution has 3.5 years of schooling, mance of ministries of education to recruit- compared with 11.6 years for a person in ment and the influence of mayors and com- the top quintile. In Brazil, a child born to a munities. When any link in the chain breaks household in the poorest 20 percent of the (i.e., children do not come to school, the distribution is three times as likely to die teacher is not present, there are no inputs, before reaching the age of five as a child and so on), the entire process may fail. born into a household in the top fifth. A Whether a system works depends on how Guatemalan family whose income is in the responsibilities are assigned, what capabili- bottom 20 percent of the distribution has a ties those in charge actually have, what 57 percent chance of being connected to the accountability mechanisms exist for those water mains and a 49 percent probability of in charge, and what incentives are in place having access to electricity, while a house- to make the accountability mechanisms hold in the top 20 percent of the distribu- effective in practice. All these factors tend to tion enjoys rates of 92 and 93 percent, be deeply embedded in social and political respectively. Throughout the region, the relations. bulk of social protection spending goes to The WDR conceptualized an approach formal social security systems that primar- that focuses on accountability mecha- ily benefit those in the upper part of the nisms and power relations between policy income distribution. Such inequalities are makers, providers, and citizens/users. The particularly noteworthy between indige- approach emphasizes the role of three such nous and nonindigenous people across relationships: Introduction 3

1. The “political voice” of different citizen Figure 1.1 A Stylized Presentation of Accountability Relationships groups over policy makers in shaping policy design;

2. The “compact” between policy makers POLITICIANS AND and service providers, whether they be POLICYMAKERS public, private, or nongovernmental; and 3. The “client power” (either through the exercise of a more direct form of voice or through choice) of citizens vis-à-vis the voice/ compact/ service providers. politics policies Figure 1.1 illustrates these relationships, and box 1.1 provides a glossary of terms relevant to this framework. CITIZENS, RICH PUBLIC, PRIVATE, Within such a framework there are two AND POOR, POWERFUL client power CIVIC PROVIDERS; inﬂuence pathways for service delivery for AND WEAK FRONT-LINE WORKERS the poor (or nonpoor). The so-called “long route” involves the inﬂuence of citizens on Source: World Bank 2003a.

BOX 1.1 A Glossary of Terms for a Service Delivery Accountability Framework

Actors: Individuals, households, communities, Compacts: The broad, long-term accountability while policy makers implement those direc- firms, governments, and other public or non- relationship that connects policy makers to tions and also set and enforce the governmental organizations that finance, organizational providers.This relationship conditions under which service providers produce, regulate, deliver, or consume can, but need not, take the form of a operate. Accountability subrelationships services. contract. between politicians and policy makers are Accountability: A set of relationships among Long and short routes of accountability: usually derived from the constitution or service delivery actors with five features: the Clients may hold service providers account- administrative law. understanding that a service or goods able for performance in two ways. First, the Service delivery framework or chain: The four embodying a service will be supplied (dele- short route connects clients and providers service-related actors—citizens/clients, gating); paying for the service (financing); directly.When such direct client power is politicians/policy makers, organizational supplying the service (performing); obtain- weak or not possible to use, clients must use providers, and front-line professionals—and ing relevant information and evaluating per- voice and voting as citizens to hold politi- the four relationships of accountability that formance against expectations; and being cians accountable. In turn, politicians and connect them: able to impose sanctions for inappropriate policy makers must use their compacts with • Voice and politics: Connects citizens and performance or provide rewards for appro- service providers to hold those providers politicians priate performance (enforcing). accountable.The long route of accountabil- • Compacts: Connects politicians/policy Client power: The accountability relationship ity combines the actions of clients pressur- makers and providers. connecting clients to the front-line service ing politicians and policy makers/politicians providers, usually at the point of service pressuring providers. • Management: Connects provider organizations with front-line professionals delivery, based on interactions through Organizational providers: Public, private, and which clients monitor suppliers and express nonprofit entities that provide services. • Client power: Connects clients with their service demands. These can range from a government min- providers. Clients/citizens: Citizens participate in political istry with thousands of employees to a com- Voice and politics: This is the most complex processes to shape and attain goals. Clients munity-run school serving a village. accountability relationship. It connects citi- receive services to satisfy their household Politicians and policy makers: The service zens and politicians and includes formal and demands. Individuals can act as both clients delivery actors that discharge the state’s leg- informal processes such as voting, electoral and citizens islative and regulatory responsibilities. Politi- politics, lobbying, propaganda, patronage, Clientelism: The tendency of politicians as cians may be elected or may reach their clientelism, media activities, information patrons to respond to political competition positions through undemocratic means. access, and others. Citizens delegate to by excessively favoring one group of They can also be policy makers—for exam- politicians the functions of serving citizen clients over another in return for a political ple, a general who is president, but also runs interests and financing government advantage. Politicians often practice clien- the military. Policy makers are more through taxes. Politicians perform by provid- telism by providing narrow supporter commonly the highest unelected officials, ing services. Citizens enforce accountability groups with free public services or public either from the civil service or appointed. through the short and long routes discussed employment. Politicians set the general policy direction, above. 4 Citizens, Politicians, and Providers: The Latin American Experience with Service Delivery Reform

policy makers—working through political with freedom of expression and association. institutions—who then provide the policy The shift from authoritarian to constitu- and resources for providers (organizations tional rule and more competitive electoral and individuals) to run services, with systems constitutes a major development, important relationships of “management” particularly when seen in historical per- within providers, especially with respect to spective. The spread of democracy has pro- influences on front-line workers’ behavior. vided a large number of citizens, hitherto When such a long route works, it is the excluded from the political decision-mak- classic model of an effective Weberian ing processes, with mechanisms through bureaucracy that is ultimately accountable which they can voice their demands for bet- to citizenry through an elective democ- ter services, though whether government racy. The so-called “short-route” involves will actually respond to these demands is the more direct influence of the ultimate still subject to a range of political and social beneficiaries, when clients directly control factors that often limit actual exercise of cit- how services are delivered or have the izens’ rights. effective capacity to choose from whom Second, governments in the LAC region they get services. have become significantly more decentral- Such a framework helps understand ser- ized in political, administrative, and fiscal vice delivery problems. When governments terms. This is best understood as a particu- do not feel the pressure to respond to citi- lar dimension of democratization. The zen demands, when they are incapable of number of popularly elected subnational enforcing basic performance rules on those governments has expanded throughout the directly responsible for service delivery, and region, achieving what has been called a when citizens have no control or choice “silent revolution” (Campbell 2003). These over service providers, typically services will governments have acquired growing service fail. Unless there is an enabling environ- delivery responsibilities and channel a ment (one that provides opportunities for, larger share of total public spending on key rewards and recognizes “good” behavior, services’ provision. This has also entailed a that supports those that try, that controls growing transfer of revenue authority to abuse of various types, that learns from local levels, in the form of transfers, taxes, mistakes, and has the chance to adapt), or access to debt. individual effort by committed providers Third, there was a process of redefining and users will go to waste. The (public pol- the state’s role that involved increased con- icy) trick is to help establish such an tracting out to private and nongovernment enabling environment. service providers, wholesale privatization (especially in the infrastructure sectors), Latin American Political and more generally, a movement toward and Institutional Context enhancing the role of competition in many Within this type of framework for inter- economic activities, including the provision preting service delivery, it is fundamental to of key services. understand the broader political and insti- Fourth, there has been rising activity by tutional context where delivery mecha- civil society actors, from NGOs to identity- nisms operate. And the last two decades based movements (notably amongst indige- have seen a deep transformation in Latin nous groups), both in the delivery of ser- 1 America and Caribbean (LAC) countries. vices and as actors seeking different First, the region has moved from being a set policies. This has sometimes been associ- of countries where democracy was the ated with efforts by national or local gov- exception, to one in which it is the rule. In ernments to open up government decision- 1980, a majority of LAC countries did not making to citizens and civic organizations. have democratically elected governments. Today, most of these countries are political Interpreting service delivery successes and democracies with national authorities failures in Latin America. If accountability elected in free and competitive elections, relationships between citizens, policy mak- Introduction 5

ers, and service providers are key to effective We also analyze these experiences through service delivery, and there have been both an additional prism that examines the extent systemic reforms (expanding national and to which the poor (and the nonpoor) are local democracy) and an array of specific “empowered” in their relationships with both experimentation (privatization, increased policy makers and service providers— choice), why is the current service delivery whether service-providing organizations or condition still problematic, especially in front-line workers. In essence, we see poor terms of inequality and quality? Chapter 3 people’s empowerment as the product of the systematically explores this issue. interaction between what we term the The general message is that institutional “agency” of the poor (i.e., their capacity to reform is a long-term endeavor, and there exercise choice, to envision alternative are no magic bullets. The political and insti- courses of action, and to act upon their tutional changes noted above are important choices) and the social and political opportu- and are broadly positive for service delivery. nity structure (i.e., the institutional context However, history has bequeathed the region within which poor actors operate) that weak and unequal formal and informal enables the exercise of that agency. This institutions. While conditions vary consid- opportunity structure depends on a variety erably across countries, institutional struc- of factors, including the openness of formal tures are often clientelistic and, especially in or informal institutions, the power and cohe- countries with ethnic and social cleavages, siveness of elite groups, and the state’s socially exclusive. Throughout the region, bureaucratic capacity. The extent of empow- the poor are typically disempowered. erment is fundamental to both the overall Greater formal democracy and decentral- service coverage and quality, whether under ization have opened political opportunities, public, private, or nongovernment auspices, but procedural democracy is only one ele- and the workings of particular institutional ment in the shift to broadly accountable design experiments. governments, and decentralization some- While context is central—and there are times only leads to relocating patronage no magic bullets—it is possible to interpret and the consequent capture of benefits by success and failure patterns. Overall, we find local elites. that “long-route” reforms have brought An important corollary of this analysis is mixed results when used to strengthen both that service delivery’s effectiveness for both policy makers’ responsiveness to citizens poor and nonpoor people is highly depen- and the internal performance of service dent on the institutional context where the delivery mechanisms. Such deep institu- delivery takes place. Similar designs and tional reform is proving to be a much spending levels on delivery of education, longer-term endeavor than was originally health, roads, water, electricity, or other ser- envisioned, especially when the poor have vices can have very different effects depend- historically been not empowered or have ing on how they work in practice, both not been incorporated in local or national across countries and across localities within polities through clientelistic structures. In countries. For example, decentralized deliv- some cases, however—within some social ery with local participation may lead to funds or privatization experiences, for great service quality for the poor in one sit- example—there are important positive uation, but be captured by clientelistic prac- experiences, but this greatly depends on the tices of local elites in another, or rendered political and institutional context. By con- ineffective by unmotivated and underpaid trast there are many positive cases of “short- front-line workers in yet another. When we route” mechanisms that have brought gains turn to specific experiments in service through either increased choice of clients design, we see considerable variety. Privati- (for example, as a product of increased zation can lead to greater service efficiency, competition for telecommunications) or coverage, and quality for the poor in one increased client power (for example, in case, but corruption and increased prices in experiments in greater parental influence another. on schooling). 6 Citizens, Politicians, and Providers: The Latin American Experience with Service Delivery Reform

In chapter 4 we turn to implications for and those in Brazil’s southeast different future policy design and formulation. Draw- from the northeast. ing on the interpretation of past successes Finally there are important conclusions and failures, we argue for approaches that for the policy design process. Integrated seek both to work through the long path within democratic debate, systematic explo- to strengthen the long route of greater ration and evaluation processes are key for accountability to all citizens and strengthen both the “technical” design issues and for the workings of core government services, the broader process of empowering all citi- and continued experimentation with a vari- zens and making decisionmakers more ety of short-route mechanisms. These two accountable. External actors can play a routes are seen as highly complementary. In valuable complementary role, but this both it is fundamental to link speciﬁc tech- needs to be shaped to support internal pol- nical designs with measures to empower the icy experimentation and accountability poor, both through strengthening their processes, by exploring opportunities to agency and improving the institutional support institutional change, strengthen opportunity structure where they live and evaluation and debate processes, and facili- potentially inﬂuence policy makers, organi- tate systematic sharing of experiences. zations, and front-line workers. In designing policies, it is important to shape designs and Notes approaches in light of the country or local 1. Salamon et al. (1999) present a quantitative context—the best changes in Chile may be assessment of the service delivery role that quite different from those in Guatemala, NGOs played in Argentina, Brazil, Colombia, Mexico, and Peru. Service Delivery and Social Outcomes: A Story of Successes and Failures

Latin America has a long history of eco- in very basic services are closed, new ser- nomic, political, and social inequality that vices become a binding constraint for equi- has had a potent influence both on the table development in the region (for exam- chapter question of service delivery for the poor ple, secondary and tertiary education). In and on overall economic and social perfor- addition, more complex challenges, besides mance. The last two decades, though, have coverage, become apparent, the most been a period of change for most countries important being the quality of the services 2 in the region. There has been a dramatically offered. In summary, while there has been different scene in country after country significant progress, there are enough sig- stemming from the overwhelming shift nals suggesting that there is still room for from authoritarian to constitutional rule, substantial improvement; service delivery the movement to more decentralized gov- reform remains an open agenda. ernments, the rising activity by civil society actors outside the state, the growing role of NGOs (both for- and not-for-profit) in Progress in Basic Service delivering a variety of services, and the Coverage increased activity of various LAC countries Primary and secondary education coverage in the world economy. increased significantly between 1980 and During this period, coverage of key ser- 2000 in the region. Considering the 20 vices (education, health, water and sanita- largest countries, average primary enroll- tion, electricity, and telecommunications) ments of the relevant age group went from increased—often substantially—in most 81 to 92 percent, while the corresponding countries in the region. However, progress figures were 40 and 58 percent for sec- has been unequal between countries, ondary education. Most of the countries regions, services, and social groups: large analyzed displayed primary education cov- inequities in access to services remain a fea- erage rates close to or above 90 percent in ture in most LAC countries. The improve- 2000. Despite the noted increases, sec- ments in social outcomes resulting from ondary education coverage remains below service coverage expansion have not been 70 percent in most countries in the region. large and spread out enough to break with In addition, coverage differences between the resilient pattern of inequality that has countries are larger for secondary than for characterized the region. primary education. For example, the mini- Often, recorded service delivery progress mum secondary education coverage in the masks large inefficiencies and deep inade- sample corresponds to Guatemala (26 per- quacies in the service quality and relevance. cent), while the largest corresponds to The noted service coverage expansion was Argentina (79 percent). This gap is almost often achieved by relying on increased pub- three times larger than the corresponding lic spending from a very low base relative to one for primary education. Turning to the region’s income level. Furthermore, the international comparisons, only a couple of “service delivery reform” agenda that LAC Latin American countries underperform countries face can be appropriately charac- (considering their income levels) in terms terized as a moving target; as coverage gaps of primary education, while in the case of 7 8 Citizens, Politicians, and Providers: The Latin American Experience with Service Delivery Reform

secondary schooling, about half the coun- for their income level—immunization rates tries in the region experience gaps relative and professionally attended births. to their income levels. Figure 2.2 shows that there have also Figure 2.1 provides a summary view of been significant increases in coverage for the secondary education case. Three findings basic infrastructure services (water, sanita- are immediately apparent: (i) the absolute tion, electricity, and, in some cases, tele- coverage increases have been remarkable; (ii) phone service). In the decade from 1986 to most of the region still lags behind the mid- 1995, coverage in a sample of 12 countries dle income countries, and the gap remained increased from 76 to 82 percent for the almost unchanged throughout the period water service, 67 to 80 percent for sanita- under study; and (iii) very few countries tion, and 82 to 89 percent for electricity, (Argentina and Chile) come close to the cov- while telephone penetration doubled from erage levels of high-income countries, 5 to 10 percent. However, the figure also including that of the United States. Although shows that coverage is still unequal with the gap closed in the majority of the cases, large differences between households this is due in part to the low relative starting according to their income levels. level observed in the Latin American coun- Importantly, coverage increases have fre- tries in the sample, relative to their income quently been larger for the population’s levels. most disadvantaged segments. This is par- Coverage of health services also increased ticularly true for services where coverage significantly during this period. For exam- rates were relatively high to start (see, for ple, for the same group of countries the per- example, Foster 2003). Figures 2.2 and 2.3 centage of births attended by a trained pro- show these results for water, electricity, and fessional increased by 10 percentage points primary education. Despite this progress, and immunization rates for children up to about half the countries considered showed one year of age by 16 points. By 1999, these coverage rates below those predicted by countries displayed higher than expected— their income level.

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

100

90 1970 1999 80

coverage (%) 40

0 a r y r o a a s e la u o c a a o B ia ic a ru e ia zil a ile a g d li ic u d R b x m e liv ra n ic h tin te m a b g a e a P m o ya a C n o m ra v u R u , m n o B u m e ta c te a l p ta ra c la lo M a c B a S a ic a e s a E e o P in G J rg d in u N o P zu C A h G l S R C e le ite ig E n n d n H a e id U ic V in r m m e o p D p U Source: World Development Indicators (WDI). Service Delivery and Social Outcomes: A Story of Successes and Failures 9

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

Water Electricity 100 99 95 97 90 85 95 80 93 75 91 89 access rates (%) 70 access rates (%) 65 87 60 85 1986 1989 1992 1995 1996 1986 1989 1992 1995 1996 Income 1357 Income 1357 Decile Decile 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.

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

45.0

40.0 Around 1990 35.0 Around 2000

30.0

25.0

20.0 relative gap (%) 15.0

10.0

5.0

0.0 a a a a a ic ay ic la or ay or a eru hile tin m razil d bia livi ras P C n na R zue B exico a gu o am rugu a e M cu B ndu alvad J U rge P sta E ara olom o A o en P C H C V El S 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.

The many successes of LAC countries in haps, the main differentiating factor expanding coverage of basic services over among the region’s social groups. In partic- the last two decades have not always ular, the last decade has seen a sharp resulted in leveling the playing ﬁeld among increase in the returns to higher education different groups in society; many coverage in the majority of the countries. This and quality gaps are reproduced for less development is responsible for a high share basic services, which, increasingly, become of the region’s recent rise in income binding constraints to equal opportunities. inequality.1 However, access to higher edu- The most notable example is in higher cation continues to be highly restricted in education. As basic and secondary educa- most countries and associated directly with tion gaps have diminished in many coun- income status. Furthermore, access gaps tries, higher education has become, per- between the rich and the poor appear to 10 Citizens, Politicians, and Providers: The Latin American Experience with Service Delivery Reform

have increased over the last decade in most services over the period considered. In many countries (see figure 2.4). cases, such efforts were associated with Another increasingly important gap can changes in the way services are delivered. be seen in social protection mechanisms in Chapter 3, below, analyzes those changes. the region. The traditional social insurance But more generally, the service coverage instruments that command the largest share expansion was typically associated with of public budgets in many countries in the changes in the spending level and composi- region have, by design, been associated with tion. For social services, this was largely formal sector employment. As a result, the achieved through a substantial increase in poor, who often lack formal employment, public resources, while for infrastructure have tended to be excluded from the benefits services, public expenditure declined and of pensions, social health insurance, unem- the key mechanisms were private financing ployment, and family allowance systems, and improved cost recovery. among others. Today, the so-called “trun- CEPAL (2001) estimates that the increase cated welfare state” constitutes one of the in total per capita social spending during the key bottlenecks to inclusive development in 1990s was about 50 percent, if pensions are the region (De Ferranti et al. 2004). While included, and 30 percent, if they are not. the 1990s saw an expansion in targeted Public education spending for the 20 largest social assistance programs, these remain countries in the region grew from 3.3 per- small in size relative to the other programs, cent of GDP in 1975 to 4.2 percent in 1999. which appear to be particularly resilient (as The increase in health expenditures is even in the case of higher education) to change.2 more marked. Several countries have large Beyond all the noted limitations and dif- public spending increases (above 50 per- ferences across countries and services, it is cent) in these two services, including fair to say that LAC countries have made Argentina, Bolivia, Brazil, Chile, Colombia, important efforts to expand coverage of key Nicaragua, Paraguay, and Uruguay. In most

Figure 2.4Tertiary Education Enrollment Differences in Selected Countries 8 0.9 Around 1990 Around 2000 0.8 7 Enrollment Q5 Enrollment Q1 0.7 6

0.6 5

0.5 4 0.4

3 enrollment rate (%) 0.3 enrollment Q5/enrollment Q1

2 0.2

1 0.1

0 0 a bia aica Peru Brazil Chile exico Bolivia M icaragua Ecuador Jam Panam Uruguay N Honduras Paraguay Colom ArgentinaVenezuela El Salvador Costa Rica 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 ordered according 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 the latest period (around 2000). Service Delivery and Social Outcomes: A Story of Successes and Failures 11 cases, the expanded coverage of social ser- Service Quality vices was done fundamentally by increasing The evidence on quality is much less system- budgets (with little or no change in the atic than that on coverage, but there are clear financing of some large ticket items in gov- signals that suggest both poor average qual- ernment budgets that continue to benefit ity of services and significant differences relatively well-off groups in society). This among social groups. In education services, suggests the presence of large inefficiencies, for example, the few Latin American coun- which, in some cases, may question the sus- tries participating in international education tainability of the overall efforts. assessments such as PISA (Program for During the 1990s, public expenditures on International Student Assessment) and infrastructure services fell from 1.2 percent TIMSS (Trends in International Mathemat- of GDP to 0.6 percent of GDP. During the ics and Science Study) have shown average mid-1990s, this decline was more than offset low scores (suggesting overall poor educa- by surging private investment. However, tion quality) and, sometimes, high disper- these tended to be concentrated in middle sion (suggesting unequal access to quality income countries and the telecommunica- education). Recent PISA examination results tions and energy sectors. Even at their peak, indicate that in many LAC countries, educa- private capital flows covered barely half of tion quality is lower than what would be estimated infrastructure investment needs expected given their income levels. The Latin and less than one-tenth of needs in the water American Laboratory (LAL) for Assessment and sanitation sector (Fay 2000). Moreover, of Quality in Education’s First International the decline in private capital flows since Comparative Study of Education suggests 1997 has led to an overall reduction in the that only a minority of students achieve sat- resources available to expand these services. isfactory results (for example, read, recognize This is a serious concern, particularly given meanings, and understand and interpret that remaining coverage gaps tend to be information) in most countries in the region concentrated in commercially unattractive and those tend to be concentrated in the rural and peri-urban areas, where it is unre- larger cities; towns and rural areas show a alistic to expect that private initiative will significantly worse performance. Figure 2.5 solve the coverage problem without some presents the reading results; similar results degree of public subsidy. were obtained for mathematics.

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

100 100 93.15 90.83 90 90 82.48 80 75.84 80 70 70 64.45 60 60 54.16 50 48.93 50 44.63 40 40 31.81 30 30 20 20 10 10 0 0 Mega-city Urban Rural Mega-city Urban Rural Mega-city Urban Rural a. Level I b. Level II c. Level III Simple Literal Reading Literal Reading Using Paraphrasing Reading Leading to Inference

Minimum expected percentage Source: Laboratorio Latinoamericano de Evaluación de la Calidad de la Educación (Latin American Laboratory for the Evaluation of Quality of Education). 12 Citizens, Politicians, and Providers: The Latin American Experience with Service Delivery Reform

The data from the Laboratorio Latino- (social outcomes), such as average educa- americano de Evaluación de la Calidad de la tional attainment, literacy rates, life Educación (Latin American Laboratory for expectancies at birth, and child mortality the Evaluation of Quality of Education) rates.3 For example, for the largest 20 coun- show that test results are strongly associated tries, life expectancy increased from 59.7 to with a few school-related factors that can be 70 years between 1970 and 2000, and most seen as more direct quality measures: educa- of the countries show greater than 15 per- tion level of teachers, instruction hours, cent growth in this category. In addition, teacher-student ratios, and degree of school the countries with the lowest initial level autonomy/participation. Unfortunately, few (Bolivia, Ecuador, El Salvador, Honduras, countries collect such indicators systemati- and Peru) show the highest increases, catch- cally. However, there is evidence that Latin ing up with the regional average. By 2000, American countries tend to have yearly the dispersion in this indicator had hours of instruction, which are below the declined. A very similar pattern emerges Organisation for Economic Co-operation when analyzing infant mortality rates. All and Development (OECD) average. Given these countries increased their average edu- the high incidence of teacher absenteeism in cational attainment in a significant manner many LAC countries, the gap in actual class in the 1970–2000 period. It is also notewor- hours is likely to be even larger. Similarly, thy that the countries that experienced the LAC pupil-teacher ratios are above the highest increases were the ones lagging in OECD average. 1970. Thus, in 2000 there is less dispersion Of course, the poor quality problem is in educational and health outcomes in not unique to education. There is scattered Latin America. However, relative to its evidence that quality is low in the provision income level, many Latin American coun- of some infrastructure services in Latin tries are underperforming. In particular, the America when compared to countries with infant mortality levels and rates of access to similar development levels. For example, safe water in many countries are below their waiting time to get a telephone line or the expected levels; in some cases, the differ- number of faults in this service per unit of ences are quite large (see figure 2.6). time in Latin America is more than double Unfortunately, the noted service cover- those observed in East Asia. Relatively high age increases have not been sufficient to water coverage rates often hide serious eliminate the very large inequalities inadequacies in both the potability and ser- observed in many key social indicators. For vice continuity. Inequity in the quality of example, in Mexico, the average person in some infrastructure services, especially the bottom quintile of the income distribu- sewage and telephone service, also seems to tion has 3.5 years of schooling, compared be high (Estache, Foster, and Wodon 2002). with 11.6 years for a person in the top quin- There is anecdotal evidence that the health tile. In Brazil, a child born to a household in services that the poor receive are of much the poorest 20 percent of the distribution is lower quality than those received by the three times as likely to die before reaching population’s better-off segments (see, for the age of five as a child born into a house- example, World Bank 2002b). This is espe- hold in the top fifth. In Bolivia, that ratio is cially true in regard to physician training greater than four, with children in the bot- and experience, the quality of the infra- tom fifth experiencing under-5 mortality structure where health services are offered, rates as high as the South Asian average. A and waiting times to get appointments Guatemalan family whose income is in the (DNP 2002, for the case of Colombia). bottom 20 percent of the distribution has a 57 percent chance of being connected to the Social Outcomes water mains and a 49 percent probability of The trend of pronounced improvements in having access to electricity. The corre- access to basic services has contributed to sponding probabilities for a household in significant progress in the evolution of key the top 20 percent of the distribution are 92 indicators of well-being and capabilities and 93 percent, respectively. Figure 2.7 por- Service Delivery and Social Outcomes: A Story of Successes and Failures 13

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

a. Percent Difference from Predicted Infant Mortality b. Percent Difference from Predicted Life Expectancy 120% 25% 100% 80% 20% 60% 15%

40% 10% 20% 5% 0% 0% –20% –5% –40% –60% –10% –80% –15% 0 0 2000 4000 6000 8000 2000 4000 6000 8000 10000 12000 14000 16000 18000 10000 12000 14000 16000 18000 GDP/capita, PPP adjusted GDP/capita, PPP adjusted

c. Percent Difference from Predicted Adult Literacy d. Percent Difference from Predicted Water Access 40% 40% 30% 30% 20% 20% 10% 10% 0% 0% –10% –10% –20% –20% –30% –30% –40% –40% –50% 0 0 2000 4000 6000 8000 2000 4000 6000 8000 10000 12000 14000 16000 18000 10000 12000 14000 16000 18000 GDP/capita, PPP adjusted GDP/capita, PPP adjusted

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, year 2000. 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.

trays some of the most important inequali- One in four indigenous Ecuadorians never ties in outcomes quite clearly. attended school, compared to only one in Such inequalities are particularly note- 20 nonindigenous Ecuadorians. Compared worthy between indigenous and nonindige- to nonindigenous Ecuadorian children, nous people across Latin America (Hall indigenous Ecuadorian children are twice and Patrinos, forthcoming). Indigenous as likely to be chronically malnourished. In people in Mexico are four times as likely to Mexico’s predominantly indigenous be illiterate than the nonindigenous. In municipalities, the infant mortality rate is Guatemala, the average indigenous adult 41 per 1,000, compared to 24 per 1,000 in has just 2.5 years of schooling, compared to predominantly nonindigenous municipali- 5.7 years of schooling among nonindige- ties. The child mortality rate for Ecuador’s nous adults. In Bolivia, indigenous people indigenous population is double the rate have 5.9 years of schooling, but nonindige- for the nonindigenous population (10.5 nous people have 9.6 years of schooling. and 5.1 per 1,000, respectively). 14 Citizens, Politicians, and Providers: The Latin American Experience with Service Delivery Reform

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

a. Bolivia 1997 b. Percent of stunted children by income quintile 200 Under 5 mortality is 143 150 Poorest 50 quintile Peru 40 100 Nicaragua 30

50 20 Richest 10

quintile short for their age

00percentage of children 0 12345 Poorest 2nd 3rd 4th Richest years of age household income quintile

Source: De Ferranti et al. 2004.

