.4o it ? 4 'E: !VW IC .-43-tirf Zq NAVN, -4. Q-C vW W, 00-k.- Av", AP,-, 'k. A Z. m vi 006 4 -n -4 I' 2600 September 2003 f7 -I; N z-, A- Ag :,J, M, K Avz !.V[. - . , . - .3 A.. g -N, W s! 7i fl-,ntP dk _4 4L L - na JF _Z,0 :5V 4 'Rw q- - ? % 51,,!,%.-,,J;", Ic aa V %:i.4- -l, world development repori90Q4 Making Services Work for Poor People Overview The World Bank Washington, D.C. © 2003 The International Bank for Reconstruction and Development/The World Bank 1818 H Street, N.W. Washington, D.C. 20433 Telephone 202-473-1000 Internet www.worldbank.org E-mail feedback@worldbank.org All rights reserved. Cover and interior design: Susan Brown Schmidler. Cover photographs, from left to right: Nurse in Rwanda showing a newborn infant; © David Turnley. A street child in New Delhi drinks water from a tap; © Reuters NewMedia/CORBIS. Children learning in an elementary school in Myanmar; © Chris Lisle/CORBIS. This document summarizes the World Development Report 2004, a copublication of the World Bank and Oxford University Press. It is a product of the staff of the World Bank. The findings, interpretations, and conclusions expressed herein do not necessarily reflect the views of the Board of Executive Directors of the World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this publication and accepts no responsibility whatsoever for any consequence of their use. Rights and Permissions The material in this work is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The World Bank encourages dis- semination of its work and will normally grant permission 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, www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, World Bank, 1818 H Street NW, Washington, DC 20433, fax 202- 522-2422, e-mail pubrights@worldbankorg. ISBN 0-8213-5637-2 Contents of the World Development Report 2004 1 Services can work for poor people 5 Citizens and politicians but too often they fail Citizen voice and political accountability Outcomes are substantially worse for poor people for service delivery Affordable access to services is low-especially The politics of providing public services for poor people to poor people Quality-a range of failures Beyond the ballot box: citizen initiatives U ~~~~~~~~~~~to increase accountability Making services work to improve outcomes t nraeacutblt Information strategies to strengthen voice spotlight on Progresa Decentralization to strengthen voice Citizen voice in eight sizes 2 Governments should make sGvervi es n sh spotlight on the Kecamatan Development services workc Program A public responsibilityPrga A public responsibflity ~spotlight on Norway and Estonia Growth, though essential, is not enough s o More public spending alone is not enough 6 Policymakers and providers Technical adjustments without changes Compacts, management, and the "long route" in incentives are not enough of accountability Understanding what works and why-to improve Increasing accountability: separating the services policymaker from the provider spotlight on Kerala and Uttar Pradesh Limits to accountability Overcoming the limits 3 The framework for service provision Provider incentives in eight sizes An analytical framework: actors and accountabilities Scaling up, scaling back, and wising up Why establishing relationships of accountability is so complex spotlight on Cambodia* Successes and failures of the public sector and the market 7 Basic education services From principles to instruments Common problems of service provision Reforming institutions to improve services For higher-quality systems, strengthen the for poor people will be difficult relationships of accountability Citizens and clients, politicians and policymakers: spotlight on Uganda voice Policymakers and organizational providers: 4 Clients and providers compacts When will strengthening the client-producer link Organizational and frontline providers: matter most? management Increasing client power through choice Client power Policies to improve choice Getting reform going Increasing consumer power through participation spotlight on Educoe Client power in eight sizes spotlight on the Bamako Initiative iii iv WORLD DEVELOPMENT REPORT 2004 8 Health and nutrition services 10 Public sector underpinnings of service The health of poor people reform Market failures and government failures Strengthening the foundations of government Strengthening client power Spending wisely Strengthening poor citizens' voice Decentralizing to improve services Compacts: provider incentives to serve the poor Making, managing, and implementing Six sizes fit all? good policies Curbing corruption in service delivery spotlight on Costa Rica and Cuba Managing transitions: overcoming reform hurdles 9 Drinking water, sanitation, Evaluating and learning and electricity spotlight on Ceara The state of water and sanitation services Infrastructure and the accountability framework 111 Donors