Overall, the evidence reviewed is indica- Notes tive of the growing importance that govern- 1. In fact, secondary education returns remained ments in the region have given to expand- almost unchanged during the 1990s, while those ing social services. However, these trends for tertiary education were on the rise, implying raise many questions about challenges for that the skill premium grew importantly the future. There are a number of countries throughout the region. Santamaría (2003) finds that still show insufficient commitment of that this increase in the wage gap between those public resources to financing key services. with high education levels and the rest of the But overall, further progress is unlikely to labor force explained most of the increase in overall income inequality in Colombia over the happen if governments keep relying only on 1990s. increasing resources and if more effective delivery systems are not developed. This is 2. In 2000, CEPAL (2001) estimated that pension the case particularly when fiscal constraints spending was around 5.3 percent of GDP as an limit the capacity to respond by simply average for the region and that it represented adding more resources. The same is true for more than half of total social spending. When the services for which ensuring quality and rel- region’s 15 largest countries are considered, on average, only 7 percent of that expenditure bene- evance is essential and, thus, deeper changes fits the poorest 40 percent of the population. may be needed in the type of service provided. It is in this context that we stress the 3. Of course, many other (macro, micro, and importance of assessing how well today’s structural) factors have contributed to the service delivery systems are positioned to observed progress in those outcomes. For exam- address the many remaining challenges fac- ple, child mortality reductions tend to be associ- ing LAC countries, particularly in terms of ated not only with health services, (for example, distribution of hydration tools to poor house- serving the poor. The next chapter will pro- holds), but also with improved parental educa- vide an assessment of the changes intro- tion, increased coverage of water and sanitation duced in LAC service delivery systems over services, higher incomes, and reduced fertility. the last two decades. LAC Service Delivery: Assessing Two Decades of Change

The last two decades have seen some radical assessment, we have made an effort to rec- transformations in state organization and ognize and identify such gaps in the two structure and in the relationship between the decades of change in service delivery chapter state, its citizens, and markets. The spread of approaches in Latin America. democracy, state decentralization, a redefined state role, and increased activity by non- Reforming the Compact governmental actors represent important Latin America has long been characterized by 3 transformations that create opportunities a general tendency for its states to be “weak” for change in accountability relationships and “unequal,” specifically when delivering between policy makers, provider organiza- key public goods (especially in terms of tions, front-line providers, and citizens. macroeconomic stability and universal guar- This chapter seeks to explore the influence antees of property rights and of citizenship) that these various changes have had on the and provisioning quality basic economic and delivery of services to poor people in Latin social services to all (De Ferranti et al. 2004). America. We review the evidence on how Where public goods are inadequately pro- key accountability relationships were trans- vided, wealthier groups either resort to pri- formed during this period and the influence vate mechanisms (for example, with private such transformations had on service delivery. security or holding savings abroad to deal First, we look at changes in the relationships with macroeconomic instability) or actually between policy makers and public and pri- benefit from weaknesses (for example, by vate service providers. Next, we review influencing justice systems through private changes in the relationships between clients influence). Where basic goods are underpro- and providers that have sought to increase vided, it is typically the poor who are served “client power.” Finally, we consider changes last and where quality is low, richer groups in the relationships between citizens and pol- have greater means to opt out in favor of pri- icy makers and their impact on overall state vately provided services. To a great extent, accountability. In doing so, we will seek to such weakness has been associated with badly highlight the many connections between functioning state bureaucracies in charge of these accountability relationships and iden- service delivery—even though the underly- tify remaining bottlenecks to improve service ing causes of such deficient functioning have delivery systems in Latin America. typically been linked to the social and politi- This chapter draws on available studies cal structures within which such bureaucra- and assessments of specific country and cies operate. Thus, a key aspect to improving sector service delivery experiences. The service delivery is associated with the incen- growing emphasis on conducting system- tives and capacity of the policy makers to atic evaluations of policies and programs establish effective compacts—broad, long- among both academics and development term accountability relationships—with the practitioners provides us with a good organizations (state and nonstate) in charge empirical basis for our review. Nevertheless, of service provision. important knowledge gaps remain (for The last two decades have seen many some types of services, groups of countries, attempts at reforming the state’s internal and service delivery approaches). In our organization, its structure, and its roles in 15 16 Citizens, Politicians, and Providers: The Latin American Experience with Service Delivery Reform

Latin America. Some of these attempts had their structures and administrative systems1 the specific purpose of improving state ser- and introduce new public management vice delivery effectiveness, while others approaches. Many countries significantly responded to different policy goals (such as improved capacities for core economic the state decentralization process). In this management (see box 3.1). However, the section we review the most important of effects of the modernization efforts on these reforms, searching for empirical evi- improving service delivery have been lim- dence on whether and how they improved ited to date. Implementing fundamental service delivery. First we discuss the efforts reforms of government bureaucracies to to reform government bureaucracies. Next, make them more results-oriented has we explore the effects of administrative proven to be a long, arduous path rife with decentralization that is part of the multifac- political and technical difficulties. eted state decentralization process that LAC Initial public sector reform efforts countries have been implementing. Finally, focused nearly exclusively on fiscal adjust- we look at a broad range of initiatives that ment objectives and the broad redefinition innovated service delivery by establishing of the state’s roles. Attention gradually stronger, more performance-based com- shifted to modernization of core govern- pacts with actors outside the mainstream ment functions, such as macroeconomic state bureaucracies. planning, financial management, and tax administration. By the end of the 1990s, LAC government bureaucracies state institutional reform was squarely on remain generally weak, ineffective, the region’s development agenda. and largely unaccountable for results These reform efforts tended to empha- A mix of reforms emerged in the region size improvements in internal processes and over the last two decades, whereby govern- instruments. Aided with technological ments have sought to modernize and adapt advances in the 1990s, integrated financial management systems have become popu- larized as potent instruments for strengthening government financial management BOX 3.1 Macro vs. Sectoral Reforms (especially expenditure control during the budget execution stage and ex post report- Centralized approaches to strengthening By contrast, large service delivery core economic management functions domains are intrinsically transaction-inten- ing). Integrated financial management sys- have been relatively effective, and many sive and involve discretionary decisions by tem reforms are sometimes seen as building countries now boast professional central front-line bureaucrats that are difficult to blocks or as a catalyst for broader institu- banks and finance ministries aided by com- monitor and control centrally. It is this com- tional reforms of the public expenditure petent technocrats and information tech- bination that makes it particularly prone to nology (IT)-based systems to better man- clientelistic abuse and thus challenging to management systems; better information age public monies.Why was this reform reform (World Bank 2003a; Pritchett and can contribute to developing better institu- apparently more successful than others, Woolcock 2002).Widespread citizen discon- tions that facilitate better decisions and especially those in service delivery? We sug- tent with government performance has eventually, better outcomes. gest two reasons why this domain was dif- somehow failed to translate into a strong ferent from that of most front-line service drive for results-oriented public sector The introduction of these improvements delivery programs: reforms.This contrasts with the evolution of in many countries has been among the most reform efforts in OECD countries, where fis- 1. First, the huge costs of macroeconomic successful reform efforts in the region’s pub- cal pressures led not only to the general mismanagement in the 1980s shaped lic administration realm in the 1990s. As seen movement toward tighter fiscal policy and political will and practical learning to expenditure control (Atkinson and van den in Guatemala’s reform (see box 3.2), it is pos- make changes in core government Noord 2001), but also to a greater expendi- management functions. sible to use such systems in a way to derive ture efficiency focus (for example, an immediate service improvement impacts 2. Second, macro policy is an area that can emphasis on “value for money”) and conse- be effectively run by small cadres at the quently to deep public sector management (Dorotinsky and Matsuda 2001). Overall, the national level. Moreover, once the pop- reforms.We might then speculate that contribution of the region’s financial man- ulist macro policy’s failures as a redistri- either the public sector bureaucracy’s poli- agement reforms has been to establish better bution instrument became manifest, it tics or culture has to change for fiscal pres- became an area that offered relatively input control as a fundamental building sure to translate into concerted efforts to 2 few patronage opportunities and thus improve service delivery institutions. block of good public administration. is more easily delegated to technocrats. LAC countries have also begun developing the technical foundations for performance- LAC Service Delivery: Assessing Two Decades of Change 17

BOX 3.2 Public Financial Management Reform in Guatemala

At one point or another, most countries in the • By decentralizing purchases through the sites in Guatemala and through the Internet region have attempted to reform their financial IFMS, control of the procurement processes (www.siafsag.gob.gt). management systems (normally budgeting, was improved and arrears with suppliers • Physical and financial indicators are now treasury, accounting, and debt management) to were eliminated.This reduced the cost of cer- included in the budget to better measure make public expenditure more effective and tain wholesale purchases for hospitals and performance and promote a result-oriented efficient.The Guatemala Integrated Financial schools by 10 to 70 percent. focus, and line ministries have timely and Management System (IFMS) stands out as one • As part of the IMFS deployment, payments accurate information to improve physical proof the most successful. Emerging from three are electronic and checks are no longer gramming and service provision. decades of civil war, Guatemala had a desperate issued to suppliers, with the exception of • At the municipal level,the project is also imple- need not only to improve service delivery, but teachers’salaries. Once these begin to be paid menting IFMS platforms that share the national also to restore confidence in the public sector’s electronically, an estimated savings of $2.5 system’s functionality at a smaller scale.These integrity and efficiency. million per year will be achieved simply by systems will also generate substantial savings Some examples that illustrate Guatemala’s eliminating the use of physical checks. that can be directed to improving municipal success and its potential impact on the cover- • As a complement to the IFMS, the project is government service delivery. age of public sector services are: deploying an online procurement system By themselves, back-office systems such as • Elimination of the “floating debt”that arose (“Guatecompras”), which has been widely the Guatemala IFMS cannot directly affect the from inadequate budgetary controls, which praised both inside and outside the country. quality of final public service delivery; however, used to lead the Ministry of Finance to devote It is expected that in the medium run, up to a they can help increase coverage by reducing virtually the full first quarter of the year to 20 percent savings in public sector purchases input costs, be it through improved paying commitments from the previous year. will be achieved through increased procurement practices, speedier payments to It turned out that many of these transparency and more effective competition. suppliers, or reduced corruption opportunities. commitments were not legitimate. Close to • Legislators, private citizens, NGOs, and other In addition, they provide real-time financial $100 million was saved through the controls stakeholders can access proposed and exe- information that can be critical for making pol- implemented as part of the IFMS. cuted national budgets at several computer icy decisions in all sectors.

based budgeting and management with the civil service systems has proven politically use of performance indicators and the intro- difficult and technically intractable. Most duction of program evaluations. Several countries in the region have yet to estab- countries, including Bolivia, Brazil, Chile, lish professional “Weberian” bureaucra- Colombia, Mexico, and Peru, have embarked cies, which many analysts see as a neces- on efforts to develop a results-oriented man- sary precursor to introducing modern agement framework even though, with the managerial approaches (Shepherd 2003). possible exception of Chile, no country has Far from introducing performance-ori- yet developed a fully functioning perfor- ented personnel management practices as mance management system. Often these a necessary complement to strengthening efforts have suffered wavering political com- the public sector’s service delivery capac- mitments by different governments. In some ity, several countries in the region are still cases, the work done so far appears to be struggling to keep the aggregate wage bill mostly technocratic exercises with limited under control and lack sufficient capacities links to actual government decision-making to control deployment and attendance, let and management processes. alone performance, of their public ser- Arguably the most important public vants. In very few cases, this has reached sector reform from the viewpoint of the level of failed states (see box 3.3 ). improving service delivery, especially in Aside from the technical and fiscal com- the labor-intensive social sectors, is per- plexities (for example, how to resolve the sonnel management reform. After the well-known dilemma of controlling the emphasis on public sector retrenchment in aggregate wage bill for sound fiscal policy the fiscal adjustment context, the second on the one hand and paying individual civil half of the 1990s saw the emergence of civil servants competitive salaries on the other), service professionalization, or the estab- civil service reforms are riddled with politi- lishment of a merit-based civil service, as a cal difficulties. Introduction of a merit- commonly-agreed-on priority in a num- based civil service is difficult because stake- ber of countries. Yet deep reform of the holders in the political system actively resist 18 Citizens, Politicians, and Providers: The Latin American Experience with Service Delivery Reform

agement and to introduce greater perfor- BOX 3.3 Haiti: Public Service Delivery in a Failed State mance orientation. However, the reform has stalled partly due to strong union pres- Haiti’s characterization as a failed state, or rity force, they will be dependent on donor one that has been failing over decades, resources to finance anything beyond their sure (Rinne 2001). In Chile, union opposi- does not waive the rules of improving pub- operating budget and possibly even a fair tion has been cited as one of the reasons for lic service provision. However, it does make share of that. In this situation, the two high- relatively limited progress in civil service their application far more difficult.The est priorities are to regain control over reform aimed also at introducing greater state’s service provision role is minimal (for expenditures through the introduction of example, private schools provide over 80 an improved budgetary process and to find flexibility in the public sector personnel percent of education), ministries manage ways to run or oversee service delivery and management regime (Egaña 2003). much of their budgets through “current other basic government functions with the Union opposition to performance- accounts”for which there is no accountabil- depleted human resource base. Here enhancing reforms tends to be particularly ity, and public office has been used to donors face a common dilemma in an reward partisan loyalists. Over the past extreme form—whether to emphasize exe- problematic in social sectors. For example, decade, the shrinking group of donors still cution of their own programs through teachers unions typically negotiate for job operating in the country looked whatever means possible (such as by using security (as well as, obviously, for better increasingly to NGOs to implement their NGO executors, hiring the most qualified wages) and against results-based quality programs, or they managed their own Haitians to backstop their efforts, or finding 3 funds, or if they channeled monies through ways to short-circuit less-efficient govern- assessment systems and local control. In the government, often short-circuited the ment procedures) or to sacrifice some countries ranging from Bolivia, Costa Rica, highly inefficient “normal”processes. Not advances there in the interests of building the Dominican Republic, Peru and Vene- surprisingly, public sector human resource public sector capacity. Should they decide quality, quantity, and distribution were to focus some efforts on capacity building, zuela, just to name a few examples, unions totally inconsistent with service provision they will have to tailor them to the realities opposed reform measures to strengthen needs. Most staff was located in Port-au- of existing organizational and human meritocracy and teachers’ accountability Prince; a United Nations Development Pro- resource capabilities. New procedures will for learning outcomes or to introduce pri- gramme survey indicated that 68 percent have to be simple, but effective, less reliant had only a primary or secondary education, on technology than in other national set- vate provisions so as to generate competi- and even offices with technically qualified tings, and capable of use by the “average” tive pressure for public providers and offer personnel had too few of them to carry out employee, not just the upper level staff. A “exit” options to parents (Gray Molina, their mandates.The near civil war leading to mix of implementation modalities will be Pérez de Rada, and Yeñez 1999; Maceira President Aristide’s departure was accom- inevitable, including the use of NGOs and panied by the exodus of many of the more the private sector, but they should strive for and Murillo 2001). It is also true that qualified government staff, leaving an enor- an integrated rather than an either-or teacher unions—or at least factions within mous gap at the upper and middle- approach. Salary subsidies or similar unions—have sometimes been reform advo- management levels. Salaries and operating arrangements to attract more qualified staff cates and governments have worked with budgets, while never adequate, are now at should be used with caution as the few an all-time low. benefiting from them will not be sufficient them for quality and efficiency improving Assuming that the current transitional to run the government, and the more basic reforms, for example, in Minas Gerais government and an eventual elected one problem will be how to provide incentives (Grindle and Mason 2004).4 When explicitly can maintain an adequate level of civil for the ordinary state workers. order, aided by the presence of a UN secu- involved in the reform process, responsible unions can play a positive role in defining the content of performance-oriented teacher service reforms and facilitate their imple- it. Civil service unions are often opposed to mentation, as was the case in Chile over the change initiatives that are designed to 1990s (Crouch 2004).5 improve their performance incentives and Besides the frequent opposition from thus also affect their entrenched (and usu- public sector unions, the most obvious cul- ally legally sanctioned) privileges and bene- prits for lack of civil service reform are the fits. Unions typically oppose these mea- continued prevalence of clientelism and the sures, especially those geared toward fiscal politicization of public administrations. In adjustment purposes (which result in job much of Latin America, provision of public losses), better internal controls (which jobs has been a major instrument both for result in reduced room for opportunistic achieving or maintaining national level sup- behavior), and managerial flexibility (which port and/or an instrument for the clientelis- can result in both job losses and greater per- tic exchange of local level favors. This prob- formance exigency). In Brazil, for example, lem has been diagnosed in Bolivia, for the 1998 administrative reform was intro- example, where the peculiarity of the politi- duced with the aim of introducing greater cal system creates permanent and strong flexibility to public sector personnel man- incentives for political parties to divvy up LAC Service Delivery: Assessing Two Decades of Change 19

the public sector as a currency of political personnel inculcated in the ethos of rule- exchange to maintain minimum conditions bound integrity (for example, teachers who for governing (World Bank 2000b). Bolivia’s show up at classrooms every day to teach) case may be extreme in some sense, but it is or of personnel management flexibility to hardly unique. In Mexico, with a more take advantage of incentives for improved developed pool of professional human organizational performance. The latter type resources, politicization has reigned in the of problem is well illustrated by the difficul- administration of public sector personnel ties that public health sector reforms faced (Garibaldi n.d.). It is also seen in the only in seeking to improve the incentive frame- recently changed tradition of use of teaching work for efficiency and responsiveness (see jobs as a means of patronage in Colombia box 3.4). (Angell et al. 2001). Public utilities are a fur- Overall, the evidence suggests that the ther example of a patronage system of job historical weaknesses of Latin American provision and cheap, if low quality, services state bureaucracies still limit the effective- (Foster 2003). Overall, in a context where ness of conventional public sector delivery. job granting remains a preferred mecha- State modernization reforms have upgraded nism to reward political clienteles, there is a governments’ institutional capacities con- serious challenge to design effective incen- siderably—at least in some areas of public tive systems that motivate front-line workers administration systems—and it is likely that involved in service delivery.6 Furthermore, such upgrading will have substantial effects often union resistance to performance- over the longer term. But very few countries enhancing reforms is a defensive move against such practices. Effective social service delivery depends critically on the quality and dedication of BOX 3.4 Lack of Civil Service Reform as a Health Reform teachers, doctors, and nurses. Yet, organized resistance by unions and informality sus- Bottleneck tained by patronage and clientelism severely A main thrust of LAC health sector reforms limited in scope and effects (World Bank limit reformers’ options for using explicit has been to organizationally separate the 2001b;World Bank 2002b; Over and Watan- incentive mechanisms for enhancing their three key health system’s functions: regula- abe 2003;World Bank 2003c). tion, financing, and service provision. In par- For these reforms to be successful in performance orientations (for example, ticular, purchaser-provider split reforms enabling public sector performance change performance-based pay). In addition, in have been introduced in many LAC coun- requires significant flexibility in resource most countries in the region, paying the tries with the objective of improving the management and allocation by service large number of social sector workers ade- public health sector’s efficiency and respon- providers. In other words, reform implemen- siveness incentive framework. Under inte- tation requires that public providers be quately remains a significant challenge grated systems, the central authority has increasingly able to adapt their service pro- given national fiscal realities. A long-term the simultaneous missions of maximizing duction functions and cost structures to the solution seems to depend on somehow the impact of its resources on the popula- continuous evolution of price signals deter- encouraging greater professionalization of tion, of financing the system, and of ensur- mined by the new payment mechanisms. ing financial sustainability of its own There is increasing evidence of the difficul- the social sector workers (especially teach- providers. Historically, that combination of ties in implementing the purchaser- ers). But in the short run, it may be possible missions created incentives for the state to provider split (Baeza 1996). For example, the to find promising alternatives in greater focus on ensuring a stable revenue flow for slow pace of the reforms in Costa Rica and its own providers, rather than on examining Chile—two countries with relatively effec- reliance on “social control” mechanisms, the impact on the population. Chile tive public administrations—is due to a such as community vigilance of family provides the earliest case of these reforms great extent to the management rigidity in health workers in the Brazilian state of with the creation of the National Health all production factors at the provider level, Ceará (Tendler 1997) or school director Fund (FONASA) in 1981. After significant particularly in the lack of flexibility in man- legal, managerial, and financial reforms, aging human resources. Indeed, elections as practiced in several Brazilian there is almost complete separation of pur- international experience (England 1998) states. chasing and provision in Chile’s public increasingly suggests that purchaser- The consequences of the lack of civil ser- health sector. Colombia is the other para- provider split reforms are facing serious vice reform go beyond continued general digmatic case; in the mid-1990s, it initiated constraints due to the difficulties in reform- an ambitious health insurance reform that ing civil servant regulations in the public weaknesses in public administration. They implied a clear and nationwide separation health sector and the resulting rigidity in can also undermine institutional innova- of functions. Efforts at establishing public managing production factors and modify- tions in specific sector reforms by depriving hospital autonomy were widespread, but ing cost structures by public providers. the public sector of technically competent 20 Citizens, Politicians, and Providers: The Latin American Experience with Service Delivery Reform

have been able to move decisively toward goal of improving the service delivery effi- developing a results-oriented public admin- ciency. More typically, decentralization of istration or achieve significant depoliticiza- service delivery responsibilities was largely tion of the state apparatus. We interpret this determined by the design and outcomes of as being due to a combination of the intrin- both political and fiscal decentralization sically long-term nature of institutional processes. In all cases, however, the decen- change and the persistence of countervail- tralization of service delivery responsibili- ing forces from groups, from patronage- ties to local governments9 has resulted in a oriented political groups to unions, that reform of the compact between govern- benefit from existing arrangements. The ments and service providers whereby the implication is, however, that service deliv- latter are increasingly under the supervision ery quality by state agencies remains a seri- of subnational governments, as well as new ous bottleneck, particularly for those ser- compact relationships between the national vices (for example, education and health) government that still sets overall sectoral for which personnel practices are crucial policies and the subnational governments and in countries where resistance by inter- charged with service delivery. est groups is the strongest. Whether these changes result in better services depends crucially on local bureaucracies being able to function more effectively than Service delivery function the national ones. The key is for local govern- decentralization was attempted in ments to have the combination of resources, country after country with authority, and incentives needed to do so heterogeneous results (World Bank 1999).This can also be seen Over the last two decades, most countries in from the perspective of the tradeoff between the region have gone through a process of increased information at lower government decentralizing service delivery responsibili- levels and the influence of local power struc- ties. Practically all countries have under- tures and practices (Bardhan and Mookher- taken some form of decentralization of jee 2000a, 2000b). In the Latin American con- their primary and secondary education sys- text, there is a strong history of the power of tem, which involved the transfer of deci- local strongmen and of the importance of sion-making powers to actors within or, unequal local social structures. However, more likely, outside the ministry of educa- there is also a significant variety of experi- tion’s bureaucratic structure. In many ences, including in terms of the democratiza- countries, the water industry was trans- tion and decentralization responses. The evi- formed from a single national provider dence essentially points in the direction of under the central government’s control, to a mixed results: decentralization’s impact on highly decentralized structure with service service delivery varies significantly between being delegated to hundreds of new munic- countries and sectors or geographic areas ipal utilities.7 Local and regional govern- within countries. Critical to explaining the ments were also charged with growing results are local political conditions and the responsibilities for public health facilities country- and sector-specific policy frame- and programs. Decentralization also took works under which decentralized service place in the transport sector, with munici- delivery operates. palities and regional governments becoming increasingly involved in tertiary and Design issues. In many cases, the overall secondary road management. Over time, frameworks under which responsibilities some countries also delegated responsibili- were decentralized in most countries and ties for social assistance programs to subna- for most services have created overlaps and tional governments. gaps, inadequate or insufficient authorities These reassignments of responsibilities and resources, and few (enforceable) mech- followed many paths and approaches,8 but anisms to promote local government per- in few, if any, cases was administrative formance orientation. Most countries had decentralization driven primarily by the many difficulties in establishing well func- LAC Service Delivery: Assessing Two Decades of Change 21

tioning compacts between government lev- Fiscal transfers remain the principal els to ensure the necessary incentives for financing source for a large number of local effective service delivery to the poor.10 governments, creating the potential for In many cases, blurred responsibilities strong accountability lines toward the cen- have both made the tasks of newly empow- ter. In reality, though, the establishment of ered local governments complex as well as clear, transparent compacts linked to service created too many opportunities for individ- delivery outcomes is the exception rather ual governments to essentially avoid than the rule. In Argentina, for example, the becoming accountable for results vis-à-vis absence of a clear “compact” between the their citizens. In the case of health services, national level and the provinces (that is, lit- for example, in most cases decisions on tle or no specification of targets and objec- human resources (for example, salaries and tives and the use of general transfers that civil service rules) remained centralized or are not tied to any particular efficiency in the hands of regional governments indicator) implied that decentralization’s imposing restrictions on municipal man- effects on education quality and coverage agement autonomy (Bossert 2000). In those vary a lot across provinces, depending on countries (such as Bolivia and Colombia) provincial management (Galiani and Schar- that distributed responsibilities for educa- grodsky 2002). This raises both efficiency tion services among three government lev- and equity concerns. For water and sanita- els, room was created for strategic behavior tion, notwithstanding extensive decentral- for each government level to seek to expand ization, central government often retains a their power or budgets, and inconsistencies role in financing services, whether through in the assignment of responsibilities and budget transfers or credit lines for major resources made it extremely difficult for investments. This financing relationship local actors to “manage” schools in practice. potentially provides an additional mecha- Similarly, the arrangement to have national nism for increasing the accountability of regulatory agencies oversee the perfor- local service providers, since access to mance of subnational water utilities (for resources could potentially be conditioned example, in Bolivia, Colombia, and Peru) on absolute performance thresholds or has not functioned very effectively due to ranking against utilities in other jurisdic- differing sector policy goals between tions. However, it is striking how little use national and subnational governments, and has been made of this instrument in prac- even conflicting jurisdictions that under- tice, with the majority of municipalities in mine the regulatory agency’s authority.11 the region receiving fiscal transfers without Superimposing decentralization over other any accountability requirement against the sector reforms is another source of inconsis- resources spent.12 tency. For example, many countries now have Nevertheless, there are good examples of legal contradictions between policies pro- arrangements where strong intergovern- moting community-based approaches to mental compacts coexist with large fiscal rural water and sanitation services and transfers. In Chile, for example, the educa- decentralization laws. This is the case, for tion delivery responsibility was transferred example, in Honduras, where a recently to municipalities through specific agree- passed water sector law provides legal ments that defined explicitly the responsi- recognition and a mandate to community bilities, rights, and obligations of the new water committees to provide services, while providers and linked central resources to the the decentralization law gives municipal number of students attending class. As illus- governments the same responsibility. As a trated by the case of Brazil’s FUNDEF, an result, paradoxically, rural water and sanita- innovatory financing mechanism that col- tion decentralization may actually result in lects resources from state and municipal centralization of decision-making power by governments in a single fund and redistrib- transferring legal responsibility and author- utes them according to the amount of stu- ity away from the hands of rural communi- dents enrolled, thus ensuring spending min- ties to local governments. imum levels in all states and municipalities, 22 Citizens, Politicians, and Providers: The Latin American Experience with Service Delivery Reform

when the compact between the center and factors contributed to this success: local subnational governments is strengthened, electoral competition as motivation to there is great potential to generate improve- pursue reform; the fact that education ments in educational achievement out- reforms included both private and public comes and enrollment, in particular in poor schools; close parental oversight; control areas (World Bank 2003b). over human resources on the municipality’s behalf; and a new incentive frame- Local responses. While the characteristics work that the central government estab- of the overall fiscal, administrative, and lished (capitation grants instead of political frameworks under which decen- cost-driven central transfers). All these fac- tralization takes place are clearly influential tors combined provided local politicians on service delivery opportunities for local with both the incentives and the ability to decisions and actions, it is increasingly move toward results-based management. apparent that localities respond differently In other words, changes along the various to such frameworks, emphasizing the role accountability relations (that is, not only of local political systems and local social or even primarily in the intergovernmental organization’s characteristics and strengths. compact) made the observed results possi- The experience across municipalities in ble—a recurrent theme in our analysis. Bolivia is of interest here, particularly since The experience of LAC countries sug- they appear to be characterized by relatively gests that effective decentralization oppor- weak state capabilities and embedded tunities systematically differ between types patronage politics (World Bank 2000b). of services. In contrast to the education sec- Popular participation was intended as a tor’s mixed cross-country and in-country national reform to shift resources and con- evidence, there is surprisingly little evidence trol to municipalities and legalize local level that local elections have created any signifi- civil society groups, of which campesinos’ cant pressure to improve municipal water syndicates were the most important in services. Only a minority of small and many parts of the country. While the medium-size cities in Latin America have changes brought benefits, on average, undergone major reforms of their water (Faguet 2000) the dynamics of change utilities.13 It is only a matter of speculation depended on the variable local social con- as to why local level political incentives text. In some parts of the country, notably implicitly appear to be less in favor of in some of the Cochabamba valleys, where reforms in the water than in the education there was a relatively strong social mobiliza- sector. What is clear, though, is that asym- tion tradition (that is, greater agency capac- metric decentralization results by sector are ities of the poor), the legal changes in the related to the influence of size. In the water direction of greater openness appear to be sector, a serious but unanticipated decen- bringing initial benefits in the form of bet- tralization consequence has been the loss of ter public action. In the traditional sierra scale economies in service provision, as well city of Sucre, by contrast, a traditionally as the dispersion of scarce human resources more closed political culture and a stronger with the capacity to manage and maintain elite may have caused the decentralizing water systems. In commercial terms, the reforms to primarily increase the local client base for many smaller municipalities elite’s power (Gray Molina 2002; Calderon is entirely composed of low-income resi- and Szmuckler 2004; Blackburn 2001). dential customers, making it difficult to But beyond these broad, albeit power- achieve financial viability. This type of scale ful, structural forces, it is the combination factor appears to be of much less impor- of many factors that appears to be behind tance in the case of education, for example. successful local service delivery reform. Nevertheless, scale may also be a limiting For example, Bogotá city engaged in a factor for some aspects of education service comprehensive reform program of the provision, such as curriculum design or local education system, improving learn- evaluation systems, which tend to remain ing and teaching outcomes. A variety of the central government’s responsibility. LAC Service Delivery: Assessing Two Decades of Change 23

More importantly, while small scale may a discussion of this approach’s limits on not be a serious bottleneck at any particular improving service delivery. point in time, it may reduce flexibility to react to changing conditions over time: for Contracting out service delivery. Contract- example, by limiting the capacity of system ing out services to private operators, administrators to change resource alloca- whether through long-term concessions or tion decisions in response to demographic outright asset sales, is most advanced in the shifts. Of course, these challenges do not infrastructure sectors. Prior to 1980, utilities necessarily call for centralization. Rather, in most Latin American countries were they emphasize the importance of arrange- organized as publicly owned and operated ments for intermunicipal cooperation. national (or state level) monopolies. Politi- It is probably fair to say that the fruits of cians exerted their control over the sector decentralization are still not fully harvested. via the appointment (and dismissal) of util- In some cases (countries and sectors), some ity managers, and by providing subsidies to important changes (for example, in financ- support unsustainable enterprises. In ing arrangements to improve accountability return, utilities provided jobs, kept tariffs for results) are called for as a precondition down, and allocated investments and public to create a better enabling environment for works contracts on a political criteria basis. local delivery of services, and several coun- The consequences of this regime were spi- tries are already moving in that direction, raling costs, low service quality, and a short- Colombia being the most recent example. age of resources to fund badly needed ser- In other cases, more time may be needed to vice coverage expansions. From the early allow local governments and local politics 1990s onwards, there was a widespread to play out as reformers hoped. In all cases, movement to try and replace the “clientilis- however, central governments are faced tic” service provision model. The reforms with the need to provide incentives and stressed the need to provide a clear institu- implement mechanisms to promote and tional separation between the roles of policy facilitate needed actions on the part of sub- maker, regulator, and service provider. Ser- national entities: central governments have vice provision would be delegated to the pri- a central role (monitoring results, providing vate sector to ensure that services were run incentives, and even intervening when ser- on commercial principles, and providers vice delivery fails) in making decentraliza- were free to pursue service delivery objec- tion work for service delivery outcomes. tives. In other words, the contracting out provision to private firms implied that in the case of utilities, many bottlenecks in the Establishing compacts with agencies health sector from inadequate progress in or actors outside the mainstream implementing civil service reforms (see box state apparatus has overall 3.4) would not materialize. produced good (short-run) results The reforms implied a substantial LAC governments have sought alternative increase in private sector participation. A and often innovative ways to avoid the deep recent survey of private sector participation difficulties of reforming bureaucracies in Latin American infrastructure during the (central and local) to improve service 1990s found that there had been more than delivery outcomes. Many of these alterna- 200 concessions for water, electricity distrib- tive ways involve a combination of con- ution, and basic telephony services (Guasch tracting out services to private firms and 2004). However, the scope for private partic- NGOs and relying on alternative bureau- ipation proved to vary substantially across cratic structures that bypass more tradi- sectors; by the year 2000, over 90 percent of tional hard-to-reform ones. Overall, these telephone customers, 50 percent of electric- efforts produced positive results in terms of ity customers, and 15 percent of water cus- improving service delivery outcomes. tomers in urban areas in Latin America Below is a brief review of the assessments received their service from private operators. available for some experiences, followed by There is substantial evidence that private 24 Citizens, Politicians, and Providers: The Latin American Experience with Service Delivery Reform

sector participation has led to significant receive remuneration from the department improvements in the performance of utili- on a per-student basis. While the contract ties, particularly as regards connections, establishes clear standards that must be labor productivity, and service continuity.14 upheld, the concessionaire has full auton- The increased coverage of electricity, water, omy over school management and is evalu- and sewerage services primarily benefited ated solely on results. Concessionaries have lower income groups, while the surge in already produced striking management telephone penetration primarily benefited improvement results: they allocate, on aver- the middle classes (McKenzie and age, 55 percent of the per capita remunera- Mookherjee 2002). There is similar positive tion to human resources, well below the 90 evidence regarding water sector coverage.15 percent allocation by public schools. While coverage and efficiency effects were More common is the practice of con- generally favorable, price effects were more tracting out the delivery of social services varied, often depending on initial condi- to NGOs. In their efforts to improve effi- tions (De Ferranti et al. 2004). ciency and reduce costs, a growing number An essential aspect of these reforms was of ministries of health in Central America the creation of a regulatory agency for utili- began to purchase primary health care ser- ties (with the key function of insulating services from NGOs—usually private, non- vice providers from political interference, so profit entities—to provide coverage in that decisions can be taken on the basis of delimited geographic areas (Fiedler 2002). long-term public interest). At present, Latin Through a formal, competitive process, El America has more than 30 national level Salvador, Honduras, Nicaragua, and regulatory agencies for the water, electricity, Panama awarded contracts to NGOs to and telecommunications sectors. Regula- provide coverage to remote (often indige- tors for subnational jurisdictions have also nous) populations with little or no access sprung up in federal countries such as to care, in return for a fixed, annual capita- Argentina and Brazil, and for the water section payment. In most cases, the NGOs tor in countries such as Ecuador and Hon- provide care using a network of itinerant duras. In many cases, sectoral legislation teams, usually complemented with com- incorporates several mechanisms designed munity-based paramedics. The NGOs are to safeguard regulatory autonomy, in par- generally required to visit designated vil- ticular by establishing clear criteria for reg- lages or to provide care at specified delivery ulator selection and removal16 and making sites a minimum number of times per year. regulators financially self-sufficient via sec- These innovations started out as rela- tor levies. tively small-scale experiments. In some Similar, large-scale experiences of con- cases—Nicaragua, for instance—they have tracting out private concessions are less not been regarded as successful, and their common for social services. The Bogotá scale and financing were reduced. In other school concessions program (which created cases, both the clientele and authorities 18,000 new school places) is an important enthusiastically endorsed this approach, example of how private management of and its scope has gradually grown. In public schools could work if used more Panama, for example, NGO provider teams extensively. The municipal education now cover about 95,000 people, and current department constructs new, state-of-the-art plans call for expanding the program to schools in low-income areas that are each nearly 10 percent of the population in the able to accommodate between 800 and next few years. With the longest history of 1,200 students. A public procurement contracting out to NGOs among Central process offers the opportunity to manage American countries and of purchasing a the school; bidders (good-quality private greater variety of services than its neigh- educational institutions) are evaluated on bors, Guatemala has contracts with approx- their proposed management plans. Conces- imately 90 NGOs that provide services to sionaries are obliged to provide educational more than three million people, about 30 services to poor children and in return percent of the national population. Impact LAC Service Delivery: Assessing Two Decades of Change 25

evaluation studies show better results under approaches to public sector reform on an this approach than under the more tradi- agency-by-agency basis. This approach has tional public sector delivery model (La For- shown promise of relatively quick success, gia et al. 2004). including a handful of well-known examples In terms of education, the largest case of of impressive organizational turnarounds. contracting out is the Fe y Alegría schools Probably the best contemporary case is a set that operate in 14 Latin American countries of semiautonomous agencies created in Peru and cover approximately one million stu- in the 1990s. Instead of trying to reform the dents. An evaluation of the program’s oper- whole government machinery, Peru created ations in nine countries indicates that a series of “autonomous” agencies in key schools integrated into Fe y Alegría outper- areas related to sustenance of the marked- form traditional public schools, with lower oriented reform and targeted poverty and repetition, dropout, and higher overall social programs. These agencies included the progress rates in school and higher overall revenue administration agencies, regulatory retention rates, on average (Patrinos 2002; agencies, and social program agencies Swope, Latorre, and Celedón 1998). [Comisión para la Formalización de la The national Fe y Alegría office in each Propiedad Informal (COFOPRI) and Fondo country oversees the education quality de Compensación y Desarrollo Social (FON- given and offers training to its teachers and CODES)]. The agencies tended to be staffed center directors, administers the educa- with competent professionals who were gen- tional centers, and coordinates activities in erally better paid than civil servants in the such a way that it becomes a local area core ministries and operated with a degree of development strategy. Fe y Alegría, for autonomy from governmentwide rules gov- example, developed many strategies to erning personnel and financial management. reduce grade repetition and school dropout In some cases (for example, COFOPRI), they rates during the first years of schooling for also experimented with various ways of children coming from the most disadvan- involving stakeholders in decision making as taged areas. Critical to Fe y Alegria’s success a means of improving accountability (see are its parent-centered strategies. These are Cantuarias and Delgado 2004). based on an open school environment that Although the approach of creating such promotes effective parental involvement in efficiency enclaves is more common in areas the learning process, community involve- such as revenue administration and regula- ment in school management, and student- tion (see box 3.5), the model has also been centered strategies that depend on an inter- applied widely for the provision of specific active teaching method and recognize a social infrastructure services through social variety of learning styles. The dynamic at funds. These funds were first launched as play in this case is strongly associated with “social compensation” programs to soften the altruistic and “empowerment”-oriented the impact of structural adjustment policies approach that characterizes Fe y Alegría. In on the poor through the construction of other words, the effects of contracting out small-scale infrastructure projects targeted appear to be closely related to the contrac- to poor communities in accordance with tor’s nature. local priorities. Perhaps the most significant innovation Bypassing unreformed bureaucracies. When that social funds have brought about has the nature of the services or the political/his- been their role as means of investing in torical conditions do not permit govern- basic infrastructure. Whereas traditional ments to contract them out to either delivery lines—namely, the line min- private operators or NGOs, the near impos- istries—tended to be relatively inefficient sibility of a governmentwide civil service and focused on infrastructure in capital reform resurfaces as a major bottleneck cities (perhaps due to political pressures, to to improving service delivery. In this context, greater visibility of problems, or simply a number of governments in the region because of economies of scale and of have adopted, to different extents, targeted agglomeration), social funds, on the other 26 Citizens, Politicians, and Providers: The Latin American Experience with Service Delivery Reform

the case with traditional government agen- BOX 3.5 Executive Agencies in Jamaica cies or local governments, especially in centralized regimes or in societies characterized After taking various administrative reform respectively). Delegating authority to the approaches over the years, Jamaica intro- CEO to manage the financial and human by the presence of local strongmen. Social duced the creation of executive agencies resources needed to produce results was a funds’ new accountability arrangements (EA) as part of a broad public sector mod- major achievement of the reform.The CEOs have allowed communities to express their ernization program.While previous initia- could “hire, fire, and discipline”their employ- needs (including making choices among a tives attempted some redesign as part of ees; enabling them to maintain control over civil service reform, the system was still sub- the allocation of most inputs. variety of sectors) and to stay informed ject to central controls over resource inputs The Executive Agencies Act (EAA) provi- regarding investment decisions and works and did not address performance manage- sions that Parliament passed in 2002 incor- progress. Rules often furthered that ment. EA, which are characterized by orga- porate a strategic planning process to accountability by establishing social control nizational separation, offered Jamaica a way assess each agency during its restructuring, to improve public sector functioning and including its future organizational structure, mechanisms, as part of supervision—in enhance service delivery. staffing, and remuneration requirements. other words, communities had to sign off In creating the EA, Jamaica incorporated Since then, EA developed new performance before contractors were paid. essential features of new public indicators and the agencies are now evalu- management. It also made significant leg- ated regularly. All agencies have prepared Several “bureaucracy avoiding” account- islative and institutional changes to elimi- citizens’charters as agreements with their ability mechanisms operating in the case of nate impediments arising from traditional clients on the agencies’performance. Some social funds can be found in conditional cash central control over allocation of financial have incorporated customer satisfaction transfer (CCT) programs. In the recent past, and human resources in line departments. components among their key indicators 18 One such feature was to make the EA chief and have reported satisfactory levels in fol- CCT programs have become a popular executive officers (CEOs) directly responsi- low-up customer surveys.The EA improved social assistance approach as they have been ble to their portfolio ministers, with incen- operational efficiencies and was able to shown to generate important gains regarding tives to produce specific outputs (that is, charge reasonable user fees, in cases where both the provision of immediate assistance rewards and sanctions were given to the consumers are “identifiable”and it is CEOs for good and bad performance, efficient and equitable to do so. and longer-term human capital development impacts (Rawlings and Rubio 2003). Through provision of cash grants directly to poor households, conditional on human development–related activities such as children’s school or clinic attendance, CCT pro- hand, were able to channel investments into grams allow central governments to have a poorer, hitherto unreached rural areas and direct relationship with the target popula- to do so at lower costs and greater speeds tion, at least in terms of beneficiary selection, than the line ministries. An important rea- without the intermediation of local authori- son for this efficiency is that the funds con- ties or without having to use local service tract with third parties to implement the providers to reach the target population to works, which brings performance-based deliver social assistance.19 Indeed, the central incentives into play. Likewise, their staff government commonly administers CCT tends to be recruited competitively, with programs directly, including identifying ben- employment contracts more along the lines eficiaries and verifying compliance and of those offered to private sector employees. delivery of cash transfers. Making these There is considerable evidence that social functions operate in an efficient way (that is, funds have been effective in expanding ser- having a well-designed and -maintained vice delivery. They have also benefited from objective targeting system, making payments relatively careful evaluation (Rawlings et al. in time to large numbers of poor families 2004) that often showed positive effects in dispersed throughout the country, and so outcomes for the poor, with lower operating on), depends crucially on the existence of a costs than for most comparative programs. relatively small but technically solid and Social funds essentially created a new, politically protected central bureaucracy— shorter route of accountability between the something that most line agencies in the providers of small infrastructure invest- same country have failed to achieve. ment, on the one hand, and local communities, on the other.17 The relationship Limits to alternative compacts. The com- between the fund and the communities it mon denominator of these various innova- served was generally stronger than had been tions is the bypassing of well-established LAC Service Delivery: Assessing Two Decades of Change 27

traditional state bureaucracies (central and between the government and these enclave local) in favor of compacts either with pri- agencies and the latter’s exemptions from vate for- and not-for-profit actors or with many public sector regulations have some- “elite” state agencies or units for which per- times led to loss of effective central control, formance-based management is feasible (or and, at times, corresponding opportunistic easier) to establish. There are many, mostly behaviors. Some enclave agencies have well-documented successful examples of engaged in irresponsible or reckless finan- such initiatives. At the same time, far from cial activities (for example, excessive bor- being a silver bullet to service delivery rowing or overspending) or in patronage- reform, bypassing state bureaucracies also driven personnel practices. raises a series of broader (mostly longer- Similarly, some CCT program critics term) common concerns. argue that, while they represent a creative First, in most cases, success of these approach to providing social assistance, approaches depends on the consolidation they constitute an “end-run” around the of an “efficiency enclave”—for example, the more difficult task of reforming inefficient social fund, the regulatory agency, or the public education and health services. There CCT program management unit. Even are concerns that without greater attention though enclave agencies may be a good, to the provision of quality services, CCT second-best solution in some contexts, they program conditionalities run the risk of are hardly foolproof (Laking 2002). Of par- mandating the poor’s use of low quality ser- ticular concern is the potential lack of insti- vices, tying them to ineffective service tutional sustainability as often an enclave providers.20 These supply-side concerns agency’s short-term performance is sus- have led to calls for renewed attention to the tained with exceptional political support basic task of providing accessible, high- and government commitment (often by quality health and education services in the presidents themselves) and, thus, are poor areas. A related concern involves the vulnerable to waning commitments that local government’s role. Although adminis- accompany change of governments or trative arrangements vary considerably shifts in the political environment. More across programs, CCT programs (especially generally, the question as to how and why those with highly centralized administra- such enclaves are spared the typical clien- tion such as Colombia and Mexico) have telistic force that weakens the quality of been accused of undermining local govern- public sector employment remains unan- ments’ effectiveness by bypassing their swered. What type of political calculation authority. This concern has particular reso- led political leaders to provide special nance in countries where democratically administrative procedures and practices to elected local governments are in their a selected few agencies? It is possible to infancy and where central governments speculate that such approaches are moti- have a long tradition of clientelism and vated by an idiosyncratic need to ensure the paternalism. Strong centralization also lim- effective implementation of high priority its the program’s ability to address benefi- policies or initiatives (for example, estab- ciaries’ needs and build in local response lishing credibility of a privatization reform mechanisms for basic operational tasks by creating an independent regulatory such as targeting, verifying compliance with agency or creating an effective social pro- conditions, or addressing beneficiaries’ con- gram for social containment reasons). In cerns. It may also limit the program’s ability some cases, the donors’ role may be equally to effectively coordinate with health and important as a source of pressure and/or education service providers.21 encouragement for such approaches. These concerns have also led analysts to Second, even when key agency sustain- ask under what conditions will the “out- ability is not an issue, there may be negative sourcing” of functions create opportunities consequences of such islands of efficiency for strengthening mainstream government operating in the context of an unreformed agencies in charge of policy making. For public sector. The arms-length relations example, contracting out health services to 28 Citizens, Politicians, and Providers: The Latin American Experience with Service Delivery Reform

NGOs in Central America implies new shows that President Fujimori specifically types of activities by ministries of health targeted it using political criteria (targeting that require the development of new capa- potential swing areas near election time). bilities, such as the ability to manage a com- Similarly, case study work by Tendler (2000) petitive bidding process and to oversee, in the northeast of Brazil documents, in the monitor, and evaluate contract compliance. cases analyzed, how “demand-driven” ini- Indeed, the development of these purchas- tiatives were often captured, including on ing arrangements is seen as an opportunity the supply side, by private companies that to develop these skills via learning-by- offered communities particular products doing. To date, perhaps with the exception that they could deliver and then suggested of Guatemala, such changes have not mate- forming local committees to give the stamp rialized (Fiedler 2002). of participation.23 Similarly, many project documents and Problems of capture have also been iden- charters from the mid-1990s suggested that tified in the regulation of contracts with the social funds would play a role only tem- private sector, which remains a major chal- porarily in the delivery of small infrastruc- lenge, notwithstanding substantial advances ture services, until the line ministries were in creating legal frameworks and regulatory strengthened and better able to perform institutions. Key problems are the difficulty these functions themselves—when it was of ensuring an arm’s length relationship assumed that they would turn the functions between the regulator and the executive back to the line ministries. For a number of branch and the challenge of balancing reasons (including the funds’ administrative investor and consumer interests in societies efficiency which made them attractive to with a limited tradition of consumer organi- governments and donors), in most cases this zation and participation (see the next sec- transfer did not materialize. Governments in tion). Newly privatized companies become many countries began using social funds as part of new distributional coalitions, where the de facto and, at least in the case of Hon- large conglomerates play an important role duras, official principal agencies for small- in contributing to a perception of concentra- scale social and economic infrastructure. It tion of power and profits. A series of Latino remains an open question whether such spe- Barometro polls, for example, find that a cialization is in the end effective in enabling growing proportion of those questioned dis- sector ministries to concentrate their func- agreed with the statement that privatization tions on policy-making and regulatory had been beneficial for their countries. Many aspects—a stated goal of the movement to critics are concerned with how fairly privati- institutionalize the role of social funds.22 zation’s benefits have been distributed Third, there is a broader issue on the dif- among different stakeholders and, in partic- ficulty of balancing autonomy (for exam- ular, the extent to which the state and private ple, of social funds, CCT units, or regula- investors may have benefited at the expense tory agencies) and ultimate accountability. of consumers and workers. This phenome- The many successes throughout the region non can be attributed partly to biases in the of such approaches cannot obscure the real original design of privatization transactions danger of capture by clientelistic politicians that tended to emphasize fiscal concerns or by private suppliers and concessions; over social and economic considerations. It even enclaves are embedded in sociopoliti- probably also reflects deficiencies in the cal systems. For example, analyses of the ongoing regulatory process, which has not major Mexican Social Fund scheme of always been successful in transferring effi- Pronasol (or Solidaridad) document how it ciency gains into lower tariffs for customers. was systematically used for clientelistic pur- Indeed, the challenge of developing an effec- poses (Diaz-Cayeros et al. 2002; Molinar tive counterweight in noncompetitive sec- and Weldon 1994). For the Peruvian Social tors has been difficult due to classical prob- Fund of FONCODES, while Paxson and lems of regulatory capture and the common Schady (2002) documented its success in interest of governments and privatized utili- terms of pro-poor targeting, Schady (2000) ties in high profits. LAC Service Delivery: Assessing Two Decades of Change 29

As the previous review indicates, when mary objective. As a result, many of these public bureaucracies don’t work well, short- experiments have benefited from evalua- circuiting them by contracting out services tions that seek to identify their success in to private firms and NGOs or setting up achieving service delivery objectives. We parallel, more independent agencies in will first review the experience with client charge of service delivery is showing some choice and next with client participation in good results, at least in the short run. There the management of services. are many indications, however, that in the longer run both their sustainability and full In the limited areas in which it effectiveness depend on whether the rest of has been implemented, giving the state apparatus improves its capacity and becomes more performance-oriented. individuals a choice over which All in all, to initiate and sustain good, organization they get services from effective service delivery compacts (within has generated powerful pressures for or outside the state), at a minimum, results on front-line providers requires the basic commitment and support The introduction of choice as a mechanism of policy makers. As the examples reviewed to promote client power over service in this section show, LAC countries have providers has been a powerful one, albeit of been successful in developing such com- restricted coverage. In the infrastructure pacts, particularly when they transferred sector, the 1990s reforms resulted in the management responsibilities outside the gradual introduction of competition for mainstream state bureaucracy. Nonetheless, electricity only (in the case of a relatively many questions remain as to the political small number of large industrial con- incentives at play to explain the genesis, sus- sumers) and for long distance and cellular tainability, and capacity to scale up (or out) telephony. In the water sector, competition of such innovations. We speculate that the has been limited to bidding out concession overall success of compact reforms is contracts. However, even there, the extent of related to broader efforts to enhance both competition has been quite limited due to client power and citizen voice. the relatively small number of multination- als active in the sector (Foster 2002). Rela- Enhancing Client Power tively little attention has been paid to the When clients have power over providers, a competition that arises from small-scale shorter, more direct form (that is, less independent providers of water services, dependent on actions and performance of such as water tankers and secondary water other actors) of accountability for results retailers, particularly in underserved peri- may arise. Over the last two decades, many urban areas (Solo 2003).24 Latin American countries have explored In education services, two key experi- different ways of enhancing client power. In ences with choice are those of Colombia some cases, this entailed giving individuals and Chile (see box 3.6). In the first half of the power of choosing which organization the 1990s, Colombia introduced a sec- they get their services from, with the expec- ondary education targeted voucher pro- tation that this will create better (market- gram, which involved approximately one- like) incentives to providers. In other cases, fifth of the country’s municipalities and it involved giving clients a direct (collective) close to 2,000 private schools. The main role in service management, with the expec- objective was to increase the number of tation that such involvement should result poor students transitioning from primary in stronger control over the actions of to secondary education. Chile, on the other front-line providers and a corresponding hand, represents a unique case in the sense positive change in the quality of services that a nationwide system of “quasi-vouch- provided. ers” directed to all municipal schools and Unlike most of the changes discussed in nonfee-charging private schools was estab- the previous section, these transformations lished with the main objective of promoting had service delivery improvement as a pri- competition among schools. 30 Citizens, Politicians, and Providers: The Latin American Experience with Service Delivery Reform

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

Chile’s higher education reforms of the 1980s, effort, Chile’s two oldest public universi- about 200,000 students in 1983 to about instituted under the military government, were ties were subdivided into 16 new universi- 475,000 students in 2001. Also, the num- the most radical reforms of higher education ties, including pedagogical institutions. ber of Master’s students increased from undertaken in Latin America.Those reforms Overall, the purpose of these reforms was about 1,500 students in 1982 to 8,000 sought to increase learning and enrollment to offer students more choices and to students in 2000; the number of doctoral while decreasing public spending on higher force educational institutions to innovate students increased in similar proportion. education.These reforms successfully expanded in competing for students. Box Figure 1 shows that this increase schooling access, though their immediate effect Competition for funds. Chile decreased pub- was more than double the mean on schooling quality has been unclear. lic expenditure on higher education, but increase for Latin America, and slightly above the mean increase for OECD Competition between school types. In sought to increase funds from private countries.The immediate effect on qual- Chile’s initiative for diversiﬁcation, sepa- sources (that is, charging tuitions) to ity of schooling, however, was unclear. rate tiers were created for universities, compensate for the lower spending.The Partly due to the enrollment of many Professional Institutions, and Centers for government also redesigned a small lower-achieving students, learning may Technical Education.In Chile’s deregula- portion of its spending in order to allo- have decreased in the short run. In 1999, tion effort, private institutions were cre- cate some funds based on quality and to address concerns about quality of ated and allowed to operate with minimal need, while addressing credit market education, national accreditation com- oversight.In the 1980s, 40 private univer- constraints that could cause students to missions began setting program stan- sities, 78 private Professional Institutes, underinvest in their own human capital. dards and recognizing programs that and 61 private Centers for Technical Edu- Outcomes. Between 1983 and 2001, enroll- met those standards. cation were created.In its decentralization ment in higher education rose from

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

300 250 200 150 100 50 0 MexicoBrazil Argentina LA Average Uruguay New Spain OECD Colombia Chile United Portugal Zealand Kingdom Sources: OECD, 2003; Holm-Nielsen et al., 2004.