and service reform for service delivery Aid and accountability Urban water networks Strengthen-don't weaken-the compact Rural areas: network and non-network systems Let provider organizations manage Sanitation Increase client power Electricity Promote voice Moving the reform agenda forward Align aid delivery with service delivery spotlight on Johannesburg-* Why reforming aid is so difficult Bibliographical note Selected World Development Indicators 2004 *These spotlights are reprinted at the end of the overview text. Overview K l 'oo often, services fail poor people- ization to local governments, community in access, in quantity, in quality. But participation, and direct transfers to house- the fact that there are strong exam- holds. There have been spectacular suc- ples where services do work means govern- cesses and miserable failures. Both point to ments and citizens can do better. How? By the need to strengthen accountability in putting poor people at the center of service three key relationships in the service deliv- provision: by enabling them to monitor and ery chain: between poor people and discipline service providers, by amplifying providers, between poor people and policy- their voice in policymaking, and by makers, and between policymakers and strengthening the incentives for providers to providers. Foreign-aid donors should rein- serve the poor. force the accountability in these relation- Igo to collect waterfour times a Freedom from illness and freedom from ships, not undermine it. day, in a 20-liter clay jar. It's hard illiteracy-two of the most important ways Increasing poor clients' choice and partic- work! . y I've never been to school poor people can escape poverty-remain ipation in service delivery will help them as I have to help my mother with elusive to many. To accelerate progress in monitor and discipline providers. Raising her washing work so we can earn human development, economic growth is, poor citizens' voice, through the ballot box doesn'tghave ambathroom...O h ff of course, necessary. But it is not enough. and widely available information, can could alter my life, I would really Scaling up will require both a substantial increase their influence with policymakers- like to go to school and have more increase in external resources and more and reduce the diversion of public services to clothes. effective use of all resources, internal and the non-poor for political patronage. By Elma Kassa, a 13-year-old girl external. As resources become more produc- rewarding the effective delivery of services from Addis Ababa, Ethiopia tive, the argument for additional resources and penalizing the ineffective, policymakers becomes more persuasive. And external can get providers to serve poor people better. resources can provide strong support for Innovating with service delivery changes in practice and policy to bring arrangements will not be enough. Societies about more effective use. The two are should learn from their innovations by sys- complementary-that is the essence of the tematically evaluating and disseminating development partnership that was cemented information about what works and what in Monterrey in the spring of 2002. doesn't. Only then can the innovations be The World Development Report 2004 scaled up to improve the lives of poor peo- builds an analytical and practical frame- ple around the world. work for using resources, whether internal The challenge is formidable, because or external, more effectively by making ser- making services work for poor people vices work for poor people. We focus on involves changing not only service delivery those services that have the most direct link arrangements but also public sector institu- with human development-education, tions. It also involves changing the way health, water, sanitation, and electricity. much foreign aid is transferred. As govern- Governments and citizens use a variety ments, citizens, and donors create incen- of methods of delivering these services- tives for these changes, they should be selec- central government provision, contracting tive in the problems they choose to address. out to the private sector and nongovern- They should be realistic about implementa- mental organizations (NGO)s, decentral- tion difficulties. And they should be patient. 