In the health sector, once again Chile and insurers, but also introduced demand-side Colombia introduced paradigmatic reforms premium subsidization for the poor. that put competition and choice at the cen- Colombia also introduced an explicit benefit ter. In the 1980 context of a complete over- package and a risk/income equalization haul of its social security system, Chile fund, both important instruments to reduce reformed its social health insurance system, the market segmentation incentives on a introducing private-public competition for risk basis. mandatory health insurance. The reform Overall, the results of competition have allowed all formal workers, independent of been very positive in all sectors where it has their income level, to choose between a pri- been tried. In infrastructure services, intro- vate health insurer (ISAPRES) and the pub- ducing competition in cellular telephony lic national health fund (FONASA) for reduced charges and catalyzed a major cov- mandatory health insurance coverage. The erage expansion. The number of cellular reform, however, mandated that public sub- lines in service grew 500-fold during the sidies would be given only to FONASA ben- 1990s (at a 57 percent annual average eficiaries irrespective of the income and risk growth rate), to the point where, by the year characteristics of those opting for ISAPRES. 2000, cellular telephone penetration had In 1994, Colombia also introduced a radical overtaken fixed line coverage in countries as social health insurance reform that, as in the diverse as Bolivia, Chile, Guatemala, Mex- Chilean case, promoted competition among ico, and Venezuela. As a result of competi- LAC Service Delivery: Assessing Two Decades of Change 31

tion, cellular call charges fell between 30 and 50 percent in Bolivia, Peru, and BOX 3.7 User Fees Venezuela, largely as a result of the intro- An integral element of the 1990s infrastruc- Traditional cross-subsidy schemes are duction of the “calling party pays” system. ture reform process was to ensure that util- often poorly designed, failing to target sub- Similarly, long distance telephony liberal- ity tariffs reflected accurately the full eco- sidy resources effectively to the poor. In La ization drove down international calling nomic costs of service provision. Paz, Bolivia, for example, the average water costs by 50 percent in Chile, Colombia, Cost-reflective tariffs are needed to consumption of a low-income household is safeguard service providers’financial sus- around 5 cubic meters per month, and the Mexico, and Peru, and by more than 70 per- tainability and hence, political average consumption for a higher income cent in El Salvador and Guatemala. Even in independence.This principle had varying household is around 25 cubic meters per noncompetitive sectors, the move toward implications according to the sector and month, showing a clear differentiation. cost-reflective charges (see box 3.7), com- locality, and it is difficult to generalize about However, the water utility provides a subsi- the experience. Indeed, a recent study dized rate on the first 30 cubic meters per bined with the establishment of customer found that out of 10 privatization cases month, so that higher income households complaints channels, have begun to make studied, prices rose in five and fell in the obtain the largest subsidy benefit. Even in consumers more demanding about the other five (McKenzie and Mookherjee 2002). countries like Chile and Colombia, which The same study shows that in most have developed more sophisticated individ- quality of service they receive. cases, welfare losses arising from higher ual or geographical means for targeting sys- The Colombia education voucher expe- utility tariffs that the reform triggered are tems for water subsidies, significant rience was thoroughly evaluated (Angrist more than compensated for by the welfare deficiencies remain. Gomez Lobo and Con- and others 2002, 2004). Results indicate that gains associated with expanding access to treras (2000) show that about 65 percent of services (McKenzie and Mookherjee 2002). these subsidy resources go to households in the participating municipalities, the pro- The reason is that the budget share in the top 70 percent of the income distri- gram provided an effective way to increase allocated to the utilities is relatively low, and bution, while 55 to 75 percent of the benefi- educational attainment and achievement hence tariff increases in the cases studied ciaries come from the top 70 percent of the (college-entrance test scores) for the poor (Argentina, Bolivia, and Nicaragua) income distribution. accounted for less than 1 percent of house- The accumulated evidence suggests that students benefiting from the vouchers. Crit- hold budgets, while the gains from access the transition toward cost-reflective user ical were the sound criteria in the selection can be greater than 10 percent of budgets charges needs to include a strategy to evalu- of the participating municipalities and for the poorest families, as households are ate and safeguard affordability of a subsis- schools and the application of effective tar- able to abandon more expensive tence basket of utility services to the poorest substitutes to utility services. In many coun- households.In some circumstances, it may geting mechanisms, with the use of perfor- tries, the reform of service tariffs did not be much more important from a social per- mance incentives for participating schools adequately evaluate the extent to which spective to focus scarce subsidy resources on and beneficiaries. Despite its positive cost-reflective tariffs would be affordable to reducing access costs for the unconnected, results, the lack of the government’s com- the poorest clients, or investigate appropri- rather than reducing prices for those who ate compensating mechanisms. already have access to the service. mitment, the teacher union’s opposition, and administrative difficulties (disburse- ment delays and program monitoring burden) led to the program being discontinued as a national initiative in 1997. Neverthe- 1980 to 32 percent in 1990), which, control- less, some local governments continued ling for socioeconomic status, perform bet- implementing similar approaches without ter than municipal schools (Carciofi, national support. Beyond the idiosyncrasies Cetrángolo, and Larrañaga 1996; McEwan of the factors leading to the demise of this and Carnoy 2000), pointing to positive particular experience, the message appears effects of competition on the private to be that, unless the ultimate beneficiaries schools’ performance. However, it is gener- of the increased choice have sufficient voice ally recognized (McEwan and Carnoy 1999, (that is, they are sufficiently empowered) to 2000; Hsieh and Urquiola 2003) that a lack defend their gains from those whose inter- of incentives and municipal sector capacity ests such schemes affect, sustainability may to compete,25 a lack of clear information on be a serious bottleneck. relative school performance, and some The Chile “education voucher” experi- “cream-skimming” among private schools ence generated a range of evaluation studies hampered competition among private and centered on the relative performance of municipal schools. public and private schools. Overall, with the In the 1990s, further municipal auton- introduction of the per-student subsidy, omy constraints (for example, the 1991 enrollment increased considerably in the teacher statute) and centrally driven inter- private voucher schools (from 14 percent in ventions in favor of the poorest/weakest 32 Citizens, Politicians, and Providers: The Latin American Experience with Service Delivery Reform

municipal schools complicated even more encouraging. Nevertheless, this should not the working of competition between munic- obscure the fact that, as illustrated in the ipal and private schools.26 On the other cases discussed above, the conditions for hand, dissemination of standardized school choice to work (particularly for the poor) scores (SIMCE scores), increased adminis- are not minor. On the demand side, the trative and pedagogical municipal school existence of highly unequal economic and autonomy; introduction of school cofinanc- social conditions in most countries in the ing and improved monitoring of student region require active public policies to com- attendance improved the voucher system’s pensate for extensive informational asym- operation. The studies centered on the metries (for example, on the quality of edu- 1990s generally confirm that private subsi- cation services as in Chile in the 1990s) and dized schools perform better than munici- differences in purchasing power and needs pal schools (Sapelli and Vial 2002) and have (for example, by introducing income/risk improved their performance due to strong equalization mechanisms for health insur- incentives to compete (decision-making ance in Colombia). Without such compen- autonomy, reliance on the voucher or parent satory actions, increased social stratification cofinancing, and so on), therefore conclud- becomes a significant danger. On the supply ing that competition and choice were help- side, the presence of independent providers ful in improving private schools’ quality and operating in a competitive environment is, efficiency (Gallego 2002). However, compe- naturally, key to the proper functioning of a tition’s impact on municipal schools’ perfor- model based on choice. Both the existence mance remains more controversial, with of structural bottlenecks to competition some studies (Gallego 2002) finding that (for example, in the case of many utilities or improved municipal performance was in schooling in dispersed rural areas) and largely due to competition, while others the social desire to ensure minimum quality (Schiefelbein and Schiefelbein 2000; Cox standards (for example, safe water or prede- and González 1998) assert that it was due to termined learning goals) are satisfied the central compensatory programs. throughout the country require, again, the Turning to the health sector, social health need for government intervention in service insurance reform in Chile resulted in a sig- delivery, or most likely, in regulation. nificant increase of both private health insurance coverage and private provision of Client participation in the health services. At the same time, it also resulted in a severe segmentation of the risk management and operation of pool, with the high-income and low-risk services has also shown good results groups concentrating in the private for services that allow a relatively (ISAPRE) system and the low-income and small production scale high-risk concentrating in the public one Client participation in service management (FONASA).27 The Colombian reform, on is the second way that client power oper- the other hand, is expected to address a key ates. The experiences with community-run factor that determined income market seg- schools, health clinics, and rural water sys- mentation in the Chilean case through the tems in a diverse group of countries provide introduction of subsidized insurance premi- the empirical basis to assess the power of ums for the poor. Preliminary evaluations of client participation in service delivery. the Colombian reform (Bossert et al. 2003) Overall, the evidence suggests that direct indicate an increase in service utilization client involvement in the management of and coverage for the population at large. services can yield positive results when the However, the evaluations do not distinguish services involved allow a relatively small clearly the specific effects on the poor, nor production scale and rather simple man- the extent to which reforms improved the agement challenges. financial protection of households. The experience of community-run schools In general, the evaluation results from in El Salvador and Nicaragua provides solid the experiences of expanded choice are evidence on the benefits of a participation LAC Service Delivery: Assessing Two Decades of Change 33

approach to client power. Each school is cational process (proximity to the final required to establish an elected council— users, continuity in the provision of the ser- composed only of parents in El Salvador vice), the scope for user participation will be and of parents, teachers, and the school particularly important in that sector. director in Nicaragua. The school must also In contrast, the evidence on effective sign a special agreement with the Ministry experiments with community-run health of Education under which the latter decen- centers is rather limited. One interesting tralizes to the councils the management of example is that of local health administra- all funds, allowing them to manage all tion committees (CLAS) in Peru. These are teachers, take care of operating and mainte- private, nonprofit community associations nance costs, and even make academic deci- built around a health center or post (Cotlear sions. “Client power” is therefore expressed 2000). The CLAS prepare local health plans through active participation in school man- that become the basis for a contract with agement. In addition, in Nicaragua, client health authorities. The contracts are power also finds a way through the system designed to finance outputs rather than of national, departmental, and local educa- inputs, and the CLAS are held accountable tion councils set up to ensure broad partici- for reaching specific targets (derived from pation at all levels. the plans) and for the use of funds. The The key aspect of this model is the control CLAS are given flexibility in the use of that parents have over teacher management resources relative to state-run centers. While decisions—exactly the aspect that has proven there is no systematic evaluation of the the most controversial and hard to reform in approach’s impact, the data do show an state-managed school systems. In Nicaragua, improvement in health indicators (for exam- school-based decisions on teacher manage- ple, lower waiting time in CLAS centers). ment are shown to have a positive impact on There are a number of factors that seem test scores (King and Ozler 1998). In El Sal- to conspire against community participa- vador, enrollment in the rural areas increased tion in health centers compared to that in massively following Educación con Partici- school management. First, in most cases pación de la Comunidad (EDUCO), and it users of health services have a very intensive has been shown that school-based decisions (when sick), but much less frequent, rela- on teacher management have a positive tionship with providers than parents of impact on class attendance (Jiménez and children have with schools. In that sense, Sawada 1998). solving the collective action problem The El Salvador and Nicaragua cases are involved in community-run services is the most thoroughly evaluated experiments likely to be more difficult in health. Second, with community-run schools, but there are informational asymmetries tend to be very a growing number of such experiences in large in the case of health services. As a Central America (for example, Guatemala result, the opportunities for health profes- and Honduras) and elsewhere in the region. sionals capturing control under commu- For example, the much referred to experi- nity-run schemes is likely to be higher than ence of the Brazilian state of Minas Gerais in the case of schools. Overall, these factors with community councils and elected may help explain the observed differences school principals (Guedes et al. 1997) is one in the popularity of the approach between of many examples of a growing trend the two sectors. toward school-based management experi- Unlike the case of health centers, commu- ences in Brazil (see Aglaê de Medeiros nity committees (juntas de agua potable) and Machado 2002). Overall, the evidence is cooperatives throughout Latin America have rather strong in the sense that decentralizing successfully managed rural water and sanita- responsibilities to the school, particularly in tion services. Over the years, as coverage lev- the teacher management area, makes it pos- els increased in rural communities, it became sible to maximize the use of local informa- impractical and expensive for centralized tion and accountability mechanisms. It is service delivery agencies (usually located in likely that, given the peculiarities of the edu- the Ministry of Health, Public Works, or 34 Citizens, Politicians, and Providers: The Latin American Experience with Service Delivery Reform

Urban Development) to take charge of oper- appear. Similarly, the fact that ministries of ating and maintaining water systems spread education in Guatemala and Honduras had across the country. Overall, the community- less capabilities to enforce “school com- management approach has proven to be pacts” than their counterparts in El Sal- highly successful for rural water services. vador and Nicaragua may be the principal Approaches to community-based service reason for the difficulties experienced in delivery have evolved from water systems their community-run schools.29 constructed by a national agency and subse- The overall authorizing environment is quently “handed-over” to communities for also crucial when elements of community- management, to more demand-responsive run services coexist with nonparticipatory approaches where the community partici- mechanisms. As the experience with water pates in all phases of the project cycle and supply programs in rural areas (Gran et al. shares in the investment costs of projects. 1997) showed in some cases, even though Chile provides a good example of the first communities took educated decisions based approach (Alvestegui and Sara 2003), which, on the participatory program in question, while generating impressive results in terms since the construction of the systems con- of coverage,28 may have led to insufficient tinued to be the responsibility of a “nonre- community attention to finding lower cost formed” state agency, there was no way they solutions and to investing in system rehabil- could ensure that the works be conducted itation or expansion. Community participa- according to plan. In other words, demand- tion advantages are perhaps felt more orientation must take place consistently in strongly in the latter approach, followed in all the stages of the provision process countries such as Bolivia and Ecuador; this (financing, production, delivery, and so on) approach aims to ensure communities otherwise the “weaker” link of the chain obtain a system for which they have the ends up dominating the rest. financial willingness and technical capacity Beyond establishing the authorizing envi- to maintain it (World Bank 2002c). ronment, central governments can develop The details of the service-specific autho- complementary mechanisms to support the rizing environment—in particular how operation of community groups. In many flexible and consistent are the rules and ways, social funds are evolving in that direc- constraints under which community tion. The latest generation of funds is groups need to operate and the nature and becoming a mechanism that both empowers enforceability of accountability relations to communities and local governments and the authorities delegating powers—can be channels national preferences (such as a very influential in the outcomes of the poverty targeting emphasis) in the form of experiences of community-run services. incentives to local actors. Community par- In the case of community-run schools in ticipation is now viewed as an end in and of Central America, for example, both in El itself—almost on par with the goal of Salvador and Nicaragua the recorded suc- increasing access to basic infrastructure— cesses are, in part, linked to the existence of and funds have incorporated not just new a solid “compact” linking the school with rules of the game (that is, in the operations the Ministry, which seeks to ensure the ful- manuals), but also specific training geared to fillment of coverage, efficiency, and equity strengthen community capacities for plan- targets. A weak (or weakly enforced) “com- ning and implementation.30 Similarly, the pact,”on the other hand, implies that school experience with rural water and sanitation autonomy and strong parent involvement projects indicates the importance of provid- in financial and administrative matters ing appropriate community outreach, orga- could produce, for instance, cases where nization, and training programs to comple- teachers are unfairly fired or to cases of mis- ment hardware investments. The approach handling of funds. There is, in fact, some has required centralized agencies to trans- preliminary evidence from El Salvador form their roles from service providers to showing that as the Ministry weakened its facilitators. A new set of intermediaries, usu- control of schools, such cases begun to ally specialized NGOs and small contracting LAC Service Delivery: Assessing Two Decades of Change 35

firms, have emerged to provide assistance to communities in formulating their demands, BOX 3.8 The Role of Community Participation and building capacity, and implementing pro- Cultural Adaptation in Service Delivery jects, while the central government provides the normative framework, access to funding, Malnutrition in Honduras dropped rapidly has about one-ninth the cost of the first through the 1980s, but progress stalled in facility-based program and has had far more and technical oversight. 1990.The Ministry of Health required public extensive reach than that program had. Both the opportunity for and ultimate health centers to detect low growth (stunt- Community participation also implies results of client-participation schemes ing) among children, then to use stunting as the potential to incorporate adaptation of a criterion for distributing health services. service delivery to local culture. One clear appear to be context-dependent: national By 1994, the Ministry concluded that the example comes from a maternal and child and local political conditions, community program needed closer community ties to health project in the Paruro (Cusco) organization traditions, and the strength of maximize impact and designed the Com- Province in Peru where an effort was made front-line providers (for example, teacher prehensive Attention to Childhood in the to improve the number of women from Community program (AIN-C) for that pur- rural areas who attended the clinic for child- unions) matter in explaining outcomes. For pose. AIN-C targets all children less than birth. Findings from earlier interventions example, both the El Salvador and Nicaragua two years of age in poor communities.Vol- highlighted that women from remote rural education experiences started after a civil unteer monitors meet with mothers areas were unwilling to leave their children war and, therefore, were also part of a pacifi- monthly and plot the infant’s growth for an extended period, found the clinic progress on a simple bar chart. In AIN-C, it is cold, and felt uncomfortable using formal cation and democratization attempt. They communities that request the program Western birthing procedures.The project were initially implemented on a subnational before it begins operations, decide how established a “waiting house”where women scale to be extended to different extents. This many monitors will operate, and select the nearing childbirth could be lodged and fed not only reminds us of the idiosyncrasies monitors.The monitors give quarterly in a house close to the clinic, installed extra reports to the community with focus on heating in the birthing room, and provided behind these cases, but also that the political communal obstacles to good health, such an alternative bed to the standard Western climate may have been particularly welcom- as contaminated water or a lack of shared “birthing table.”The result was a dramatic ing to such initiatives given the need to con- childcare.They negotiate with mothers rise in medically attended births and a solidate democracy to which such participa- about specific needed practices to improve sharp fall in postnatal mortality of both 31 a child’s nutrition and decide whether to babies and mothers.The practices have also tion was functional. In fact, the EDUCO reach mothers by house-to-house visits, been followed by many other health cen- experience grew out of community efforts neighborhood sessions, communitywide ters in rural areas in Peru with the same dra- during the civil war: it was a case where a rel- events, or other means. matic results.This is a clear demonstration AIN-C has been shown to increase of how attention to cultural issues can pro- atively weak national state took advantage of breastfeeding to children under six months vide great advances in fighting poverty—in relatively strong communities (rather than a of age, improve provision of oral rehydration this case on the health dimensions—and in strong state empowering weak communi- solution to children with diarrhea, expand the general development effectiveness of ties) to respond to growing social needs in the giving of vaccines to infants, and spread projects and programs. other preventative health measures. AIN-C Sources: Griffiths and McGuire, 2004; Ministerio de the postwar scenario. Salud, 2000. Sustainability can be an issue when relying too much on community-based management mechanisms. Community-based schools, for example, are often quite isolated from the by unequal group-based relations. This can state’s overall institutional structure (includ- be of particular importance where there is ing the absence of links with the intermediate greater social distance between front-line and local government levels) and have devel- workers and their “clients,” as with indige- oped at the margin of the teacher unions; in nous students and nonindigenous teachers. this context, they are particularly vulnerable The ways in which poor groups interact to political change.32 Similarly, the CLAS with front-line providers will be a function experience in Peru indicates that while com- of both the larger social context and histo- munities received them with great sympathy, ries of mobilization and engagement. At the they encountered significant resistance in a same time, when communities are offered state bureaucracy that resented the implied sufficient support, client power can con- loss of control (Cotlear 2000). tribute to changing the local balance of Client power relationships are affected power, whether this is through increased by the incentives, motivation, and behavior parental influence in EDUCO schools in El of front-line workers. These relationships Salvador, the introduction of culturally are particularly embedded in the local appropriate practices to birthing and post- sociocultural conditions, the behavior pat- natal care in Quechua areas in Cusco, Peru, terns, and collective expectations produced (see box 3.8, above), or the provision of the 36 Citizens, Politicians, and Providers: The Latin American Experience with Service Delivery Reform

CCT grants to mothers, which combined mining whether service delivery systems with the election of mothers’ local represen- work for people. To be sure, whether or not tatives to serve as conduits between benefi- policy makers face the incentives to address ciary families and the CCT program, intro- the needs and preferences of citizens (that duced changes in empowerment dynamics is, whether they are accountable for results) that are playing out in households and is fundamental when it comes to carrying communities throughout the region (Adato forward and implementing the reforms to 2000; and Adato et al. 2000). the compact between policy makers and The evidence reviewed indicates that providers discussed above. And even if the client participation can compensate for the client-power alternatives discussed above weakness of long-route mechanisms. Ulti- require less “hands-on” involvement by mately, however, the effects and sustainabil- state agents in the actual delivery of ser- ity of such client power mechanisms are vices, whether and how such reforms are dependent on both the existence of capaci- implemented and supported over time is ties (agency) among the poor and minimum clearly influenced, once again, by the nature state capacities to initiate and regulate their and intensity of the accountability relation- implementation. When and if the condi- ship linking policy makers and different tions to initiate such approaches exist, a vir- groups of citizens. Said differently, politics, tuous circle may be generated whereby political institutions, and citizen voice con- client participation continues to generate stitute fundamental determinants of capacity-building opportunities for the whether services work for people—particu- social organizations of the poor and disad- larly poor ones. vantaged and of individuals and families— In this section we review the evidence on factors that should contribute to making the how some fundamental political and social state more accountable overall to citizens. transformations in Latin American coun- Overall, the evidence strongly suggests tries over the last two decades have affected that taking decisions away from public state accountability for service delivery. bureaucracies (central or local, old or new) Indeed, the sharp expansion of electoral into the hands of individual or group of democracy and constitutional rule (and the beneficiaries is proving to be a powerful associated expansion in freedom of expres- idea. The growing number of evaluation sion and association) is one of the most studies available shows solid results regard- striking transformations that took place in ing the benefits of initiatives oriented to Latin American over the last two decades. enhance client power (through choice and How these changes affected state account- direct management control of front-line ability is the subject of this section. service providers), even though the spread (in terms of the number of services and Despite the widespread number of people that benefit from them) democratization process at the of such approaches is still not too wide. The evidence also suggests that client power as national and local levels, many service delivery mechanisms is also influ- bottlenecks remain to the full enced by the overall strength of state operation of mechanisms of state accountability: weak overall state account- accountability ability is likely to undermine service deliv- Universal suffrage came late to Latin Amer- ery improvements that are achieved ica compared with the United States or through enhanced client power. Europe, excluding poorer groups for longer, via literacy, wealth, or gender (De Ferranti et al. 2004). But suffrage has now been univer- Enhancing State Accountability sal for several decades and certainly so for to Citizens the latest democratic transitions. Arguably, Ultimately, the extent to which the state is LAC is better placed than the other develop- accountable to citizens (particularly poor ing regions to take advantage of the so- ones) constitutes the key factor in deter- called long route to accountability—citizens LAC Service Delivery: Assessing Two Decades of Change 37

pressuring politicians, and politicians and political accountability mechanisms pre- policy makers pressuring providers— vailing in Latin American democracies do because, on average, democracy is better not appear to be strong enough to make the established in the region. Democracy would achievement of effective service delivery to be expected to be of particular importance all citizens a central goal of government for service delivery in Latin America in light actions. of the ineffectiveness of most of the region’s Indeed, there is a seeming consensus authoritarian regimes to deliver broad- among political scientists that the demo- based services. For a number of reasons, cratic regimes that have been established in East Asian–style developmental states that Latin America present serious institutional were effective both at promoting develop- bottlenecks, particularly in relation to the ment and in including most of the popula- development of adequate and effective tion in the growth process were not a mechanisms of accountability (for example, notable feature of Latin American develop- O’Donnell 1993, 1994, 1996; Shifter 1997; ment. This can be attributed to a range of Schedler, Diamond, and Plattner 1999; factors, including the lack of external threats Smulovitz and Peruzzotti 2000). Latin that were forces for state-building in East American democracies have been referred to Asia (and earlier in Europe) and greater as “delegative” rather than “representative” social cleavages in Latin America that made (O’Donnell 1994), “thin” (Evans 2002), national institution-building more difficult “illiberal” (Zakaria 1997), and “insufficient” and reduced the responsiveness of elite- (Méndez 2000), and characterized by “citi- dominated governments to poorer groups. zenship of low intensity” (O’Donnell 2001). There were partial exceptions,33 but even In part, those bottlenecks are generic these cases fall far short of development and not specific to Latin America. Repre- action by authoritarian states in East Asia, sentative democracy requires the existence from predemocratic Korea to Vietnam in the recent period. The ineffectiveness of the authoritarian regimes has naturally heightened general BOX 3.9 Does Democracy Lead to More Social Spending? expectations of democratization to bring Building on a previous set of studies look- ment rates.With data for Latin American about major improvements in inclusive ing at the relationship between democracy countries in the 1973–97 period, Kaufman governance. However, an overview of the and economic growth, there is a growing and Segura-Ubiergo (2001) find no effect of record suggests that the democracy wave of literature that analyzes empirically the “domestic politics”(whether a country is the 1980s and 1990s has not yet ushered in a question of whether democracies value democratic or not and whether more social services than authoritarian government’s support base is strongly significant transformation of service deliv- regimes do. Using data on coverage of key linked to unions) on aggregate social ery. To be sure (as discussed above), there education and health services and democ- spending levels.They do find, however, that has been a marked expansion of social ser- racy indicators (for example, the degree of democratic governments spend more on vices in the last two decades that reflects at competitiveness of political participation to health and education, while “popularly the degree of constraints on the chief exec- elected”governments protect expenditures least partly the response of governments to utive), Lake and Baum (2001) find that on pensions and other transfers that tend citizen demands. This pattern is confirmed increases in the democracy level are posi- to benefit their support base. Using data by the growing empirical literature seeking tively related to levels of public services from 17 Latin American countries between to establish causal relationships between both in cross-section and time series analy- 1980 and 1992, Brown and Hunter (1999) ses.They also find that the effects of regime find that in times of economic crisis, democ- political regimes and economic and social changes (that is, democratization) on the racies tend to protect social spending rela- outcomes (see box 3.9, below). But electoral level of public service provision were both tive to nondemocracies. democracy alone has been less effective in rapid and disproportionate. Baum and Lake In one of the few country level analyses (2003) find that democracy is associated available, Brown (2002) explores education situations where what was required to with higher life expectancy and education spending trends in Brazil and links them to improve service delivery to the poor levels, which explain higher economic electoral competition patterns. He finds that involved tackling vested interests (for exam- growth rates. increasing electoral competition in Brazil ple, in reforming higher education financ- Using data from countries in Africa, Asia, has been associated with an increasing and Latin America, Brown (1999) finds that share of the federal budget allocated to ing or reforming “truncated” social insur- democracy (measured both as a dichoto- education and a shift from university to pri- 34 ance systems), reducing state “capture” by mous and a continuous variable) has a mary education spending. powerful firms, organizations, or individu- strong influence on primary school enroll- als, or resolving distributional conflicts. The 38 Citizens, Politicians, and Providers: The Latin American Experience with Service Delivery Reform

of institutional mechanisms that bridge the tion and market-oriented reforms in the gap between citizens and their political rep- 1980s and 1990s with relatively weak insti- resentatives—accountability mechanisms tutions for the articulation, aggregation, that ensure that public officials are answer- and negotiation of the interests of poorer able for their actions. Elections constitute groups and inadequate mechanisms for res- the primary form of political accountability olution of distributional conflicts. This may even though they are recognized as rather have left the new democracies too exposed restrictive and limited, among other rea- to capture by all sorts of special interests. sons, given that citizens only get one shot to A related characteristic of new Latin punish or reward politicians for numerous American democracies is an observed decisions and actions (Przeworski et al. weakness in systems of checks and balances 1999; Keefer and Khemani 2003). National that the executive power is subject to. The elections are a rather crude mechanism for political science literature speaks of weak citizens to enforce policy makers’ responsi- “horizontal” (O’Donnell 1994, 1998) or bilities for the delivery of specific services. “intra-state” accountability (Mainwaring But there are reasons specific to Latin 2003), associated with the inadequate role America that may help explain why the lat- played by legislatures, judiciaries, and state est wave of democratization has not yet control organisms in most Latin American delivered on the promise of radically countries. The idea (see figure 3.1) is that enhanced political accountabilities. The state accountability takes place not only functioning of democracy in Latin Ameri- through elections (“vertical accountabil- can countries appears to be deeply influ- ity”), but also through the checks and bal- enced by the depth and resilience of clien- ances among different organs of the state. telistic structures, whereby politicians Legislatures in most Latin American generally draw support from relatively nar- countries show some worrisome signs of row, geographically defined constituencies weak capacities to act as effective counter- to which they provide tangible material balances to strong executives, for example, goods (that is, patronage) in exchange for in terms of budget formulation. Rare are the electoral support. An additional feature of instances where Congress tries to hold the Latin American political history is a pattern executive branch accountable for the perfor- of “corporatism,” in which governments mance of particular policy programs and for brought working class or peasant move- the budget. Instead of working through the ments into national, institutionalized formal budget process on the basis of the patronage-based systems, as in the case of constitutionally vested power to oversee the Mexico with the PRI, Argentina with Peron- executive branch, Congress often resorts to ism, and to a certain extent, in Bolivia with passing legislation with serious budgetary the MNR and Brazil under Getulio Vargas implications without submitting them to (De Ferranti et al. 2004). rigorous fiscal analysis. This then creates a In many cases, clientelism and corpo- situation whereby a country effectively has ratism had a direct negative impact on pub- dual budget processes, one formally gov- lic sector service delivery systems,35 making erned by the budget law, but informally difficult the implementation of reforms in managed with executive discretion, and the state bureaucracies or the transfer of power other developed through substantive legisla- to alternative service providers (private section with budgetary impacts that only tor firms, NGOs, or communities). At the become apparent ex post.36 same time, clientelism and corporatism Prevailing electoral and party systems have also worked against the development often do not contribute to creating pressure of programmatic (idea-based) parties that to change this behavior pattern. In many was an important part of the political evo- countries, the less-than-democratic process lution of most European countries and the of selecting congressional candidates within United States (Shefter 1994). One conse- parties (particularly when elections are run quence has been that most Latin American based on closed party lists, as practiced in countries have faced democratic consolida- Bolivia and Venezuela) leads to weak voter LAC Service Delivery: Assessing Two Decades of Change 39

Figure 3.1 Accountability Mechanisms in Presidential Democratic Systems

Horizontal Accountabilty

Attorney General/ Public Prosecutor Supreme Audit Human Rights/ Institutions Public Ombudsman

Executive

Legislative Judiciary

Vertical Accountabilty

Elections

Citizens Civil Society Media Source: Payne et al. 2002. representation (Shugart et al. 2000). At the the legal norms, but they cannot exercise other extreme, weak party control of candi- this power except through a direct request dates (as in the case of Colombia and Brazil, (by a party with standing). Even when cases for example) leads to personalistic repre- are heard and ruled on, their ability to force sentation that, in many cases, contributes to compliance with their decisions tends to be “pork-barrel” congressional politics. As a limited, especially when directed against result, efforts to modernize legislatures have government itself (Henderson et al. 2004). shown so far limited impact (Mendez In the best of situations they are most effec- 2000). The implication is double. Legisla- tive in providing relief to individuals, pun- tures remain weak to oversee actions by the ishing rule violators, and determining the executive branch. They also remain a weak legality of actions and the constitutionality actor in terms of promoting legal reforms of laws. Since resorting to judicial means is oriented to improving incentives for more costly, this “reactive” characteristic of the effective service delivery systems. courts has tended to beneﬁt those who can The judiciary is another key actor in the afford judicial recourses, that is, the rela- presidential system of checks and balances tively well-off. Since the rulings typically adopted in Latin America. Still a rarity in apply to individual cases one-by-one, even the region is an effective judiciary that bal- favorable rulings for upholding citizen ances independence from undue political rights to public services, for example, tend interference with external accountability for not to beneﬁt the majority poor who have its decisions. Courts, by design, are the least not participated in the particular court accountable to external actors for their cases. Courts can hold political leaders actions, and thus the least proactive among accountable for breaches of the law, but it is the three branches. Their role is to encour- rare that they can have much effect in forc- age rule-based behavior by resolving con- ing them to do what they will not or cannot crete cases with a consistent application of undertake. 40 Citizens, Politicians, and Providers: The Latin American Experience with Service Delivery Reform

Against this general backdrop, judicial having independence is hardly sufficient to independence remains the subject of have well-functioning courts. Yet despite intense criticism and controversy in the reform efforts, such as court modernization region (Galindo 2003). More activist courts programs, independence of the courts that have taken an atypically aggressive role remains questionable in most Latin Ameri- in reviewing government policies and can countries. Many countries in the region demanding provision of second and third still experience limited capacity to provide generation rights to aggrieved citizens have equal access to justice to all citizens been accused of unwarranted meddling in (Buchanan 2001; Garro 1999; Pinheiro executive affairs and even a sort of judicial 1999; O’Donnell 2001). Historical legacies populism (Wilson and Handberg 1998). In of arbitrary behavior by the state and the a few countries where the judiciaries have ability of powerful elites to operate above or high degrees of institutional independence outside the law contribute to negative per- from the executive (for example, Brazil, ceptions of the justice system (Domingo Colombia), judicial decisions have some- and Sieder 2001). As a result, courts tend times struck down important policy mea- not to play an active role as enforcers of sures with effects that may have been wel- state accountability vis-à-vis citizens, par- fare-enhancing for the general population ticularly poor ones (Correa Sutil 1999; (for example, by curtailing legally sanc- Pasara 2002; Dodson and Jackson 2000). tioned privileges and by imposing unaf- In the last 20 years, legal and constitu- fordable fiscal costs on the government, tional reforms across Latin America have both current and future). Courts sticking to generated quasi-autonomous institutions a more conservative definition of their that monitor the exercise of public author- functions, and often with less political inde- ity. The creation of an ombudsman office pendence, have been charged with an (“Defensor del Pueblo”) has probably been antipoor bias and an overly formalistic one of the most successful efforts to com- approach to their work (Buchanan 2001; pensate for weak judiciaries. In 1985, Garro 1999; Pinheiro 1999; O’Donnell Guatemala was the first Latin American 2001). Both types have been accused of cor- country to create an ombudsman’s office, ruption, self-interested rulings, and extra- and in the 1990s, many countries followed ordinary inefficiency. From the well-docu- suit. 37 These offices investigate complaints mented case of Peru in the late 1990s to the over human rights, environmental protec- ongoing frequent accusations of Venezuela tion, freedom of the press, and elections. today, the judiciaries that lack indepen- Although institutional design varies from dence from the executive branch are seen as country to country, this office’s decisions or threats to the integrity of democratic insti- recommendations are never binding. In tutions and the rule of law. Methods used to most countries, the legislative branch deprive the judiciary of institutional inde- appoints the ombudsman, though the pendence vary from outright political ombudsman operates as an independent purges of judges (for example, Bolivia went entity. But, in most countries, ombudsman through 16 massive judiciary purges offices have very limited (legal and other- between 1936 and 1982), appointment of wise) capacity to enforce their findings temporary judges whom the executive can (Shugart et al. 2000) and their effectiveness remove arbitrarily as in Peru and Venezuela depends mostly on political dynamics and at the time of this writing, or blatant use of not on legal status.38 partisan criteria for appointing and remov- Audit institutions (such as comptroller’s ing judges as in Honduras, where until and inspector general offices) oversee bud- recently, the Supreme Court and the major- geting and expenditure of public funds. ity of the bench was renewed at each gov- Recent reforms have often involved changes ernment change. from simply publicizing budget figures to In many countries, establishing a mod- evaluating efficiency or cost-effectiveness. icum of judicial independence is a priority Public auditors have also gained increased for assuring a minimum rule of law, though institutional autonomy. Prosecutorial or LAC Service Delivery: Assessing Two Decades of Change 41

investigative agencies (often the Ministerio this challenge is exacerbated by Latin Amer- Público or Procuraduría General) fulfill sev- ica’s sociopolitical conditions that have lim- eral roles, but the most important for ited the scope for effective and equitable democratic accountability has been their public actions. Widespread elections and increasing role overseeing prominent gov- political decentralization are important ernment corruption investigations. In both radical changes that have taken place in cases, there are indications of weak capaci- Latin America over the last two decades. In ties in these institutions (Payne et al. 2002). particular, they have created opportunities Decentralization has moved issues down for creativity, innovation, and experimenta- to the local level, adding a new layer of tion in country after country. But patron- political relations that influences opportu- age, elite rule, and group-based exclusion nities and capacities for citizens to influence appear to be resilient. While elections can policy makers. Seemingly dramatic transi- be established overnight, rule of law and tions to new political and social equilibria representative institutions require changes have occurred in some cities or municipali- in political culture and habits that are ties, leading some observers to see more harder to change. Ultimately, whether and potential for rapid change at local levels how the service delivery institutions will be (Campbell 2003). When this happens, bet- transformed as a result of the transitions ter service delivery systems typically from authoritarianism to democracy result.39 However, it is still unclear how depends on the strength of mechanisms for widespread the benefits of local level vertical (electoral) and horizontal (within democratization, as exemplified in these state) as well as social accountability. cases, will be throughout the region. As the case of Bolivia previously discussed (see A virtuous circle may be in the above) indicates, even a sharp national reform as Popular Participation may have making, whereby social limited impact on political influence of the accountability mechanisms enable poor, depending on local context. The over- more effective state actions in sight and participatory instruments it service delivery introduced—that is, the vigilance commit- The mixed record of the new democracies in tees and the constructive vote of censure— transforming the historical patterns of sometimes became attractive targets for unequal and ineffective service delivery sys- political parties and interest groups, thereby tems may be one of the reasons behind the weakening accountability and transparency. growing uneasiness with political institu- In those localities where the traditional par- tions identified in opinion polls throughout ties remained strong (and civic associa- the region [United Nations Development tional activities weak), vigilance committees Programme (PNUD) 2004]. The “discon- tended to function less effectively either tent” with the observed weak state account- because the latter were co-opted by the tra- ability mechanisms has led, in some cases, to ditional elites or because the municipal conflict and political instability, which in governments and the vigilance committees extreme circumstances may reduce even fur- became arenas for open conflict. The con- ther the state’s ability to guarantee mini- structive vote of censure, instead of being mum service standards even in basic ser- used as an ultimate means of holding may- vices. At the same time, there is some ors electorally accountable for their perfor- evidence that the actions of organized citi- mance, was often abused as a tool of parti- zens (for example, monitoring and exposing san conflict. wrongdoing by the state) can trigger positive The “insufficiency of democracy” for changes in government actions. In other service delivery for the poor is partly a words, the tension created by weak state reflection of the intrinsic complexity of accountability does not necessarily mean shaping institutions for effective services— there is a zero (or worse, negative) sum bal- an issue that is on the agenda even in rich ance of power between state and citizens. societies with relatively efficient states. But Rather, when civic actions trigger reactions 42 Citizens, Politicians, and Providers: The Latin American Experience with Service Delivery Reform

still rather unstable (Garretón 2002). For BOX 3.10 ¿Cómo Vamos? Report Cards, Scorecards, and example, human rights organizations and Citizen Monitoring of Service Quality civic networks that monitor elections to prevent fraud or police actions to prevent abuse A report card is a survey soliciting citizen evaluating those services? The resulting sys- and violence have emerged as important feedback on the performance of public ser- tem, Bogotá Como Vamos, has collected sur- vices. Cards can provide feedback from users vey data from citizens on the coverage and actors in countries such as Argentina, Brazil, on: availability, reliability, and coverage of quality of health, education, security, and Chile, Colombia, Mexico, and Peru (Peruz- services; user satisfaction with service qual- other services.The data have been collected zotti and Smulovitz 2001). The development ity; responsiveness of service providers to both from an annual survey of citizens and of the public interest law movement problems; prevalence of hidden costs like city administrators, and from focus groups corruption; and user willingness to pay. including citizens from all socioeconomic (McClymont and Golub 2000) is another Report cards enable citizens to signal strata. El Tiempo and the local city television good example of how the emergence of civil needed reforms to public agencies and station disseminate survey results—each society organizations—in this case, those politicians and may exact accountability medium reaches about 3 million people— through media coverage and civil society and Cómo Vamos sends a quarterly bulletin that demand due process and “proper” pro- advocacy of service quality. Although their to 3,000 citizen organizations, libraries, cedures—is affecting accountability mecha- impact has not been rigorously evaluated, research centers, and universities.The proj- nisms. A free(er) and more inquisitive press they show potential for increasing account- ect costs about $75,000 annually—half for (Peruzzotti and Smulovitz 2001) and inde- ability and improving service quality. office and staff costs and the remainder for Report cards were first implemented in obtaining and disseminating data.The pro- pendent think tanks are additional facts in Bangalore, India, as a response to growing gram’s main visible effects have been on postauthoritarian Latin America. demand on essential public services— publicizing information. This more active civil society has pro- demand that service agencies could not Colombia’s secretary of education now vided the basis for what has come to be effectively meet. Beginning in 1993, Banga- presents performance data collected by lore gathered citizen feedback through a Cómo Vamos on the secretary’s public Web known as social accountability (Smulovitz citywide survey.That feedback pushed site. Cómo Vamos also published a citizens’ and Peruzzotti 2000): the actions by a mul- providers to improve the quality and avail- guide for the 2000 mayoral election with a tiple array of citizens’ associations and ability of public services. It also empowered summary on trends and achievements. movements and the media to monitor gov- citizens to directly evaluate civil servants Citizen report cards are being planned and hence altered the relationship between in Argentina, Bolivia, Ecuador, Honduras, ernment actions and expose wrongdoing. citizens and local government. and Mexico. A related tool is community The exercise of this type of accountability LAC Report Cards.Peru was the first Latin scorecards, which focus on monitoring per- has typically followed three complementary American country to implement report formance of services at the facility or com- strategies (Peruzzotti and Smulovitz 2001): cards to assess national nutrition, munity level—the level of the school, health 40 education, health, and employment post, treatment plant, or other service facil- legal approaches (involving the courts), programs. Bogotá’s Cómo Vamos initiative ity.The scorecard process includes an inter- social mobilization, and mediatic—the lat- soon followed in Colombia. In 1997, leaders face meeting between service providers ter two as a means of calling attention to from Bogotá’s El Tiempo newspaper, the and community members that allows for Corona Foundation, and Chamber of Com- immediate feedback and hence has the public wrongdoing. In all cases, these merce decided to found a reporting system potential to be a strong instrument for actions seek to activate the operation of on public services.The logic was simple: if empowerment. Scorecards will be intrastate accountability mechanisms citizens paid taxes to support local services, implemented in 12 municipalities in directly (that is, by decisions of the relevant why should citizens not play a direct role in Nicaragua, among other locations. state control agencies, including the legal system) or indirectly by pressure operating through the political system. In other words, civic action in response to state fail- of the more conventional vertical, and par- ures appears to be directly oriented to ticularly intrastate, accountability mecha- encouraging responses by formal state nisms, a synergic process may take place. accountability mechanisms. Partly enabled by the democratization Of particular interest and relevance for process, the last two decades have seen rising service delivery systems are a growing num- activity by civil society actors outside the ber of cases of civil society organizations state, from NGOs to identity-based move- involved in participatory monitoring and ments (notably amongst indigenous evaluation of service delivery (for example, groups). Classical social actors (those exist- through the use of report cards as exempli- ing before the new wave of democratiza- fied in box 3.10, above), public policies, and tions) appear to have lost some of their budgets as well as of the transparency in the social significance and tended to corpora- performance of public servants—a good tize, while new ones have appeared but are sample of which were surveyed in World LAC Service Delivery: Assessing Two Decades of Change 43

Bank (2004).41 This has sometimes been and capable organizations of civil society associated with efforts by national or local and formally established state institutions governments to open up government deci- (see World Bank 2003b). In other words, civil sion making to citizen and civic organiza- society pressures, demands, and actions mat- tions in response to major governability ter the most when they enable and empower crises (see box 3.11 on Peru). In some coun- the state’s own accountability mechanisms tries, regions, and cities, efforts to open up (Fox 2000). government decisions to citizens and civic The latter point is well illustrated by the organizations—unheard of until relatively experience and difﬁculties encountered by recently—are emerging in very different civil society in its interaction with the regu- settings. latory and control agencies set up by the Overall, these various cases reinforce the state in the case of infrastructure services point that the success and sustainability of [see Asociación de Entes Reguladores de experiences of social accountability are heav- Agua Potable y Saneamiento de las Améri- ily dependent on the development of syner- cas (ADERASA) forthcoming; Collado gistic relationships between autonomous 2003]. Under the reforms implemented

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

The problem. In 2002, 34 percent of Peruvians citizens and NGOs to design agency • Transparent financial Accounting for pub- listed corruption as the biggest national prob- budgets as part of Peru’s decentraliza- lic expenditures: Peru required national lem, making it third after unemployment and tion process.This process has involved and subnational governments to record poverty. Most Peruvians expressed distrust of citizens in the prioritization and alloca- expenditures through a transparent political parties (81 percent), Congress (73 per- tion of local investments and the design financial management system (SIAF) cent), and the police (66 percent). Furthermore, of local development projects that that regularly compares public spending Peruvians described receiving poor service from municipal budgets finance. Regulations to budgeting and physical goals. For municipalities (19 percent), the police (33 per- and a manual on participatory budget social programs, the information has cent), and judges (42 percent). formulation were issued, and initiatives details down to the district level.The Peru’s Response: Building Accountability. to improve quality of participation are system is transparent and open to all cit- In response to the lack of confidence in demo- under way. izens through the Internet, so that any- cratic institutions, Peru instituted a variety of • The National Agreement: Upon arriving in one can monitor spending. In 2003, the mechanisms seeking to improve government power in July 2001, newly elected Presi- site was made user-friendly to encour- operations and government accountability. dent Alejandro Toledo appealed to all age its use by civil society through a portal called “Consulta Amigable.” • National Roundtable for Poverty Allevia- political parties represented in the tion: In January 2001, the transitional gov- National Congress and civil society to • Report cards on social programs: Peru’s ernment of President Valentin Paniagua design a “National Agreement”of long- National Household Survey institutionalized a roundtable of intellec- term policies.The agreement, consisting incorporated questions that capture tuals, government agencies, local govern- of 30 state policies in a variety of areas, information about users’satisfaction ments, representatives of civil society, reli- was signed in September 2002; a matrix with health and food programs. Repre- gious groups, and international of indicators was also developed.In May sentative regional level data gathered representatives to develop a consensus 2004, a decision was made to establish a through the survey will be available to about social policies as well as to provide monitoring system to track social policy the public in 2005.The information will oversight to their implementation. implementation.The system involves be used by the national monitoring sys- both government and civil society actors. tem, which tracks the quality of public • Local poverty alleviation roundtables. The expenditures. National Roundtable helped organize • Freedom of Public Information Law. In district- and provincial-level roundtables 2003, Peru passed a Law on Trans- These social accountability mechanisms of public officials and civil society repre- parency and Access to Information, sought to improve the quality of public services, sentatives to cooperatively design local allowing citizens and public officials to and at the same time, to give legitimacy to a social programs and strategic develop- access information on government democratic government by involving citizens in ment plans. By 2004, over 1,000 roundta- operations at all levels.There is a clear program design, implementation, and monitor- bles were in operation. need to improve the government effi- ing. However, given that the reforms have only ciency in responding to an exponential • Participatory budgeting: Since the recently occurred, the extent of their effects will increase in demand for information from November 2002 regional and municipal not be clear for some time. citizens, political parties, and civil society elections, all subnational governments organizations. have been required by law to work with Source: Felicio and John-Abraham 2004. 44 Citizens, Politicians, and Providers: The Latin American Experience with Service Delivery Reform

during the 1990s, the regulatory agency is A further concern is the frequent prox- intended to act as an arbitrator between the imity between consumer associations and conflicting interests of consumers and ser- political parties, which may potentially vice providers. However, for this model to undermine their ability to act as indepen- function effectively, it is necessary for con- dent representatives of consumer interests. sumers to have a spokesperson that can In a number of countries, the creation of an articulate their interests, and counterbal- ombudsman (“Defensor del Pueblo”), as an ance the industry perspective, which is usu- independent publicly funded spokesperson ally very strongly represented by the service for consumer interests, has helped to com- providers. Without this countervailing pensate for some of these problems. A lack influence, there may be a significant risk of of tradition in consumer activism remains, regulatory capture by industry interests. In nonetheless, a bottleneck particularly for other words, without social accountability, poorer and subordinate groups. For exam- formal “horizontal accountability” mecha- ple, a case study of the Bolivian experience nisms are unlikely to be fully effective, at with complaint mechanisms after privatiza- least not in a systematic manner. tion found that there were at least three In Latin America, there was considerable times more complaints per 1,000 customers growth in the number of consumer organi- originating from the relatively higher zations during the 1990s. However, in many income city of La Paz than from its poorer countries their presence is still incipient, twin city of El Alto, both served by the same and in general, they suffer from a shortage utility companies (Foster and Irusta 2003). of funding, which in turn limits their ability Overall, the evidence reviewed presents to access the human resources they need to us with a “glass is half-full, glass is half- engage in complex regulatory debates. Most empty” situation. Clearly, the political countries in the region continue to lack the transformations of the last two decades legal framework that would give regulators have positioned countries in Latin America a clear basis for interacting with civil soci- much better to enhance the citizen voice’s ety.42 There are very few examples where the role and with it the strength of the “long regulator has established a permanent route” to greater accountability. At the same channel of interaction with civil society43 time, a long history of weak states and social and in those cases in which such interaction and economic inequalities presents itself is formalized, ensuring the independence as a heavy burden in most countries. There and sustained credibility of both consumer are signs suggesting that stronger social representatives and regulators remains a accountability mechanisms—like the ones challenge. Following the U.S. model, public described above—create pressures that hearings to consult on major regulatory could trigger the operation of horizontal decisions are already widely used in a num- (intrastate) accountability mechanisms ber of countries including Argentina, Brazil, and, more broadly, pressures for reforming Colombia, Costa Rica, and Peru. In coun- the state. If sustained, these changes could tries such as Bolivia and Panama, more further enable civil society (and particularly informal public consultation meetings are the poor) to actively and productively held. However, consumer associations often engage in influencing policies and out- complain that regulators do not provide comes. As suggested by Fox (2000), this type accessible information sufficiently far ahead of virtuous circle requires rejecting the still to allow them to prepare for these events.44 widely held assumption that state and soci- Moreover, public hearings are sometimes ety are necessarily engaged in a zero-sum perceived as a formality that does not ulti- balance of power. This points us in the mately affect the decisions that are taken. direction of complementary efforts to Furthermore, there is little evidence that, enhance various accountability mecha- when held, such hearings may have led to nisms: long- and short-route approaches inquiries or actions by, for example, the rel- complementing and reinforcing each other. evant congressional committee—a key The next chapter explores this idea in more force for accountability in the U.S. model. detail. LAC Service Delivery: Assessing Two Decades of Change 45