2 WORLD DEVELOPMENT REPORT 2004 The proMbem where many of the world's poor live.' But Poverty has many dimensions. In addition the world is off track in reaching the goals to low income (living on less than $1 a day), for primary education, gender equality, and illiteracy, ill health, gender inequality, and child mortality. environmental degradation are all aspects To reach all of these goals, economic Figure 1 Progress in human of being poor. This is reflected in the Mil- growth is essential. But it will not be development: off track lennium Development Goals, the interna- enough. The projected growth in per capita People living on less than $1 a day tional community's unprecedented agree- GDP will by itself enable five of the world's Percent 30 ment on the goals for reducing poverty (box six developing regions to reach the goal for 1). The multidimensional nature of poverty reducing income poverty (table 1). But that is also reflected in the World Bank's two- growth will enable only two of the regions 20 \pronged strategy for development-invest- to achieve the primary enrollment goal and ing in people and improving the investment none of them to reach the child mortality 10 climate. That five of the eight goals and one goal. If the economic growth projected for of the two prongs of the strategy for devel- Africa doubles, the region will reach the 0 opment concern health and education sig- income poverty goal-but still fall short of 1990 1995 2000 2005 2010 2015 nals how central human development is to the health and education goals. In Uganda, Primary school completion rate human welfare. despite average annual per capita GDP Percent But progress in human development has growth of 3.9 percent in the past decade, loo lagged behind that in reducing income child mortality is stagnating-and only poverty (figure 1). The world as a whole is partly due to the AIDS epidemic.2 90 on track to achieve the first goal-reducing Because growth alone will not be enough by half the proportion of people living on to reach the goals, the international com- less than $1 a day-thanks mainly to rapid munity has committed itself-in a series of 80 economic growth in India and China, recent meetings in Monterrey, Doha, and 70 1990 1995 2000 2005 2010 2015 B 0 X 1 The eight Millennium Development Goals Ratio of girls to boys in primary and secondary school With starting points in 1990, each goal is to be S. Improve maternal health Girls as a percent of boys reached by 2015: Reduce by three-quarters the maternal mortality 100 1. Eradicate extreme poverty and hunger ratio. 95 Halve the proportion of people living on less 6. Combat HIV/AIDS, malaria, and other diseases than one dollar a day. Reverse the spread of HIV/AIDS. 90 Halve the proportion of people who suffer from 7. Ensure environmental sustainability hunger. Integrate sustainable development into country 85 2. Achieve universal primary education policies and reverse loss of environmental Ensure that boys and girls alike complete resources. 80 primary schooling. Halve the proportion of people without access 1990 1995 2000 2005 3. Promote gender equality and empower to potable water. women Significantly improve the lives of at least 100 mil- nDeathspera 1 livebirt hsy rateEliminate gender disparity at all levels of educa- lion slum dwellers. Death pr100lvbitstion. 8. Develop a global partnership for 4. Reduce child mortality development 80 Reduce by two-thirds the under-five mortality Raise official development assistance. 60 \ rate. Expand market access. 40 Three points about the Millennium Development Goals: First, to be enduring, success in reaching the goals must be based on systemwide reforms to support progress. Second, focusing on these outcomes does not 20 imply focusing on education and health services alone. Health and education outcomes depend on too many other factors for that to work-everything from parents' knowledge and behavior, to the ease and safety of 0 reaching a health clinic or school, or the technology available for producing outcomes.Third, in countries that 1990 1995 2000 2005 2010 2015 have already achieved universal primary completion or low infant and maternal mortality rates, the spirit of Note: Blue line is the trend line to reach the the Millennium Development Goals-time-bound, outcome-based targets to focus strategies-remains Millenium Development Goal. The red line important. shows the actual progress to date. Source: www.developmentgoals.org. Overview 3 Table 1 Economic growth alone is not enough to reach all the Millennium Development Goals People living on less than $1 a day Primary school completion rate Under-five mortality Annual average Target 2015 growth Target 2015 growth Target 2015 growth GDP per capita (percent) alone (percent) (percent) alone (percent) (per 1,000 alone (per f growth births) 1,000 births) 2000-2015* (percent per year) East Asia 5.4 14 4 100 100 19 26 Europe and Central Asia 3.6 1 1 100 100 15 26 Latin American and the Caribbean 1.8 8 8 100 95 17 30 Middle East and North Africa 1.4 1 1 100 96 25 41 South Asia 3.8 22 15 100 99 43 69 Africa 1.2 24 35 100 56 59 151 *GDP growth projections from World Bank (2003a). Note: Elasticity assumed between growth and poverty is -1.5; primary completion rate is 0.62; under-five mortality is -0.48. Sources. World Bank (2003a), Devarajan 12002j. Johannesburg-to greater resource trans- public system. But at other times, they are fers by developed countries and better poli- clearly failing poor people. cies and