Notes regional rather than municipal level, and a number of Central American countries 1. For a succinct historical account of public sec- (Panama and Nicaragua), which still retain sin- tor reforms in Latin America, see Spink (1999). gle national providers. In Brazil, the decentral- 2. Better financial management systems, however, ization was from the state to the municipal have proven rather powerless in controlling level and is still only partial. questionable expenditure management when 8. The decentralization of education services, for powerful officials are determined to circum- example, presents a variety of experiences. A vent rules of good financial management. Such first group of countries (Argentina, Brazil, was the case in Peru in the 1990s, when the Chile, and Mexico) placed a subnational level executive branch circumvented its effects on of government at the center of the decentral- internal control by resorting to various infor- ization process. A second group of countries mal means ranging from highly centralized (most notably Bolivia and Colombia) have cash management to the decision to leave cer- operated, de facto, with a system of shared tain “political” expenditures outside SIAF cov- responsibilities between local, regional, and erage’s scope and the use of presidential national governments. More specifically, in the decrees to modify budgets with minimal con- case of both countries, the management of gressional oversight. See World Bank 2001a. human resources was attributed primarily to 3. The extent of union power depends on both the the regional level and the management of economic and political markets. In areas with school infrastructure primarily to the munici- less competition—education more than health, pal level, with the responsibility for the plan- old-style public monopolies, and national oil ning, pedagogical, and curricular aspects of the companies—workers have more leverage. educational process shared mostly across the central/regional/local level. A third group of 4. In the case of the Minas Gerais reform ana- countries (principally in Central America) lyzed by Grindle and Mason (2004), there were have decentralized responsibilities to the a number of factors leading to a more positive school level, adopting an approach of school- role of teacher unions than in most of the other based management with little influence of sub- cases she studied. In particular, the union was national governments. supportive of the competitive selection of school principals and of introducing broader 9. We use the term “local government” loosely community participation, as it counteracted here to refer to a broad range of subnational previous abusive employment practices by state actors. clientelistic local politicians. 10. In a number of countries including Argentina, 5. Among the reasons cited for the positive role Bolivia, Brazil, Colombia, Ecuador, Peru, and the Chilean teacher’s union played is its rela- Venezuela, conflictive relations and political tively high degree of professional and technical rivalries between the national and subnational competence, related to its reliance on its pro- governments have reduced the scope for effec- fessional reputation for its legitimacy. Com- tive intergovernmental coordination. The pared to most teacher unions in the region that essentially political nature of these relations tend to concentrate their focus on bread-and- has also made it difficult for these countries to butter issues of remuneration and work condi- resolve the perennial problem of vertical tions, the Chilean teacher’s union openly and imbalance in resource and responsibility systematically engages in serious and consid- assignments across government levels, as re- ered debates on education policy and is thus a assignment of resources and responsibilities respected voice in Chile’s education policy net- often implied zero-sum games between com- work (Crouch 2004). peting political actors. See, for example, Abru- cio (1999) for a thesis on “competitive and 6. It should be recognized that conventional predatory federalism” in Brazil and, similarly, “command and control” service delivery Tommasi et al. (2001) for the case of Argentina. approaches do work in the case of specific ser- 11. In many cases, only a small minority of munic- vices, whereby informational problems are ipalities have taken any measures to modernize limited and social demands for immediate their water utilities, whether by private sector results strong. A good example has been polio participation or by corporatizing the public eradication in LAC (Olive et al. 1997). provider. This has led to situations where the 7. The only exceptions to this general pattern are national regulatory agency applies modern Chile, which decentralized water to the incentive-based regulation to utilities that are 46 Citizens, Politicians, and Providers: The Latin American Experience with Service Delivery Reform

still managed along traditional “clientilistic” ever, for the water sector, the improvement is lines and have no real incentive to meet regula- more modest. There have also been significant tory targets. In extreme cases, there may even improvements in the continuity of service, par- be jurisdiction conflicts. For example, Peru’s ticularly in the electricity sector. national regulator found that when it autho- 15. McKenzie and Mookherjee (2002) find that the rized municipal water companies to increase increase in access to water following private their tariffs towards cost-reflective levels, in 30 sector participation in La Paz and El Alto in percent of cases the boards of these companies Bolivia was significantly higher than the previ- voted to keep their tariffs below the regulatory ous historical trend for those cities and to the price cap for political reasons. trend observed in nonprivatized cities over the 12. Brazil and Mexico have federal banks (Caixa same period. Moreover, a recent study from and Banobras), which to some extent condi- Argentina showed that (controlling for other tion access to credit against financial indica- relevant factors) municipalities that had priva- tors. Ecuador is currently developing a mecha- tized their water services had significantly nism that would allocate central government lower infant mortality rates than those that capital transfers across municipalities based on retained the service in public hands, the main a formula that captures the level of financial explanation being improved service quality and institutional performance of the utilities, (Galiani et al. 2002). taking poverty levels into account. A similar 16. Foster (2002) found that the typical period of scheme has been proposed for Bolivia. tenure of water regulators was two to four 13. In countries where decentralization has taken years, compared with five- to six-year legal place, there are a number of cases where local terms. This kind of evidence illustrates that it is mayors have played a pivotal role in water sec- difficult to legislate against the prevailing polit- tor reform processes, such as Cartagena in ical culture and that regulators may come Colombia. In countries that have thus far under pressure to “resign voluntarily” long retained a centralized service provider, such as before their legal term has expired, particularly Honduras and the Dominican Republic, there following changes of administration. Never- has been an increasing tendency for individual theless, while two- to four-year terms may fall municipalities to opt out in the hope of being short of expectations, they still represent a able to provide a better service at the local level. greater degree of stability than is the case for An interesting example is the city of Puerto sector ministers or public utility managers, Cortes on the Atlantic Coast of Honduras; fol- who typically have a higher turnover rate. lowing a hurricane that interrupted the city’s 17. It should be emphasized that responsiveness to water supply, the mayor succeeded in negotiat- “communities” is not always, or not necessarily, ing a temporary transfer of operation and the same as responsiveness to the articulated maintenance responsibilities from the national preferences of poor groups in the population. water utility and reached an agreement with “Communities” are heterogeneous and embed- the population that tariffs could be raised as ded in unequal social and political structures. long as services improved. Thereafter, a full There is undoubtedly substantial variation in transfer of responsibilities to the municipality the ways in which different social funds engage was achieved by means of a congressional with (different groups in) local communities. decree, and the utility was converted into a limited company with mixed municipal and work- 18. While Bolsa Escola in Brazil was probably the force ownership. first in the region, Oportunidades (previously PROGRESA) in Mexico has become the most 14. Comparing the decade leading up to reform to famous example (with similar schemes in the decade that followed reform, the expansion poorer countries, including Honduras and rate of service connections increased by a factor Nicaragua). of two for water and sewerage services, a factor of four for fixed-line telephony, and a factor of 19. The basic alteration of traditional accountabil- 27 for cellular telephony, while it declined ity relationships described above is comple- slightly for electricity. A similar comparison for mented by a series of innovations that CCT labor productivity shows that the improvement programs have adopted to address many of the rate of this index increased by almost threefold shortcomings of traditional social assistance for electricity, by more than fivefold in the programs: addressing both current and future fixed-line telephony sector, and by more than poverty, more sophisticated systems for target- tenfold in the cellular telephony sector; how- ing the poor, providing cash, fostering syner- LAC Service Delivery: Assessing Two Decades of Change 47

gies in human development, and using evalua- mentation on the financial protection of, and tions strategically. service use by, the poor (Massad 1996; Baeza and Cabezas 1998; Bitran et al. 2000; Vega et al. 20. This is related to concerns about the geographi- 2003). It should be noted that, particularly dur- cal selection criteria that most CCT programs ing the 1990s, increased targeted public financ- applied, which can leave out poor areas with ing for FONASA resulted in significant, pro- limited health and education supply capacity or gressive anti-poverty effects of public health limited financial infrastructure. For a discus- expenditures. sion of some of the targeting concerns in the specific case of Oportunidades-PROGRESA, 28. Over 20 years, water services have been pro- see Adato (2000) and Adato, de la Brière, Min- vided to 1.3 million rural communities, there- dek, and Quisumbing (2000). by reaching full coverage of the rural concentrated population. 21. To address many of these concerns, CCT programs have set up channels through groups of 29. In Honduras, for example, insufficient capacity mothers’ representatives that act as informa- in the Ministry of Education contributed to tion conduits between the centrally adminis- substantial delays in salary payments to tered program and the local beneficiaries. This autonomous schools and teachers being more system may have led to women’s empowerment absent for training needs (Di Gropello and and provided a communication channel with Marshall 2004). local communities, but it operates indepen- 30. Development of this emphasis was affected (not dently of elected local officials. without endogeneity) by many factors— 22. For a review of the institutional development including the rise of democratization through- impact of social funds, see Carvalho et al. out the region; the involvement of “community (2002). participation” specialists to help improve methods in social funds, but who had a tendency to 23. Tendler also argues that there are some struc- see community participation as a good in and tural features of many social funds that make of itself; and rising emphasis in the interna- them particular conducive to patronage—the tional development community on notions delivery of “obras,”the queues, or lists of appli- such as social capital, empowerment, and so on. cants to be chosen from, and the capacity to turn on the money in preelection periods. 31. In the case of Nicaragua, decentralization to the school level was also functional to the process of 24. These providers often provide the only realistic “de-Sandinization” that the Chamorro admin- short-term supply alternative for low-income istration initiated. See Grindle and Mason customers, yet they often operate in a climate (2004) and Kaufman and Nelson 2004. of informality. This limits the possibility of regulating these activities, for example, to 32. As, for example, in the case of Guatemala with ensure that safe drinking water is provided. It PRONADE, which was put under severe strain also complicates the formation of potentially when there was a change in government. helpful partnerships between formal utilities 33. Cuba has had exceptional results in delivery of and independent providers, whereby, for quality education and health services, under example, the utility provides a bulk supply of authoritarian auspices (World Bank 2003a)— potable water that independent providers then although possibly at extremely high costs in distribute to peri-urban areas. other dimensions of development. Chile under 25. Namely, the existence of soft budget constraints Pinochet, while presiding over distributional at the municipal level, restrictions on municipal worsening along some dimensions, did have decision making, and the lack of autonomy and areas of relatively effective public action for the accountability of municipal schools. poor (Sen and Drèze 1989). Both were building on previous histories of relatively effective 26. See Di Gropello (2002) for the negative impact social development. of the teacher statute and Sapelli and Vial 34. In Brazil, for example, the democratic transi- (2002) for the negative impact of central inter- tion involved as much as anything a failure to ventions on competition through the implied resolve distributional conflicts—with the 1988 differences in per-capita budget between the constitution reflecting such a multiplicity of two types of schools. pressures. Expansion of debt and an unusual 27. There is still significant controversy about both capacity to raise taxes (by Latin American stan- the sources and long-term impact of this seg- dards) instead supported a mix of measures 48 Citizens, Politicians, and Providers: The Latin American Experience with Service Delivery Reform

that sought to provide something to all groups, ing their associational autonomy and weaken- including expanded education for the poor. ing civil society. But this was done without tackling the multi- 40. Both the Brazilian (1988) and Colombian ple areas of opportunity-hoarding or vested (1991) constitutions introduced legal figures interest in the society. Highly subsidized col- (Acciones de tutela and Acao Direta de Inconsti- lege education has traditionally been one of tutionalidade) that allow for citizens’ legal these elite preserves in Brazil, and the numbers actions in response to failures to enforce rights. indicate a sharp rise in per student spending. 41. Some innovations have targeted the typically 35. Recent work that tries to model clientelism’s hard-to-reach legislatures (Mendez 2000). A effects on policy outcomes under different good example is the case of Poder Ciudadano in assumptions all point to the same conclusion: Argentina, with the creation of databases with clientelism results in less than optimal levels of the voting records of members of Congress, public goods production (Keefer 2002; Robin- which has been taken up by civil society orga- son and Verdier 2002; Ames 2003). nizations in Colombia, Dominican Republic, 36. An example of budget management with exec- Ecuador, Guatemala, Mexico, and Panama. utive discretion is Peru in the 1990s. Between 42. Relatively little attention was given to con- 1994 and 2001, the Executive issued 748 sumer issues in the original regulatory frame- urgency decrees, 27 percent of which directly works. An important reason for this is that, amended the budget and 41 percent had a clear with the exception of Brazil, utility reform effect on the budget or public finances (World predates the establishment of broader con- Bank 2001a). An example of substantive legis- sumer defense legislation. In a number of lation affecting budget allocation and even fis- countries (such as Argentina, Colombia, Costa cal sustainability was the passage of a teacher Rica, and Peru), consumer defense legislation statute in Honduras in the late 1990s; this was introduced a few years after utility reform, granted generous salary increases and other obliging regulators to review their current employment benefits without corresponding practices in this area, as well as providing them measures to enhance teacher performance. with new legal instruments to facilitate civil 37. Ombudsman offices were established in society participation. Argentina, Bolivia, Brazil, Chile, Colombia, 43. An interesting exception is the case of ETOSS, Costa Rica, Ecuador, El Salvador, Guatemala, the water regulator in Buenos Aires, which in Honduras, Mexico, Panama, Peru, and 1999 established a Consumers Commission Venezuela (Smulovitz and Peruzzotti 2000). that was made up of one representative from 38. In Peru, for example, Alberto Fujimori’s gov- each consumer association active in the water ernment blocked the Peruvian legislature from and sanitation sector. ETOSS funds the com- acting on the Ombudsman’s recommendation mission and the commission has the opportu- to prosecute actions of the armed forces and nity to review and comment on all of the major intelligence organizations. decisions that the Board of Directors makes. In a few isolated cases, such as the state regulator 39. For example, Porto Alegre in Brazil is a case of in Rio Grande do Sul in Brazil, the provincial an apparent transition in which the participa- regulator in Cordoba, Argentina, and (for a tory budgeting “instrument” played a role. The brief period) the telecom regulator in Peru, design of the participatory budgeting process— consumer associations have been invited to by which delegates are allocated to the budget join the board of directors of the regulatory council in proportion to the participation level agencies. of associations in open forums—has stimulated associational activity. It has also had the effect 44. In response to these criticisms, Costa Rica has of promoting horizontal coalition building, in introduced a consumer advisor in the regulatory contrast to the past when neighborhoods agency staff; the advisor’s role is to help con- processed demands vertically through political sumer groups analyze and understand the infor- patrons with the effect of severely compromis- mation issued ahead of each public audience. Conclusions

Over the last two decades Latin American These transformations created opportu- countries have seen some radical transfor- nities for change in the accountability mations in the state’s organization and struc- mechanisms that link policy makers, citi- chapter ture as well as in the relationship between the zens, users of services, and providers. Figure state, citizens, and markets. First, the region 4.1 graphically illustrates some of these has moved from being a set of countries changes. The accountability relationship where democracy was the exception, to one between citizens and policy makers has 4 in which it is the rule. Second, LAC govern- been altered by the generalization of ments have become significantly more national elections (and by the various decentralized in political, administrative, mechanisms of intrastate controls implied and fiscal terms. Third, there was a process of by constitutional rule) and the strengthen- redefining the state’s role that involved ing of civic rights; it has also been altered by increased roles of private service providers the introduction of local elections (and the and a movement toward enhancing the role increased role of local governments enabled of competition in many areas of economic by administrative decentralization) and the activity, including the provision of key ser- growth and increased activity of civil soci- vices. Fourth, there has been rising activity ety organizations acting as watchdogs and by civil society actors, both in services deliv- intermediaries between citizens and politi- ery and as actors seeking different policies. cians. The evidence suggests that expanded

Figure 4.1 A Network of Inﬂuence and Accountability Mechanisms

National Public sector elections modernization NATIONAL POLICY MAKERS Administrative decentralization

Contracting out/ LOCAL privatization GOVERNMENTS Advocacy watchdogs Local elections INTERMEDIARY REGULATORY ORGANIZATIONS AGENCIES Civil society development

PROVIDERS CITIZENS Public & Private Competition and community-based management 49 50 Citizens, Politicians, and Providers: The Latin American Experience with Service Delivery Reform

political voice by citizens was an important across countries and localities depending factor in expanding service coverage under on political and social context. government auspices that took place over Using the WDR language (see box 4.1, the last two decades, particularly for social below), we can say that Latin American coun- services. tries have “experimented” with both long and The accountability relationship between short routes to accountability. Together, policy makers and providers has been democratization and state reform efforts affected both by contracting out services have created more opportunities for the long previously delivered by state bureaucracies route (citizens exercising control over and by changes in the way those bureaucra- providers through elected politicians or pol- cies operate, brought about by public sector icy makers) to operate well in practice. But modernization efforts. The accountability many weaknesses remain in the long route, relationship between providers and users particularly in terms of the citizen–policy has been transformed, in many cases, as a maker relationship, which in turn weakens result of the introduction of competition policy makers’ incentives to hold government and community management of services. and nongovernment providers accountable The evidence reviewed in chapter 3 shows for services delivered. Overall, short route that, when successfully implemented, mechanisms (clients making providers changes in accountability relations between accountable for results through the exercise policy makers, providers, and clients do of choice or through direct management of have an impact on the quantity and quality services) have, on balance, been successful in of services delivered. However, the evidence improving the delivery of many services, also shows that a particular service delivery albeit with signiﬁcant variation depending approach may have very different results on the local sociopolitical context. Moreover,

BOX 4.1 Lessons for Donors and International Organizations

How can the World Bank and others in the inter- change rather than on inputs.Tradition- to longer-term strategic goals; monitor national community contribute to improving ally, donors often set up parallel and evaluate progress toward reaching services for the poor? This work has identified processes to ensure effective outcomes those goals and use the information to three main areas for innovation: on isolated projects, but coordinated inform program and policy decisions. efforts to strengthen (not weaken) exist- 3. How to do it? Engage clients 1. What to do? Reform lending and analytical ing service delivery systems are key to Strengthen accountability and client “voice” work engaging clients in results-oriented ser- through access to information. Public Mainstream institutional analysis as a key vice delivery reforms. access to transparent information on component of the project cycle. Beyond 2. Tools that can help? Monitoring and government performance and, traditional economic and technical evaluation conversely, supporting government’s analyses, a direct implication of our Promote impact evaluations. Promoting access to public opinion are key to analysis is that “donors”would be well- structured learning from rigorous strengthening the accountability of pol- served by a better understanding of the impact evaluations of key interventions icy makers and service providers to political economy characterizing is critical to generating knowledge on beneficiaries. accountability relationships between development effectiveness—especially policy makers, provider organizations, Process-oriented,responsive client where this is integrated into front-line providers, and citizens. Project partnerships. There are no quick and mechanisms for a variety of forms of work would be more relevant and effec- unique fixes to service delivery debate on the results, from communities tive if donors were to systematically con- problems.Technical advice alone is no to technical seminars to parliaments or sider the incentives and individual coun- longer sufficient; instead, donors can congresses. Getting the basic material to try and sector contexts that affect increase their effectiveness by adopting feed such debates requires more invest- service delivery. Seemingly “neutral”or a partner role based on regular, open, ment in both evaluation design and “technical”analysis will often fail to take sensitive, and in-depth policy dialogue implementation, notably because of the account of the institutional with clients.This process-oriented (as public good aspects of these determinants of success or failure. opposed to a task-oriented) approach is evaluations, which remain likely to become more appropriate as More flexible lending instruments and opera- undersupplied without subsidization institutional reform issues become more tional practices are a precondition to Establish performance management systems important in moving forward the reform focusing on outcomes and institutional and key performance indicators linked agenda. Conclusions 51

the weaknesses encountered along the “long that the state delivers. But very few coun- route” may impose often serious bottle- tries in the region have been able to move necks even to apparently successful innova- decisively toward developing a results-ori- tions empowering service users, particularly ented public administration or achieve sig- poor ones. nificant depoliticization of the state appara- Looking forward, further progress is tus. Some “islands of efficiency” (such as unlikely to happen if governments keep some social funds) have been created in the relying only on increasing resources allo- public sector with consequent positive ser- cated to the provision of services and more vice delivery results. But their ability to sus- effective delivery systems are not developed. tain those results is often limited without This is the case particularly when fiscal con- reforms in the operation of “mainstream” strains limit the capacity to respond by sim- state agencies. Decentralization to subna- ply adding more resources and for services tional governments has changed structures, for which ensuring quality and relevance is but in most cases without radically chang- essential. It is in this context that the report ing this conclusion—at least not so far. stresses the importance of assessing the Many local governments have indeed extent to which service delivery systems are grabbed the opportunities created by today better positioned to address the many decentralization and introduced a variety of remaining challenges facing Latin American innovative service delivery approaches. But countries, particularly in terms of serving overlaps and gaps in responsibilities, inade- the poor. In this final chapter, we summa- quate or insufficient authorities and rize our assessment of how changes in key resources, few (enforceable) mechanisms to accountability relations that have occurred promote performance orientation, and the in Latin America since the early 1980s persistence of clientelistic relationships affected service delivery, and discuss the both at local and central levels (as well as implications from this assessment for between government levels) remain bottle- future reform efforts. In some cases this necks across the region. involves entering a more speculative terrain. As documented in chapter 3, the many experiments Latin American countries have attempted with alternatives to conventional Routes to Better Service public sector delivery have, overall, pro- Delivery duced positive results, though with a wide Overall, public bureaucracies remain a bot- variation, depending especially on the tleneck for effective service delivery in most sociopolitical and institutional context. countries. As shown in chapter 2 of the Contracting out to the private and NGO report, in most LAC countries governments sectors, promoting competition and choice, have increased the amount of resources and enabling consumer participation in the allocated to expanding coverage of basic management of services have been tried— services. Conventional, centralized public with some notable successes—in various sector service delivery (involving pure pub- sectors throughout the region. While the lic financing, delivered through public sec- empirical evidence is hardly conclusive, it is tor organizations and subject to standard fair to say that (i) the balance of evidence public employment practices) has played a between contracting out and service deliv- role in the substantial expansion in service ery in-house in the state is mostly in favor of coverage in many sectors and countries. But the first, under current conditions in most the weak quality of public LAC bureaucra- countries; (ii) the experience shows that cies has not shown much improvement. there is considerable scope for expanding This is despite the significant progress in private sector participation in service provi- state modernization reform that upgraded sion, provided this occurs in a sound regula- governments’ institutional capacities con- tory environment; (iii) giving individuals siderably, in some areas. It is likely that such choice over who they get their services from upgrading will have substantial effects over has generated, in the limited areas where it the longer term in the quality of services has been implemented, powerful pressures 52 Citizens, Politicians, and Providers: The Latin American Experience with Service Delivery Reform

for results on front-line providers; and (iv) among others. In the end, it is unrealistic direct client involvement in the manage- and unlikely that LAC countries can get ment of services can yield positive results away without both stronger state account- when the services involved allow a relatively ability and complementary action to create small production scale and rather simple effective core public bureaucracies, even management challenges. when deciding to move decisively in the There is often a temptation to try to direction of delegating major responsibili- “pick winners” and point our evaluative fin- ties for delivery to nonstate actors. ger in the direction of one service delivery The political transformations of the last approach—a silver bullet. Specifically, the two decades have positioned Latin Ameri- difficulties in reforming state institutions can countries much better to enhance state and the relative success in the various alter- accountability. At the same time, a long his- native delivery models reviewed above may tory of weak states and social and economic lead some to argue that the solution to Latin inequalities presents itself as a heavy burden America’s many remaining service delivery in most countries. While elections can be gaps resides solely in a scaling up of the established overnight, rule of law, an effec- “client power” and “contracting out or pri- tive system of checks and balances, and rep- vatization” experiences. While it is easy to resentative institutions require changes in become enthusiastic with the results of political culture and habits that are harder to individual positive experiments, the evi- change. This will no doubt require more cit- dence is too strong regarding the impor- izen influence on government decisions and tance of context to simply argue for scaling actions, including from poor and excluded up or replicating “best practices.” First, it is citizens. Stronger voice (as a precondition often hard to replicate the specific condi- for more accountable politicians) is needed tions that enabled the initiation and effec- not just to reform the state bureaucracy, but tive implementation of the innovative also to ensure that contracting out, choice, efforts reviewed in chapter 3 (be they PRO- and participation do benefit the poor. This GRESA/Oportunidades in Mexico, water does not mean that until and unless state privatization in La Paz and El Alto in accountability is fully ensured—that is, until Bolivia, or EDUCO schools in El Salvador). Latin American countries become fully How such reforms will be shaped and mature and consolidated democracies— implemented will depend crucially on the efforts to reform service delivery are bound social, political, and public sector context at to fail. On the contrary, there are many ways national and local levels. Put another way, in which short- and long-term goals can be the reasons for success often lie in the par- made complementary. To give one example, ticular ways that groups interact in each community participation approaches to ser- society, rather than the specifics of technical vice delivery (as in community-run schools design, however important that may be. in El Salvador) cannot only have good Second, in the long run most such innova- short-term results but, depending on how tions would, in the end, be ineffectual with- they are structured, can also help build indi- out reforms in the operation of “main- vidual, community, and state capacities over stream” state agencies that are more deeply the longer run. beholden to the fundamentals of interac- In the last analysis, using the WDR ter- tions between citizens, policy makers, and minology, the short route (clients exerting service delivery agencies (public or private). direct control over providers) should be In situations where the state’s capacity to conceived as a complement rather than a establish and enforce regulations, to priori- substitute for the long route to accountabil- tize and coordinate policy actions across ity (citizens exerting control over providers ministries, sectors, and government levels, through politicians), a potential shortcut and to establish and sustain compensatory that will only be consolidated if linked to mechanisms is weak, there are dangers to more fundamental changes. Moreover, our these routes—a lack of sustainability, new reading of the evidence makes us skeptical forms of capture, and social stratification, of an approach whereby reformers put all Conclusions 53

their eggs in one transformational basket. barriers in the citizen and policy maker Rather, we are inclined to support the relationship. As a result, in most countries notion of “strategic incrementalism” the case is weak for ambitious new public (World Bank 2003a): continued efforts to management–type reform measures being deepen societywide conditions for account- the solution to poor service delivery sys- ability (deepen democracy, citizen partici- tems over the short to medium term. pation, and state reform), while at the same This assessment, however, should not be time pursuing opportunities for specific construed as an argument against public service delivery reforms and program inno- administration modernization efforts. vations that push for choice, client involve- Rather, the implication is that Latin Ameri- ment in decision making, and perfor- can countries should indeed continue mance-based contracts both in and outside investing in general state capacities, but with the public sector. In other words, pragmatic stronger links to specific in-sector reforms incremental reforms while unlikely to fully and linked to service delivery results. Specif- address service delivery problems, can con- ically, we hypothesize that the likelihood tribute both to short-term results and to that investments in improving the quality of creating conditions for deeper and more administrative and financial systems will favorable change over time. In this context, have significant and/or lasting impact is complementarities will arise to the extent greater if and when they are linked to that such innovations are designed in ways reforms in specific service delivery systems. that contribute to strengthening both state Indeed, management reforms tied to better and individuals’ capacities as citizens and service delivery outcomes can create new consumers. Seen from this perspective, the constituencies and the consequent political experiences reviewed in this report provide support without which their sustainability valuable lessons for the design of alternative would be questioned. This hypothesis is cor- service delivery approaches. roborated by the relative success of administrative reforms tied to substantive policy reform agendas. Examples include financial Public sector management reforms: management reforms, such as the introduc- seeking stronger links with concrete tion of computerized financial management and observable service delivery information systems as a potent instrument improvements for expenditure control and thus broader The evidence suggests that reforming state fiscal adjustment efforts in the region, or the bureaucracies and developing a perfor- establishment of regulatory agencies as part mance orientation are naturally difficult of a broader privatization policy in several endeavors, from which quick results should countries in the region. Similar examples not be expected. The primary reasons for related to service delivery as such are more state reform failures are political and less so limited, but the same logic would apply. technical. Addressing the problems associ- In some circumstances, the strengthen- ated with weak traditions of bureaucratic ing of an “enclave” in the administration autonomy and the highly politicized man- may be required to achieve a specific service agement of government offices involves delivery goal. For example, this could be the touching on vested interests. But when the case when political or social conditions political system does not force accountabil- require quick results on the ground that ity mechanisms to function effectively, tech- mainstream agencies would be hard pressed nocratic approaches to state reforms are to achieve. Experience, though, suggests unlikely to produce lasting results. It is hard that the creation of islands of efficiency may to aspire to a Weberian bureaucracy in result in the perpetuation of parallel charge of service delivery when state bureaucracies with consequent efficiency accountability mechanisms are not fully losses. The response cannot be simply developed and functional. This represents a rejecting altogether the approach. Instead, major challenge for Latin America’s young what is required is facing the dilemma democracies that still experience too many between short-run results and long-term 54 Citizens, Politicians, and Providers: The Latin American Experience with Service Delivery Reform

institutional development, and designing handled in a more centralized manner. In upfront clear guidelines and strong incen- many countries, there is great heterogeneity tives for transitions from islands of effi- in the size of subnational entities (for exam- ciency to the mainstreaming of administra- ple, municipal governments ruling over tive reforms. large cities and very small localities with dispersed populations), which are in a very different position to deliver the various Service delivery decentralization: aspects of particular services. Considering creating better conditions for the limited experience with decentraliza- performance-orientation at all tion programs that follow asymmetric rules government levels among subnational entities with the same Decentralization is no panacea. It can be as legal status, addressing these heterogeneities susceptible to (local) elite capture and cor- may require the active monitoring and ruption as national service delivery. But it is a engagement of central authorities. political fact of life, and for many services, it The latter comment highlights another has the potential for both more effectively important lesson: making decentralization responding to local demands and fostering work is not a question of getting the blue- local structures of more equal and effective print right and then letting things happen. accountability. Moreover, under the right A more proactive role of central govern- conditions, decentralization can provide a ments and NGOs to improve information context for innovation and experimentation. flows between subnational entities and At the core of successful decentralization is a among citizens could contribute to combination of an improved compact strengthening the agency of the poor at the between government levels and strengthened local level—a key factor for decentralization local accountability to all citizens. Depending to materialize in welfare improvements for on the country and sector, this may require the poor. Central governments (and a range simpler and more transparent use of inter- of nongovernmental actors) can also governmental transfers, the enforcement of enhance the benefits from decentralized hard budget constraints for local govern- innovation by facilitating cross-jurisdic- ments, clear definitions of responsibilities tional learning and exchange of experi- (including through legal and/or administra- ences. At the same time, central govern- tive instruments), strengthening the capacity ments may often need to be proactive to of central government monitoring and help changing behaviors amongst local auditing, and developing information and actors, for example, through the use of evaluation systems. It may also require spe- financial and other incentives, such as con- cific intervention instruments (by national ditional matching grants, or even direct authorities or federal bodies) when service engagement in recruitment, in sharing delivery failures threaten the well-being of information, and in support for mobiliza- citizens. Last but not least, it requires attention. In the end, however, more time may be tion to the other side of empowerment: the needed for decentralization’s benefits to capacity of poorer groups to participate in materialize in a more generalized fashion. local decision making. Characteristics of different services make Delivery by nonstate providers: them more or less prone to local delivery, particularly depending on the presence of ensuring adequate regulatory and economies of scale. In fact, it is typically compensatory mechanisms are in specific aspects or functions in broad ser- place to increase service delivery vices that involve such economies of scale. effectiveness and equity To give just one example, while day-to-day There is solid evidence showing the positive operations of schools can be managed in a impact of reforms oriented to increasing very decentralized fashion, aspects such as the participation of nonstate providers (pri- curriculum design or student learning vate for-profit and not-for-profit organiza- assessment systems are more effectively tions) in service delivery, when the condi- Conclusions 55