institutions in developing coun- tries. The level of resource transfers is Services arefailingpoorpeople difficult to calculate precisely. Some esti- in four ways mates are converging around a figure of $40 How do we know that these services are fail- billion to $60 billion a year in additional ing poor people? First, while governments foreign aid-so long as the money is devote about a third of their budgets to accompanied by policy and institutional health and education, they spend very little of reforms to enhance the productivity of it on poor people-that is, on the services domestic and external resources.3 poor people need to improve their health and Focusing on the human development education. Public spending on health and goals, this Report describes the reforms in education is typically enjoyed by the non- services needed to achieve them. Ensuring poor (figure 3). In Nepal 46 percent of educa- basic health and education outcomes is the tion spending accrues to the richest fifth, only responsibility of the state (box 2). But many 11 percent to the poorest. In India the richest governments are falling short on their fifth receives three times the curative health obligation, especially to poor people. In Armenia and Cambodia, child mortality rates for the poorest fifth of the population are two to three times those for the richest B O X 2 Services-apublic responsibility fifth. Only about 60 percent of the adoles- cents in the poorest fifth of the population By financing, providing, or regulating the rights.The Universal Declaration of Human services that contribute to health and edu- Rights asserts an individual's right to"a stan- in the Arab Republic of Egypt and Peru cation outcomes, governments around the dard of living adequate for the health and have completed primary school, while all world demonstrate their responsibility for well-being of himself and of his family, those from the richest fifth have (figure 2). the health and education of their people. including. .. medical care . .. [and a right to To meet this responsibility, govern- Why? First, these services are replete with education that is] . . . free, at least in the ele- To meet this responsibility, govern- market failures-with externalities, as when mentary and fundamental stages." No mat- ments and citizens need to make the ser- an infected child spreads a disease to play- ter how daunting the problems of delivery vices that contribute to health and educa- mates or a farmer benefits from a may be, the public sector cannot walk away tion-water, sanitation, energy, transport, neighbor's ability to read. So the private sec- from health and education.The challenge is health, and education-work for poor tor, left to its devices, will not achieve the to see how the government-in collabora- healh, .level of health and education that society tion with the private sector, communities, people. Too often, these services are failing. desires. Second, basic health and basic edu- and outside partners-can meet this funda- Sometimes, they are failing everybody- cation are considered fundamental human mental responsibility. except the rich, who can opt out of the 4 WORLD DEVELOPMENT REPORT 2004 Figure 2 The poor are less likely to start school, more likely to drop out care subsidy of the poorest fifth.4 Even 15- to 19-year olds who have completed each grade or higher though clean water is critical to health out- Niger 1998 Egypt2000 comes, in Morocco only 11 percent of the Percent Percent poorest fifth of the population has access to 100 100 - safe water, while everybody in the richest fifth 80 80 Richest fif does (figure 4). 60 60 Second, even when public spending can Richestfifth 60% be reallocated toward poor people-say, by 40 40 shifting to primary schools and clinics-the 20 6.4% 20 Poorestfifth money does not always reach the frontline corn leted service provider. In the early 1990s in 0 0 Uganda the share of nonsalary spending on 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 primary education that actually reached India 1998-99 Grade Peru2000 Grade primary schools was 13 percent. This was Percent Percent the average: poorer schools received well 100 100 below the average.5 80 80 Richestfifth Third, even if this share is increased-as Richest fifth > the Ugandans have done-teachers must be 60 60 67% present and effective at their jobs, just as 40 40 doctors and nurses must provide the care 36%o~ \ that patients need. But they are often mired 20 20 Poorest fifth in a system where the incentives for effective 0 Poorest fifth service delivery are weak, wages may not be 1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9 paid, corruption is rife, and political patron- Grade Grade age is a way of life. Highly trained doctors Note:The grade number boldfaced denotes the end of the primary cycle (grade5 for Egypt and India; seldom wish to serve in remote rural areas. grade 6 for Niger and Peru). Source: Analysis of Demographic and Health Survey data. Since those who do serve there are rarely monitored, the penalties for not being at work are low. A survey of primary health care facilities in Bangladesh found the absentee rate among doctors to be 74 per- Figure 3 More public spending for the rich than for the poor Share of public spending that accrues to the richest and poorest fifths All health spending Primary health All education spending Primary education Guinea Nepal 1994 LL 11996L Armenia L199, Kosovo / Percent richest Armyenia Kosovo q uintile 1999 -.