tions are right. Privatization and contract- seeking to reverse them (as occurred in the ing out reforms can improve efficiency and Colombia education voucher case) and quality of service delivery, particularly ensuring their sustainability. when clients’ choice is enhanced. At the There are likely to be important differ- same time, privatization has sometimes led ences between services in the extent (and to increased concentrations of market (and nature) of potentially useful public interven- perhaps political) power in conglomerates tions. The more that a service is like a classic and has sometimes been associated with commodity, the more likely are choice and corruption. Not all forms of concessions competition models to work. This is most have proved to be sustainable. Some coun- true of cellular telephony, but becomes of tries have experienced difficulties in manag- rising difficulty in areas where market power ing contracts over time. Increased social is intrinsic (for example, local water), there stratification may be an undesirable out- are major informational asymmetries come of increased competition and choice (health, perhaps education), and significant when informational asymmetries are signif- probabilities that social, income, and icant and differences large in purchasing provider structures will lead to greater strati- power among consumers. The answer is not fication (health and education). In these to throw out the baby with the bath water, areas choice can play a role, but it is no as suggested in some critical analyses of pri- panacea: it is more likely to be a complement vate sector participation in service delivery, than substitute for state effectiveness— but rather to reap the lessons of experience whether in public sector service delivery or to improve the way that such arrangements effective regulation—and for stronger citizen are designed. voice, whether influencing policies or effec- First, it is essential to ensure that the tively representing consumer interests. benefits of private sector participation are equitably distributed between all the key Community-based service delivery stakeholders, namely, current customers, potential customers, employees, investors, approaches: seeking designs and the state. This is likely to require a com- that strengthen capacities of bination of regulations and subsidies, par- organizations of the poor while ticularly to ensure affordability among low- ensuring accountability for results income households. Second, greater efforts Overall, when market mechanisms are hard must be made to improve the efficacy of to develop—for example, due to a lack of regulatory agencies that are key to provid- competition—community-run services can ing a stable operating environment for con- be a powerful approach to generate results tractors and enforcing contract provisions (even in the short run) for services that do (including price and quality of services) to not necessarily involve large scales (and benefit all consumers. Third, there needs to thus both requiring sophisticated technical be greater transparency throughout the skills and involving significant collective entire process, from the transaction design, action challenges) and for which there are to the allocation of any resulting proceeds, not substantial asymmetries of information to the subsequent regulatory decisions that between clients and providers. But their affect the evolution of the contract. This long-term impact is likely to be the greatest will involve providing the public with better when they contribute to changing the local and more accessible information, establish- balance of power in favor of the poor and ing formal channels for citizen participa- disadvantaged—simply giving more influ- tion, and supporting the capacity of citizen- ence or resources to local communities based groups to interpret the policies and may, in itself, risk just pushing problems of practices of privatized firms that often elite capture, inefficiency, or patronage to a require technical analysis. Stronger voice by more local level. In the end, the key is citizens benefiting from choice and compe- whether and how they help empower the tition may be an important factor in pro- poor through both creating capacities for tecting such reforms from interest groups choice, aspiration, and organization (that is, 56 Citizens, Politicians, and Providers: The Latin American Experience with Service Delivery Reform

building agency) among the poor as well as effective social accountability is not about changing the opportunity structure where bypassing the formal channels of state they live to increase the probability of their accountability (elections, parliamentary voice and actions making a difference. How activity, independent courts, and so on) this is done will depend on social and polit- but, rather, about supporting them. ical context. In particular, to the extent that In that sense, investments in social the experience of running services leads to accountability mechanisms at the national the strengthening of associational auton- and local levels, focused on both overall omy among poor communities, even a nar- state performance and specific service deliv- row exercise of client power can contribute ery issues, can potentially have an extremely to an enhanced ability of citizens to hold high payoff. It is in this context that the governments (local and central) account- growing number of cases of civil society able for a broader set of actions and poli- organizations involved in participatory cies. At the same time, the government’s monitoring and evaluation of service deliv- role in the promotion of community-based ery (for example, through the use of report approaches may need to be more proactive cards), public policies, and budgets as well than simply “transferring resources” to as of the transparency in the performance groups of organized citizens. A solid “com- of public servants constitute a powerful pact” between the authorizing state author- instrument for change. ity and the empowered community groups In a region characterized by a long tradi- will typically be required to ensure the ful- tion of exclusionary public policies, the fillment of coverage, efficiency, and equity actual enforcement and application of rights targets, as illustrated in the case of EDUCO, appears as a key challenge for improving for example. This brings up again the com- state accountability, particularly vis-à-vis the plementarities between state modernization poor. The expansion of political rights asso- and client power approaches. ciated with the new wave of democratization offers a unique opportunity to expand civil Social accountability: seeking and social rights—particularly of the poor, the excluded, the discriminated against—an opportunities for synergistic essential step to counteract the current pat- relations to strengthen the capacity tern of policies against poverty and inequal- and accountability of state ity being captured and distorted by organizations ingrained clientelism and paternalism prac- Enhancing the accountability of policy tices. But the active and explicit enforcement makers to citizens requires more than regu- of rights could be an even more direct lar and clean elections. It is through the vir- accountability instrument when it is linked tuous interaction of electoral (national and to specific services. An example is the case of local), intrastate (legislatures, judiciaries, the recent efforts to establish guaranteed and control agencies), and social account- health insurance benefits in several Latin ability mechanisms that Latin American American countries. Although all countries countries are most likely to reverse the include the right of citizens to good health inherited pattern of clientelism, elitism, and and/or access to health services as a constitu- corruption characteristic of most of its tional mandate, until the mid-1990s, effec- states. In particular, given the difficult path tive instruments for the poor to demand to the strengthening of rule of law in coun- compliance by the state or other actors did tries with such a long history of weak public not accompany such mandates. First Colom- institutions, stronger social accountability bia in 1994, and since then Bolivia (1997), mechanisms may help trigger a virtuous Chile (2001), Mexico (2002), and Argentina circle by creating the pressures needed to (2003), have introduced legislation and sec- mobilize intrastate checks and balances tor reforms to guarantee a package of ser- and, eventually, create the opportunities for vices to all citizens. Such reforms also include stronger and more inclusive political repre- legal mechanisms for beneficiaries to com- sentation mechanisms. In other words, plain and demand state compliance. Of Conclusions 57

course, whether this approach is enforceable way specific services are being delivered to depends crucially on how fiscally and opera- improve performance. tionally reasonable the mandates really are. 1. The first dimension is the state’s overall strength and formality. We have argued The Centrality of the “Fitting that there is a general tendency across Process” the region for states to be weak in the Our review of the Latin American experi- sense of doing a poor job delivering key ence supports the view that there cannot be public goods and quality basic economic a “one-size-fits-all approach” to service and social services. Naturally, there are delivery: differences in types of services, differences between countries. State characteristics of the population, and the weakness is manifested in various ways nature of the political system make any but, for service delivery, the nature of the attempt to define the perfect “size” futile state bureaucracy (ranging from “dys- (World Bank 2003a). functional” to “Weberian”) and the Seeking the “size that fits” is at the core of degree to which the rule of law is estab- any service delivery reform effort. Character- lished (ranging from informality to dif- istics of different services—such as the degree ferent degrees of formality) appears to of discretion in decision making and the be critical (see, for example, Burki and intensity of transactions involved in the Perry 1998). Chile, for example, is a delivery process, the existence of economies country with a relatively well-function- of scale, and so on—make them better or ing bureaucracy; significant parts of worse suited for alternative delivery models Brazilian’s bureaucracy work effectively (Pritchett and Woolcock 2002; World Bank despite high levels of clientelism in parts. 2003a). For example, as discussed above, it is Poorer countries such as Bolivia or much easier to create the conditions for a Guatemala have much weaker and less well-functioning competitive market for rule-bound bureaucracies. The further some services than for others (for example, away from having a “Weberian” bureau- cellular telephones compared to rural water cracy a country (or a sector in a country) systems). Similarly, some services (for exam- is, the less likely sophisticated public ple, vaccination) are much better suited for administration reforms seeking to estab- massive, standardized approaches than oth- lish performance orientation in the state ers that require much more customization are to succeed. Similarly, in countries (for example, the pedagogical aspects of where the rule of law is not well estab- schooling). An important aspect of the “fit- lished, the enforcement of contracting ting” involves disentangling the nature of dif- out schemes is likely to be weak, and thus ferent aspects or components of each service such approaches will be more vulnerable (for example, how easy it is to monitor to abuse. results) to determine the most appropriate 2. The second dimension is the prevailing approach. type of political representation. We have But service characteristics are just one argued that Latin American politics have element in the equation. Probably more historically been characterized by clien- important is the country context. Of telism and corporatism, which have a course, countries differ along many differ- direct negative impact on public sector ent dimensions and any systematization of service delivery systems and make it dif- such differences runs the risk of oversimpli- ficult to develop alternative delivery fying complex realities. Nevertheless, based methods that may fit better changing on the review of experiences presented in realities and needs on the ground. Once chapter 3, it is possible to identify four key again, there are differences among coun- dimensions that help characterize a country tries in the extent to which patronage context and provide a useful orientation to and authoritarianism continue being the the range of feasible reform options that pattern by which politicians and citizens policy makers face in seeking to alter the relate to each other, as well as in the 58 Citizens, Politicians, and Providers: The Latin American Experience with Service Delivery Reform

degree to which programmatic (idea- tion unless the state is capable of enforc- based) parties have replaced corporatist ing strict quality regulations and/or sub- or populist movements. Efforts to reform sidizing access to quality services by tradi- intragovernmental compacts (for exam- tionally excluded groups. ple, by establishing performance-based The previous three dimensions are reward systems in the state) are likely to strongly interlinked in what has been be captured by insiders in the presence of characterized as the syndrome of “weak strong corporatist movements. Similarly, and unequal institutions” (De Ferranti et in highly clientelistic environments, al. 2004), which affects a majority of schemes involving community-managed LAC countries. These “structural” fac- services may be highly vulnerable to tors play out in various ways depending manipulation and capture by local on historical trends. patrons. Many countries are in transi- 4. Thus, a fourth dimension is the point of tion, especially at local levels: Colombia departure regarding how services are is traditionally highly clientelistic, but has being delivered. The history of how spe- experienced major changes in Bogotá, for cific services have been delivered matters example. to the extent that it creates constituencies 3. The degree of social and economic for the status quo as well as ideological inequality is the third dimension. It is well reference points for society’s conception known that Latin American countries of what are appropriate roles for different experience high levels of income inequal- actors and thus influences the costs of a ity relative to other parts of the world. It is transition to a different approach. In par- also increasingly recognized that several ticular, the historical pattern of state countries in the region experience sharp, involvement in the delivery of specific group-based social inequalities (particu- services differs between countries (for larly linked to race and ethnicity). High example, countries with a strong tradi- levels of social and economic inequality tion of publicly provided basic education typically are reflected in large gaps in versus others where private schools have access to quality services and thus historically represented a large share increase the cost of meeting the goal of of the system) and with it the strength ensuring minimum service levels for all. and attitudes of key stakeholders (includ- Social inequality and stratification are ing, of course, citizens as consumers also likely to weaken the government’s of those services) to alternative reform ability to generate necessary consensus approaches. Overall, path dependence for service reforms that often involves plays a big role in defining feasible service tradeoffs in the short run. Managing delivery reforms: forces for inertia tend inequality is particularly challenging to be high. where it is associated with group-based differences and social exclusion—as with Of course, it is the combination of these indigenous groups in the Andes and parts (and other) factors that together define of Mexico and Central America, and Afro country context and help determine which groups in Brazil, Colombia, and Vene- type of service delivery approach is most zuela, and many in the Caribbean. likely to succeed in improving results. As Besides, societies stratified along ethnic an illustration of how context matters, in and racial lines tend to be vulnerable to table 4.1 we provide a simple set of criteria populist appeals of politicians who appeal derived from our assessment of the Latin to specific societal groups by providing American experience. These are meant as them with selective private goods rather illustrations of how to analyze country than public goods that benefit society as a context in assessing prospects for specific whole (Keefer and Khemani 2003). In a service delivery reform approaches. context of high social and economic In a country where state bureaucracies inequality, an emphasis on choice and operate broadly along Weberian lines and competition may yield further stratifica- the rule of law is well established (most Conclusions 59

Table 4.1 Fitting Approaches to Country Conditions

State characteristics

Weberian bureaucracy and Acceptable rule of law but weak Weak bureaucracy and limited rule acceptable rule of law bureaucracy of law Strong tradition of state Fruitful to try results-based Small-scale experimentation with Very modest administrative reforms presence in service delivery management. Need for results-based management (for linked directly to client power complementary actions to promote example, an enclave) within overall initiatives (see below). social accountability (and long-term efforts at increasing Decentralization problematic as decentralization) in cases of pressures (demand) for reform at a national standards will be hard to authoritarian tradition and weak larger scale. Decentralization is an enforce. citizen participation. option for large countries.

Point of departure Tradition of diversiﬁed service Low inequality: competition and Contracting out a good option as Client power may be the only option delivery choice. long as opposition from vested in the short run. If inequality is high, interests (political patrons and community management may be High inequality: contracting out with corporative interests) allows it. less costly than choice/competition strong compensatory design. unless clientelism is pervasive.

likely, countries where patronage has weak- In the other extreme, in countries with ened significantly as the primary form of dysfunctional state bureaucracies and political relationship), the choice between where rule of law is limited, the set of deepening reforms of public bureaucra- choices available to policy makers is much cies—seeking to make them more efficient more limited. In such settings, reforming and performance-oriented—and expand- public sector delivery systems is a long- ing alternative service delivery models term endeavor that will face major techni- involving nonstate actors will be highly cal and political constraints. Ambitious dependent on the departure point. In sec- efforts to modernize administrative sys- tors where there is a strong tradition of state tems are likely to go wasted in the face of presence, the costs of radically shifting weak civil service, and attempts to reform approaches may be too high. In those cases personnel practices are likely to generate a cost-benefit analysis is likely to support a strong resistance from those benefiting strong push for results-based management from the prevailing use of government jobs (as is currently being tried in the case of as instruments of political patronage. Thus, Chile). This may need to be complemented in these settings, administrative reforms are by actions to open up the state to feedback likely to be a long-term affair that will not and influence from clients or citizens, if no bring major short-run service delivery ben- such tradition exists. For those sectors in efits. It will often be necessary to comple- which there is a tradition of private (for- or ment such reforms with parallel efforts to not-for-profit) participation in service enhance service delivery by nonstate actors. delivery, a key factor to consider in selecting Limited rule of law is likely to be a serious alternative delivery mechanisms is the constraint to the enforcement of contracts, degree of social and economic inequality. unless contractors have very strong intrin- Choice and competition are likely to oper- sic motivations to deliver quality services to ate better either under conditions of low clients (as in the case of some faith-based inequality or where there is the potential for NGOs, such as the noted example of Fe y developing means to ensure that poorer Alegría schools). The choice between an groups are not excluded from the widening approach that builds on community-man- possibilities. Under conditions of high aged services and one based on competi- inequality, stronger state involvement is tion and choice as alternatives to public required and, thus, contracting out using sector provision depends on social and strong compensatory schemes may be a bet- political context. In these countries, the ter alternative (for example, targeting pub- ability of governments is likely to be low to licly subsidized concessions to low-income design and implement compensatory mea- areas as in the Bogotá school concessions sures to address informational asymme- program). tries and differences in purchasing power. 60 Citizens, Politicians, and Providers: The Latin American Experience with Service Delivery Reform

Thus, under conditions of high inequality, hard-to-reform state bureaucracies. The community management schemes may be rule of law is also a supportive factor for a better alternative (as, for example, in the contracting out service provision with non- case of EDUCO in El Salvador). On the state actors. This may be the most appropri- other hand, when local level clientelism is ate approach as long as opposition from pervasive, the danger that such schemes vested interests is not strong enough to end up being captured by local patrons is block it. This suggests that contracting out significant, and thus competition among is likely to work better in countries in which private providers may be the best option. patronage, if not completely eliminated, is Where the rule of law is weak, there are also in the process of being weakened. greater risks of privatization being cap- In practice, though, the policy maker tured or corrupted—although the problem choice is constrained and conditioned by a is that it is often precisely in these cases that multiplicity of factors, which we would find public delivery is also weakest and most difficult to predict exante. Finding the “right captured. approach” is both a political and a techno- Naturally, there is a whole range of cases cratic process. This is true both at the macro in between these two extremes. In table 4.1 or national level, where there is growing evi- we have identified a particular configura- dence of the political complexity of service tion characterized by an acceptable rule of delivery reforms (Grindle and Mason 2004; law but weak bureaucracy—a configuration Kaufman and Nelson 2004), and at the that may approximate the case of a large micro delivery point, where specific design number of Latin American countries where choices interact with provider and client strong traditions of clientelism and/or behaviors in often-complex ways. As Robin- strong corporatist movements still have a son (2003) has argued, the issue is much less large negative effect on the quality of state the specific design choices or the new magic bureaucracies, even though rule of law has bullet (that quickly loses its shine) than the been strengthened over the last two processes that generate policy adjustments decades. Unlike the case of countries with to changing outcomes in response to soci- “Weberian” bureaucracies, the opportuni- etal and political pressures. This may take ties are much reduced for large-scale place at the national level or at local levels— reforms seeking to improve service delivery under the conditions of substantive political by the state through the introduction of and administrative decentralization that performance-based management, which exist in Latin America. Paraphrasing Fox suggests that smaller scale experimentation (2000, 2001), we can say “the challenge can may be more appropriate. It is in these cases be understood as a kind of political Feng that the tension between short-term fixes Shui—the art of placing things in balanced and longer-term reforms may be the relationship to one another.” strongest. Unlike the case of countries That is why the process by which service where the rule of law is not yet established, delivery reforms are defined, designed, and there are likely to be opportunities to create implemented is so crucially important. In “enclaves” or “islands of efficiency” that other words, identifying the “size that fits” bypass the overall weak state bureaucracies country and sector circumstances is likely (as illustrated by the experiences of various to require the participation and involve- social funds). At the same time, as discussed ment of different voices in society. As above, unless those experiments have some Pritchett and Woolcock (2002) suggest, built-in mechanisms through which experi- unlike some macroeconomic policies, ser- ments are, somehow, mainstreamed, the vice delivery reforms can hardly be man- short-term impacts of innovations may be aged by “ten smart people.” This may very lost over time. The fact that the rule of law likely involve more open and deliberative is more established suggests that there may forms of governance than what is presently be room to experiment with decentraliza- the case in most Latin American countries, tion of service delivery responsibilities— with the consequent requirements in terms possibly another way of bypassing some of institutions for consensus building and Conclusions 61

conflict management. Of course, strategic mental actors involved in service delivery, incrementalism applies also in this area: one of the biggest challenges ahead of there is scope for many partial changes and Latin American countries is to find simple payoffs in different parts of the accountabil- ways by which the many advantages of ity structure that will often involve specific, “learning by doing” can be leveraged partial forms of opening and deliberation, through mechanisms that facilitate and and that will support the broader shift in encourage “learning from others,” without institutions. the artificiality implicit in the traditional But to be effective and avoid becoming a calls for “replication.” synonym for “chitchatting” debate, consen- Where would the leadership for such an sus building and deliberation need to be fed approach come from? In the very graphic by rigorous analysis based on solid empir- terms used in a recent review of the litera- ics. This reinforces the importance of ture on state accountability in Latin Amer- robust and systematic evaluations, few of ica, “somebody has to kick the status quo which are readily available, that feed the from its point of equilibrium” (Schedler policy-making process in a timely fashion. 1999, 347). Who? Improving state account- In other words, strengthening the links ability will most likely require reform from between policy evaluation, debate, design, above (politicians), below (citizens), inside and implementation is fundamental to achieve a better process. That is, a learning from experience approach is needed. In this approach, policy design is the outcome of a BOX 4.2 The Role of Evaluation dialogue among interested parties, and A daunting challenge facing most countries results-based approaches among particular implementation is organized in ways that in the region is to develop an institutional sets of programs that foster demands for a enable the elaboration of regular and solid arrangement whereby results of technical greater role for evaluation in other parts of policy evaluations, which frame and sup- analyses are effectively used to inform pol- the government. icy, budgetary, and managerial decisions. The LAC experience with the rapid port further dialogue and debate. Establishing a process where evaluation introduction of conditional cash transfer The practice of “informed dialogue” results naturally feed into the process of programs also illustrates the powerful role (Reimers and McGinn 1997), whereby budget negotiations is only one piece of that can be played by the strategic use of research-based knowledge is used to shape the puzzle. An even greater challenge is to sound impact evaluations (Rawlings and foster effective demand for good analysis as Rubio 2003). Unlike many other social ser- policy, requires a strong sensitivity to a basis for government decision making. vice delivery reforms, impact evaluations process and the engagement of all relevant Consistent with its reputation for good were included as an integral part of the stakeholders. Implementing such an governance and competent public adminis- development and application of some CCT approach would undoubtedly involve capa- tration, since 1994 Chile has gradually programs.These evaluations have served developed a robust system of performance not only a technical purpose in informing bilities on the part of governmental and measurement and, increasingly, program expansion and modification deci- nongovernmental actors that, in most cases, management. Although the Chilean sions, but also a political purpose that do not yet exist. Fox (2001), for example, approach is decentralized (that is, no single allows policy makers to protect effective identifies a key role that civil society organi- entity is responsible for developing and programs during political transitions, as maintaining a national system as in Colom- PROGRESA/Oportunidades in Mexico exem- zations can play in their advocacy work, by bia), the processes are relatively well coordi- plified.This latter case was one in which a producing policy and program evaluations. nated by the Interministerial Committee at rigorous evaluation was rare in the country, He emphasizes both the capacity and orga- the cabinet level.The Ministry of Finance’s but the high profile of the PROGRESA/Opor- Dirección de Presupuestos (Budget Direc- tunidades evaluations—and their nizational challenges of such an initiative torate) also plays a key role in coordinating importance in underpinning the program and the large expected payoffs that would various decentralized initiatives and uses continuation and extension despite the result from it. To the extent that such invest- evaluation results as a basis for defining the change in government—has led to a rising ments do materialize, the synergistic civil budget framework (Marcel 1997; Armijo demand for evaluations in other parts of 2003). Performance is evaluated at the level the government and in Congress.The diffi- society or state relations discussed earlier of policies and expenditures, organizations, culties that CCT program administrators could well develop in this area as well. Thus, and individual public servants.The govern- overcame in ensuring that sound a key way to support service delivery ment has begun to link performance evalu- evaluations were conducted are also part of reforms is to strengthen the capacities of all ations to organizational incentives. this important lesson, underscoring the Moving to a comprehensive results- need to secure a solid commitment from relevant actors to participate effectively in based management system will be difficult policy makers from the program design policy evaluation and dialogue. in most countries. A more realistic objective stage forward to maintain the integrity of Furthermore, particularly given the is to build areas of evaluation, debate, and the program and evaluation designs. increasing range of state and nongovern- 62 Citizens, Politicians, and Providers: The Latin American Experience with Service Delivery Reform

(state bureaucrats), and outside (interna- advocated above help turn it into a more tional actors). The recent Latin American effective source of change? We humbly think experience provides sufﬁcient evidence that so. We see processes of systematic explo- such leadership exists within and outside the ration and evaluation, integrated within state. Often it may have become ineffective democratic debate, as key both for the “tech- by single-mindedly seeking to push a magic nical” design issues and for the broader bullet solution to service delivery. Shouldn’t process of empowering all citizens and mak- a “learning from experience approach” as ing decision makers more accountable. References

Abrucio, F. 1999. “Reforma do Estado e Contexto Federativo Baeza, C. 1996. “Recursos Humanos: La Agenda Pendiente en la Brasileiro.” Pesquisa 12. Konrad-Adenauer-Stiftung. Modernización del Sector Público de Salud en América Latina.” Adato, M. 2000.“The Impact of PROGRESA on Community Social Revista del Programa de Administración en Salud (PIAS) 12, Uni- Relationships.” International Food Policy Research Institute, versidad de Chile, November. Washington, DC, September. Baeza, C., and M. Cabezas. 1998. “La Separación de la Función de Adato, M., D. de la Brière, D. Mindek, and A. Quisumbing. 2000. Compra y Provisión en la Modernización del Sector Salud.” Serie “Final Report: The Impact of PROGRESA on Women’s Status Documentos para el Diálogo en Salud 45. Santiago, Chile: CLAISS and Intrahousehold Relations.” International Food Policy (Centro Latinoamericano de Investigación de los Sistemas de Research Institute, Washington, DC, July. Salud). ADERASA (Asociación de Entes Reguladores de Agua y Bardhan, P., and D. Mookherjee. 2000a. “Decentralizing Anti- Saneamiento de las Américas). Forthcoming. “Las Tarifas de Poverty Program Delivery in Developing Countries.” Working Agua Potable y Alcantarillado en America Latina.”Grupo de Tar- Paper, University of California, Berkeley. ifas y Subsidios, (Working group on rates and subsidies). ———. 2000b. “Capture and Governance at Local and National Aglaê de Medeiros Machado, M. 2002.“School Autonomy in Brazil: Levels.” American Economic Review 90 (2): 135–9. General Overview and Selected States.” Photocopy, World Bank, Baum, M. A., and D. A. Lake. 2003. “The Political Economy of Brasilia, Brazil, May. Growth: Democracy and Human Capital.” American Journal of Alvestegui, A., and J. Sara. 2003. “Estudio de Sustentabilidad de los Political Science 47 (2): 333–47. April. Servicios de Agua Potable Rural en Chile.” Photocopy, World Bitran, R., J. Munoz, et al. 2000. “Equity in the Financing of Social Bank, Santiago, Chile. Security for Health in Chile.” Health Policy 50 (3): 171–196. Ames, B. 2003.“The State, Civil Society and Inequality.”Photocopy, Blackburn, J. 2001. “Popular Participation in a Prebendal Society: University of Pittsburgh, Pittsburgh, Pennsylvania. An Actor Oriented Analysis of Participatory Municipal Planning Angell, A., P. Lowden, and R. Thorp. 2001. Decentralizing Develop- in Sucre, Bolivia.” PhD diss., unpublished, Institute of Develop- ment: The Political Economy of Institutional Change in Colombia ment Studies, Sussex University. and Chile. New York: Oxford University Press. Bossert, T. J. 2000. “Decentralization of Health Systems in Latin Angrist, J., E. Bettinger, E. Bloom, E. King, and M. Kremer. 2002. America: A Comparative Study of Chile, Colombia and Bolivia.” “Vouchers for Private Schooling in Colombia: Evidence from a Ran- LAC Health Sector Reform Initiative. Harvard School of Public domized Natural Experiment.” American Economic Review 92 (5). Health 29 June. Angrist, J., E. Bettinger, and M. Kremer. 2004. "Long-term conse- Bossert, T. J., O. Larrañaga, et al. 2003.“Decentralization and Equity quences of secondary school vouchers: Evidence from adminis- of Resource Allocation: Evidence from Colombia and Chile.” trative records in Colombia." Working Paper Series (US), No. Bulletin of the World Health Organization 81 (2): 95–100. 10713: 1–33, National Bureau of Economic Research. August. Brown, D. S. 1999. “Reading, Writing and Regime Type: Democ- Armijo, M. 2003. “La Evaluación de la Gestión Pública en Chile.”In racy’s Impact on Primary School Enrollment.” Political Research Evaluación de Resultados para una Gestión Pública Moderna y Quarterly 52 (4): 681–707. December. Democrática: Experiencias Latinoamericanas, ed.N.Cunill Grau ———. 2002. “Democracy, Authoritarianism and Education and S. Ospina Bozzi. Caracas: CLAD (Centro Latinoamericano Finance in Brazil.” Journal of Latin American Studies 34: 115–141. de Administración para el Desarrollo). Brown, D. S., and W. Hunter. 1999. “Democracy and Social Spend- Atkinson, P., and P.van den Noord. 2001.“Managing Public Expen- ing in Latin America, 1980–1992.” American Political Science diture: Some Emerging Policy Issues and a Framework for Review 93 (4): 779–90. December. Analysis.” Economics Department Working Papers 285, OECD, Buchanan, J. 2001. “Judicial Reform in the Americas.” FOCAL. Paris. November. Ontario, Canada. 63 64 Citizens, Politicians, and Providers: The Latin American Experience with Service Delivery Reform

Burki, S. J., and G. Perry. 1998. “Beyond the Washington Consensus: Di Gropello, E., and J. Marshall. 2004. “Teacher Effort and Student Institutions Matter.” World Bank Viewpoints. Washington, DC. Achievement in Rural Honduras.” Photocopy, World Bank, Calderon, F., and A. Szmukler. 2004. “Political Culture and Devel- Washington DC. opment.”In Culture and Public Action, ed. V. Rao and M. Walton. Dodson, J., and D. Jackson. 2000. “Democratic Transitions in Cen- Stanford: Stanford University Press. tral America: Justice Reform and the Civic Culture.” Paper pre- Campbell, T. 2003. The Quiet Revolution: Decentralization and the sented at the Institutions, Accountability, and Democratic Gov- Rise of Political Participation in L.A. Cities. Pittsburgh: University ernance in Latin America conference, the Helen Kellogg Institute of Pittsburgh Press. for International Studies, University of Notre Dame, May 8–9. Cantuarias, F., and M. Delgado. 2004. “Peru—Urban Land Titling Domingo, P., and R. Sieder. 2001. “Conclusions.” In Rule of Law in Program.” Paper presented at the Scaling-Up Poverty Reduction Latin America: The International Promotion of Judicial Reform, Conference, Shanghai, China, May. ed. P. Domingo and R. Sieder. London: Institute of Latin Ameri- Carcioﬁ, R., O. Cetrangolo, and O. Larrañaga. 1996. “Desafíos de la can Studies. Descentralización de la Educación y la Salud en Argentina y Dorotinsky, W., and Y. Matsuda. 2001. “Reforma de la gestión Chile.”ECLAC, Santiago, Chile. ﬁnanciera en América Latina: una perspectiva institucional.” Carvalho, S., G. Perkins, and H. White. 2002. “Social Funds, Sus- Reforma y Democracia 23: 141–66. tainability and Institutional Development Impacts: Findings DNP (Departamento Nacional de Planeación). 2002. “Exposición from an OED Review.” Journal of International Development 14: de motivos de la reforma a la Ley 60 de 1993: Sector Educación y 611–25. Sector Salud.” Archivos de Economía 173. CEPAL [Comisión Económica para América Latina y el Caribe (Eco- Egaña, R. 2003. “Profesionalización de la función pública en Chile: nomic Commission for Latin America and the Caribbean)]. 2001. Contexto, avances y desafíos.” In Retos de la Profesionalización “Panorama Social de América Latina.”CEPAL, Santiago, Chile. de la Función Pública, ed. D. Arellano, R. Egaña, O. Oszlak, and Chambers, R., D. Narayan, M. Shah, and P.Petesch. 2000. Crying out R. Pacheco. Caracas: CLAD. for Change: Voices of the Poor. New York: Oxford University Press. England, R. 1998. “Health Sector Reform: Human Resources Collado, C. 2003. “Control Social: Participación de los Consumi- Issues—Underlining the Importance of Human Resource Issues dores en la Regulación y Control de los Servicios Públicos Domi- for Commonwealth Countries.” Institute for Health Sector Devel- ciliarios.” Seminario sobre la Participación de la Sociedad Civil opment 14. London. en la Regulación de los Servicios Públicos, Seminar on Civil Soci- Estache, A., V. Foster, and Q. Wodon. 2002. “Accounting for Poverty ety Participation in the Regulation of Public Services, Lima, in Infrastructure Reform: Learning from L.A.’s Experience.” Peru, August. Development Studies Series, World Bank, Washington, DC. Correa Sutil, J. 1999. “Judicial Reforms in Latin America: Good Evans, P. 2002. “Beyond Institutional Monocropping: Institutions, News for the Underprivileged?” In The (Un)Rule of Law and the Capabilities and Deliverative Development.” Photocopy, Univer- Underprivileged in Latin America, ed. J. Mendez, G. O’Donnell, sity of California, Berkeley, California, January. and P.S. Pinheiro. Notre Dame, Indiana: University of Notre Fay. 2000. “Financing the Future: Infrastructure Needs in Latin Dame Press. America 2000–2005.” World Bank Working Paper No. 2545, Cotlear, D. 2000. “Peru: Reforming Health Care for the Poor.” World Bank, Washington, DC. LCSHD, Working Paper Series 57, World Bank, Washington, DC, Felicio, M. and I. John-Abraham. 2004. "Perú: Toward a System of March. Social Accountability," En Breve No. 39, World Bank, Washing- Cox, C., and P.González. 1998.“Educación: de Programas de Mejo- ton, DC. ramiento a Reforma.” In Construyendo Opciones: Propuestas Fiedler, J. 2002. “Resource Allocation and Purchasing in Latin Económicas y Sociales para el Cambio de Siglo, ed. Cortázar and America and Caribbean: A Regional Review.” HNP, Discussion Vial. Santiago, Chile: CIEPLAN (Corporación de Estudios para Paper, ed. A. S. Preker, World Bank, Washington, DC. América Latina). Foster, V. 2002. “Ten Years of Water Service Reform in Latin Amer- Crouch, L. 2004. “Political Economy, Incentives, and Teachers’ ica: Towards an Anglo-French Model.” In Reinventing Water and Unions: Case Studies in Chile and Peru.” Photocopy, World Wastewater Systems, ed. P.Seidensat, D. Haarmeyer, and S. Hakim. Bank, Washington, DC. New York: John Wiley & Sons Inc. De Ferranti, D., G. Perry, H. G. Ferreira, et al. 2004. “Inequality in ———. 2003. “Notes on Infrastructure in Latin America.” Photo- Latin America: Breaking with History?” World Bank, Washing- copy, World Bank, Washington, DC. ton, DC. Foster, V., and O. Irusta. 2003. “Does Infrastructure Reform work Diaz-Cayeros, A., B. Magaloni, and B. Weingast. 2002. “Federalism for the Poor? A Case Study on the Cities of La Paz and El Alto in and Democratisation in Mexico.” Paper presented at the 2002 Bolivia.” World Bank Policy Research, Working Paper 3177, meeting of the American Political Science Association, Washing- World Bank, December. ton, DC. Fox, J. 2000. “Civil Society and Political Accountability: Proposi- Di Gropello, E. 2002. “An Assessment of the Impact of Decentral- tions for Discussion.” Paper presented at the Institutions, ization on the Quality of Education in Chile.” Economist’s Forum. Accountability, and Democratic Governance in Latin America World Bank, Washington, DC. conference, the Helen Kellogg Institute for International Studies, University of Notre Dame, May 8–9. References 65

———. 2001. “Vertically Integrated Policy Monitoring: A Tool for Hall, G., and H. Patrinos. 2004. Forthcoming. “Indigenous People, Civil Society Policy Advocacy.” Nonprofit and Voluntary Sector Poverty and Human Development in Latin America, Quarterly 30 (3): 616–27. September. 1994–2004.”Washington, DC: World Bank. Galiani, S., and E. Schargrodsky. 2002. “Evaluating the Impact of Henderson, K., A. Shah, S. Elena, and V.Autheman. 2004.“Regional School Decentralization on Educational Quality.” Economía 2 Best Practices: Enforcement of Judgments: Lessons from Latin (2), Journal of the Latin American and Caribbean Economic America.” IFES, Rule of Law Paper Series, Washington, DC, Association. April. Galiani, S., P. Gertler, and E. Schargrodsky. 2002. “Water for Life: Hicks, N., and J. Todd. 2003. “A Brief Examination of the Relation- The Impact of the Privatization of Water Services on Child Mor- ship between GDP/capita and Performance on Selected Social tality.” Photocopy, University of San Andrés, Buenos Aires, Indicators in the Latin America and Caribbean Region.” Photo- Argentina. copy, World Bank, Washington, DC. Galindo, P. 2003. “Indicadores Subjetivos: Estudios, Calificaciones Holm-Nielsen, L., K. Thorn, and J. Prawda. 2004.“Chile: Decades of de Riesgo y Encuestas de Percepción Pública sobre los Sistemas Educational Reform Deliver.” World Bank En Breve 44. de Justicia.” Centro de Estudios de Justicia de las Américas, Sis- Hsieh, C. T., and M. Urquiola. 2003.“When Schools Compete, How temas Judiciales, Resultados Recientes para las Américas 3 (6): do They Compete? An Assessment of Chile’s Nationwide School 4–35. Santiago, Chile. Voucher Program.” Working Paper Series (U.S.); No. 10008:1- Gallego, F. 2002. “Competencia y Resultados Educativos: Teoría y (43), National Bureau of Economic Research, September. Evidencia para Chile.” Cuadernos de Economía 39 (118): 309–52. Jiménez, E., and Y. Sawada. 1998. “Do Community Managed Garibaldi, J. A. N.d. “Instituciones Informales y Desempeño del Schools Work? An Evaluation of El Salvador’s EDUCO Pro- Sector Público en México.” Photocopy, FLACSO and Instituto gram.” World Bank Economic Review 13 (3): 415–41. Tecnológico Autónomo de México. Kaufman, R. R., and J. Nelson. 2004. Crucial Needs, Weak Incentives: Garretón, M. A. 2002. “La Transformación de la Acción Colectiva Social Sector Reform, Democratization, and Globalization in Latin en América Latina.” Revista de la CEPAL 76. April. America. Baltimore: John Hopkins University Press. Garro, A. M. 1999.“Access to Justice for the Poor in Latin America,” Kaufman, R. R., and A. Segura-Ubiergo. 2001. “Globalization, In The (Un)Rule of Law and the Underprivileged in Latin America, Domestic Politics and Social Spending in Latin America: A Time ed. J. Mendez, G. O’Donnell, and P.S. Pinheiro. Notre Dame, Series Cross-Section Analysis, 1973–97.” World Politics 53 (4): Indiana: University of Notre Dame Press. 553–87. July. Gomez Lobo, A., and D. Contreras. 2000. “Subsidy Policies for the Keefer, P. 2002. “Clientelism, Credibility and Democracy.” Photo- Utility Industries: A Comparison of the Chilean and Colombian copy, World Bank, Washington, DC. Water Subsidy Schemes.” Photocopy, University of Chile, Keefer, P., and S. Khemani. 2003. “Democracy, Public Expenditures Department of Economics, Santiago, Chile. and The Poor.” Policy Research Working Paper 3164, World Government of Jamaica. 2003.“Government at Your Service: Public Bank, Washington, DC, November. Sector Modernisation Vision and Strategy 2002–2012.” Ministry King, E., P.Orazem, and D. Wohlgemouth. 1999.“Central Mandates Paper 56, Kingston, Jamaica. January. and Local Incentives: The Colombian Education Voucher Pro- Gray Molina, G. 2002. “The Offsprings of 1952: Poverty Exclusion gram.” World Bank Economic Review 13 (3): 467–91. and the Promise of Popular Participation.”Paper presented at the King, E., and B. Ozler. 1998. “What’s Decentralization Got to Do conference on The Bolivian Revolution at 50: Comparative with Learning? The Case of Nicaragua’s School Autonomy Views on Social, Economic and Political Change, Harvard Uni- Reform.” Working Paper Series on Impact Evaluation of Educa- versity. Cambridge, MA. tion Reforms 9, World Bank, Washington, DC. Gray Molina, G., E. Pérez de Rada, and E. Yeñez. 1999. “La King, E., L. Rawlings, M. Gutierrez, C. Pardo, and C. Torres. 1997. Economíca Política de Reformas Institucionales en Bolívia.” “Colombia’s Targeted Education Voucher Program: Features, Cov- Working Paper R-350, Inter-American Development Bank erage and Participation.”Working Paper Series on Impact Evalua- (IDB), Office of the Chief Economist, Washington, DC. tion of Education Reforms 3, World Bank, Washington, DC. Griffiths, M., and J. S. McGuire. 2004. “A new dimension for health La Forgia, G., P. Mintz, and C. Cerezo. 2004. “Is the Perfect the reform: Atención integral a la niñez en la comunidad en Hon- Enemy of the Good? A Case Study of Large-Scale Contracting for duras.”Photocopy, World Bank, Washington DC. Basic Health Services in Rural Guatemala.” Photocopy, World Grindle, M. S., and E. S. Mason. 2004. Despite the Odds: Contentious Bank, Washington, DC, March. Politics and Education Reform. Princeton: Princeton University Lake, D. A., and M. A. Baum. 2001. “The Invisible Hand of Democ- Press. racy: Political Control and The Provision of Public Services.” Guasch, J. L. 2004. “Granting and Renegotiating Infrastructure Comparative Political Studies 34 (6): 587–621. August. Concessions: Doing It Right.” WBI Development Studies. April. Laking, R. 2002. “Agencies: Their Benefits and Risks.” Photocopy, Guedes, A., T. Lobo, R. Walker, and A. L. Amaral. 1997. “Gestión Victoria University of Wellington, New Zealand. Descentralizada de la Educación en el Estado de Minas Gerais, Maceira, D., and M. V.Murillo. 2001.“Social Sector Reform in Latin Brasil.” Working Paper 11, Grupo de Desarrollo Humano, America and the Role of Unions.” Working Paper #456, IDB América Latina y el Caribe, World Bank, Washington, DC, May. Research Department, Washington, DC. 66 Citizens, Politicians, and Providers: The Latin American Experience with Service Delivery Reform

Mainwaring, S. 2003. “Introduction: Democratic Accountability in Patrinos, H. 2002. “Contracting Out in Education.” Photocopy, Latin America.”In Democratic Accountability in Latin America,ed. World Bank, Washington, DC. S. Mainwaring and C. Welna. New York: Oxford University Press. Pasara, L. 2002. “Justicia y Ciudadanía Realmente Existentes.” Marcel, M. 1997. “Modernización del Estado e Indicadores de Política y Gobierno IX (2). September. Desempeño en el Sector Público: La Experiencia Chilena.” Paxson, C., and N. Schady. 2002. “The Allocation and Impact of Departamento de Ingeniería Industrial, University of Chile, San- Social Funds: Spending on School Infrastructure in Peru.” World tiago, Chile. Bank Economic Review 16 (2): 297–319. Massad, C. 1996. “Health Care Reform in Chile.” Revista Médica de Payne, J. M., D. Zovatto G., F. Carrillo Flórez, and A. Allamand Chile 124 (7): 880–6. Zaval. 2002. Democracies in Development: Politics and Reform in McClymont, M., and S. Golub, eds. 2000. Many Roads to Justice: The Latin America. Washington, DC: Inter-American Development Law-Related Work of Ford Foundation Grantees Around the World. Bank and International Institute for Democracy and Electoral New York: The Ford Foundation. Assistance. McEwan, P., and M. Carnoy. 1999. “The Impact of Competition on Peruzzotti, E., and C. Smulovitz. 2001.“Civil Society, the Media and Public School Quality: Longitudinal Evidence from Chile's Internet as Tools for Creating Accountability to Poor and Disad- Voucher System.”Photocopy, Stanford University, Palo Alto. vantaged Groups.” Occasional Paper 2002/13, UNDP (United ———. 2000. “The Effectiveness and Efficiency of Private Schools Nations Development Programme), Human Development in Chile's Voucher System.” Education Evaluation and Policy Report Office. November. Analysis, vol. 22, No. 3, 213–39. Pinheiro, P. S. 1999. “Introduction in Mendez.” In The (Un)Rule of McKenzie, D., and D. Mookherjee. 2002. “Distributive Impact of Law and the Underprivileged in Latin America, ed.J.Mendez, Privatization in Latin America: An Overview of Evidence from G. O’Donnell, and P. S. Pinheiro. Notre Dame, Indiana: Univer- Four Countries.” Photocopy, Departments of Economics Stan- sity of Notre Dame Press. ford University and Boston University, Palo Alto and Boston. Pritchett, L., and M. Woolcock. 2002. “Solutions When The Solu- Méndez, J. 2000.“Legislatures, Judiciaries and Innovations in Hori- tion Is the Problem: Arraying the Disarray in Development.” zontal Accountability.” Paper presented at the Institutions, Working Paper 10, Center for Global Development. September. Accountability, and Democratic Governance in Latin America Przeworski, A., S. Stokes, and B. Mannin. 1999. “Introduction.” In conference, the Helen Kellogg Institute for International Studies, Democracy, Accountability and Representation, ed. Przeworksi et University of Notre Dame, May 8–9. al. Cambridge: Cambridge University Press. Ministerio de Salud. 2000.“Sistematización de la Estrategia Integral Rawlings, L. B., and G. M. Rubio. 2003. “Evaluating the Impact of en Salud Materno Infantil en la Provincia de Paruro.”Photocopy, Conditional Cash Transfer Programs.” Policy Research Working Ministerio de Salud, Dirección Regional de Salud Cusco, UBASS. Paper 3110, World Bank, Washington, DC, August. Molinar, J., and J. Weldon. 1994. “Electoral Determinants and Con- Rawlings, L. B., L. Sherburne-Benz, and J. Van Domelen. 2004. Eval- sequences of National Solidarity.” In Transforming State Society uating Social Funds: A Cross-Country Analysis of Community Relations in México, ed. W. Cornelius, A. C. Craig, and J. Fox. San Interventions. Washington, DC: World Bank. Diego: University of California at San Diego. Reimers, F., and N. McGinn. 1997. Informed Dialogue: Using O’Donnell, G. 1993. “On the State, Democratization and Some Research to Inform Education Policy around the World. Westport: Conceptual Problems: A Latin American View with Glances at Praeger Publishers. Some Postcommunist Countries.” World Development 21 (88): Rinne, J. J. 2001. “Redesigning the State in Latin America: Pundits, 1355–69. Policymakers, and Organized Labor in Argentina and Brazil.” ———. 1994. “Delegative Democracy.” Journal of Democracy 5 (1): PhD diss. Department of Politics, Princeton University. 55–69. January. Robinson, J. 2003. “Politician-Proof Policy?” Background paper to ———. 1996. “Illusions about Consolidation,” Journal of Democ- the 2004 WDR. April 12. racy 7 (2): 34–51. April. Robinson, J., and T.Verdier. 2002.“The Political Economy of Clien- ———. 1998.“Accountability Horizontal.” AgorA 8: 5–34. Summer. telism.”Working Paper 3205, CEPR (Centre for Economic Policy ———. 2001. “Reflections on Contemporary South American Research), London. Democracies.” Journal of Latin American Studies 33: 599–609. Salamon, L., H. Anheier, R. List, S. Toepler, S. Sokolowski, and Asso- OECD (Organisation for Economic Co-operation and Develop- ciates. 1999. “Global Civil Society: Dimensions of the Nonprofit ment). 2003. “Reviews of National Policies for Education in Sector.” The Johns Hopkins Center for Civil Society Studies, Chile.”OECD, Paris. October. Baltimore. Olive, J. M., J. B. Risi, Jr., et al. 1997. “National Immunization Days: Santamaría, M. 2003. “Income Inequality, Skills and Trade: Evi- Experience in Latin America.” The Journal of Infectious Diseases dence from Colombia During the 80s and 90s.” Photocopy, 175 (suppl 1): S189–93. World Bank, Washington, DC. Over, M., and N. Watanabe. 2003. “Evaluating the Impact of Orga- Sapelli, C., and B. Vial. 2002.“The Performance of Private and Pub- nizational Reforms in Hospitals.”In Innovations in Health Service lic Schools in the Chilean Voucher System.” Cuadernos de Econo- Delivery: The Corporatization of Public Hospitals, ed. A. Preker mia 39 (118): 423–54. and A. Harding, 105–65. Washington, DC: World Bank. References 67

Schady, N. 2000. “The Political Economy of Expenditures by the Tommasi, M., S. Saiegh, and P.Sanguinetti. 2001.“Fiscal Federalism Peruvian Social Fund (FONCODES).” American Political Science in Argentina: Politics, Policies and Institutional Reform.” Econo- Review 94 (2): 289–304. mia 1 (2): 157–201. Schedler, A. 1999. “Restraining the State: Conflicts and Agents of Vega, J., P. Bedregal, et al. 2003. “Gender Inequity in the Access to Accountability.” In The Self-Restraining State, ed. A. Schedler, Health Care in Chile.” Revista Médica de Chile 131 (6): 669–78. L. Diamond, and M. Plattner. Boulder: Lynne Rienner Publishers Wilson, B. M., and R. Handberg. 1998. “Opening Pandora’s Box: Inc. The Unanticipated Political Consequences of Costa Rican Legal Schedler, A., L. Diamond, and M. Plattner. 1999. “Introduction.” In Reform.” Paper presented at Midwest Political Science Associa- The Self-Restraining State, ed. A. Schedler, L. Diamond, and tion, Chicago, April 23–25. M. Plattner. Boulder: Lynne Rienner Publishers Inc. World Bank. 1999. Beyond the Center: Decentralizing the State. Schiefelbein, E., and P. Schiefelbein. 2000. “Three Decentralization Washington, DC: World Bank. Strategies in Two Decades: Chile 1981–2000,” Journal of Educa- ———. 2000a. World Development Report: Attacking Poverty. New tional Administration 38 (5). York and Washington, DC: Oxford University Press, IBRD, Sen, A. K., and J. Drèze. 1989. Hunger and Public Action. Oxford: World Bank. Clarendon Press. ———. 2000b. “Bolivia from Patronage to a Professional State: Shefter. M. 1994. Political Parties and the State: The American His- Institutional and Governance Review.” Report 20115-BO, World torical Experience. Princeton: Princeton University Press. Bank, Washington, DC. Shepherd, G. 2003. “Civil Service Reform in Developing Countries: ———. 2001a. “Peru. Institutional & Governance Review.” Report Why Is It Going Badly?” Paper presented at the 11th Interna- 22637-PE Latin America and The Caribbean, Poverty Reduction tional Anti-Corruption Conference, Seoul, Republic of Korea. and Economic Management Unit, World Bank, Washington, DC, Shifter, M. 1997. “Tensions and Trade-Offs in Latin America.” Jour- August. nal of Democracy 8 (2). April. ———. 2001b. “Argentina: An Assessment of Institutional Capac- Shugart, M. S., E. Moreno, and B. F. Crisp. 2000. “The Accountabil- ity for Social Sector Reform in Argentina.”Report 21557-AR: 81, ity Deficit in Latin America.” Paper presented at the Institutions, Latin America and The Caribbean: Argentina, Chile, Paraguay Accountability, and Democratic Governance in Latin America and Uruguay Country Management Unit, World Bank, Washing- conference, the Helen Kellogg Institute for International Studies, ton, DC, September. University of Notre Dame, May 8–9. ———. 2002a. Empowerment and Poverty Reduction: A Sourcebook. Smulovitz, C., and E. Peruzzotti. 2000. “Societal Accountability: The Washington, DC: World Bank. Other Side of Control.” Journal of Democracy 11 (4). October. ———. 2002b. “Uruguay—Health Sector Development Project Smulovitz, C., M. Walton, and P. Petesch. 2003. “Notes on Evaluat- Implementation Completion Report.” Report 38, World Bank, ing Empowerment.” Photocopy, World Bank, Latin America and Washington, DC. Caribbean Region, Washington, DC, September. ———. 2002c. “Rural Water Supply and Sanitation Toolkit for Solo, T. M. 2003. “Independent Water Entrepreneurs in Latin Multisector Projects.”Photocopy, World Bank, Washington, DC. America. The Other Private Sector in Water Services.” Photo- ———. 2003a. Making Services Work for Poor People. World Devel- copy, Private Sector Development and Infrastructure, Energy, opment Report 2004. Washington, DC: World Bank. and Water Department, and Department of Finance, World ———. 2003b. “Brazil Municipal Education: Resources, Incentives Bank, Washington, DC. and Results.”Photocopy, World Bank, Washington, DC. Spink, P.1999.“Possibilities and Political Imperatives: Seventy Years ———. 2003c. “Venezuela—Health Service Reform Project Imple- of Administrative Reform in Latin America.” In Reforming mentation Completion Report.” Report 54, World Bank, Wash- the State: Managerial Public Administration in Latin America, ed. ington, DC. L. C. Bresser Pereira and P. Spink. Boulder: Lynne Rienner ———. 2003d. “State-Society Synergy for Accountability.” Photo- Publishers. copy, Civil Society Team, Latin America and the Caribbean Swope, J., M. Latorre, and F. Celedón. 1998. “Estudio Eficiencia Region, World Bank, Washington, DC, April. Interna de las Escuelas Formales de Fe y Alegría.”Santiago: CIDE ———. 2004. “Voice, Eyes and Ears: Social Accountability in Latin (Centro de Investigación y Desarrollo de la Educación). America.” Civil Society Team, Latin America and the Caribbean Tendler, J. 1997. Good Government in the Tropics. Baltimore and Region, World Bank, Washington, DC. London: The Johns Hopkins University Press. World Development Indicators (WDI). World Bank Database. ———. 2000. “The Rise of Social Funds: What Are They a Model http://devdata.worldbank.org/dataonline. Of?” Photocopy, MIT, Boston. Zakaria, F. 1997. “The Rise of Illiberal Democracy.” Foreign Affairs 76 (6): 22–43.

ocial and infrastructure services have expanded significantly in Latin America since the early 1980s. But the poor continue to experience coverage gaps and S poor-quality services. The authors of Citizens, Politicians, and Providers: The Latin American Experience with Service Delivery Reform argue that access and quality issues are closely tied to the behaviors of both the people providing services and those seeking them. Are the former accountable to the latter?

Building on the accountability framework developed in the World Bank’s World Development Report 2004: Making Services Work for Poor People, this study analyzes how two decades of political, administrative, and social transformations have changed accountability relations in Latin American countries. It describes how democratization, decentralization, privatization, and civil society participation have improved service delivery and identifies the causes of service failures.

By shedding light on the factors constraining delivery of the most essential services to the most vulnerable populations, Citizens, Politicians, and Providers provides guidance to policymakers and development practitioners on how to shape public action to get better-quality services for all.

ISBN 0-8213-6089-2