~. 2000 *'~- Percent poorest Ecuador Nicaragua quintile 1998 1998 India Cambodia 1995/96 1 1 1996/97 -_. Cote d'lvoire Brazil 1997 - 1995 (NE&SE} E = Madagascar Morocco 1993 L 1998/99 Bangladesh Mexico ] 2000 L = 1996 Li Bulgaria [ 1 1 Kenya E: 1995 L 1992 ea Costa Rica Romania - 1992 E 1994 ^ 0 20 40 0 20 40 0 20 40 0 20 40 Percent Percent Percent Percent Source: Compiled from various sources by World Bank staff. Overview 5 cent.' When present, some service providers Figure 4 Water, water everywhere, nor any drop treat poor people badly. "They treat us like Percent of households who use an improved water animals," says a patient in West Africa. source, poorest and richest fifths By no means do all frontline service Poorest Richest providers behave this way. Many, often the fifth fifth majority, are driven by an intrinsic motiva- Ethiopia 2000 6 tion to serve. Be it through professional Morocco 1995 0 o pride or a genuine commitment to help Guinea 1999 * poor people (or both), many teachers and health workers deliver timely, efficient, and Cambodia 2000 * - courteous services, often in difficult cir- Kazakhstan 1999 cumstances-collapsing buildings, over- flowing latrines-and with few resources- Nicaragua 1998 clinics without drugs, classes without Indonesia 1997 O textbooks.8 The challenge is to reinforce this experience-to replicate the professional Tanzania 1999 ethics, intrinsic motivation, and other Philippines 1998 * incentives of these providers in the rest of the service work force. The fourth way services fail poor people India 1998-99 is the lack of demand. Poor people often Uzbekistan 1996 - don't send their children to school or take 0 20 40 60 80 100 them to a clinic. In Bolivia 60 percent of the them toa clini. In Boivia 60percentof the Note:The poorest fifth in one country may correspond to the stan- children who died before age five hiad inot dard of lining in the middle fifth in another country. Within-country seen a formal provider during the illness inequalities reflect inequality in access to water and in the variable used to construct quintiles. An "improved" water source, as culminatung tn their death. Sometimes the defined by UNICEF, provides adequate quality and quantity of reason is the poor quality of the service- water (that is, a hosehold connection or a protected well, notesn unprotected well or bottled water), missing materials, absent workers, abusive Source: Analysis of Demographic and Health Survey data. treatment. At other times it is because they are poor. Even when the services are free, history and aroundtheorldsocietieshave many poor rural families cannot afford the tried different arrangements-with mixed time it takes to travel the nearly 8 kilome- results ters to the nearest primary school in Mali or *Some governments contract services the 23 kilometers to the nearest medical out-to the private sector, to nongovern- facility in Chad.9 mental organizations (NGOs), even to WAfeak demand can also be due to cultural other public agencies. In the aftermath of factors, notably gender. Some parents refuse a civil war Cambodia introduced two to send their daughters to school. Husbands forms of contracting for the delivery of have been known to prevent their wives primary health care ("contracting out" from going to clinics-even for deliveries. whole services and "contracting in" some And the social distance between poor peo- services). Randomly assigning the ar- ple and service providers (70 percent of rangements across 12 districts (to avoid nurses and midwives in rural Niger had systematic bias), it found that health indi- been raised in the city) is often a deterrent. cators, as well as use by the poor, increased most in the districts contracting out.'° Alternative service delivery Whether this can be scaled up beyond 12 arrangements districts in Cambodia is worth exploring. Ensuring access to basic services such as *Governments also sell concessions to the health, education, water, energy, and sanita- private sector-in water, transport, tion is a public responsibility today, but it electricity-with some very good and has not always been. Nor do governments some very bad results. Privatizing water discharge this responsibility solely through in Cartagena, Colombia, improved ser- central-government provision. Throughout vices and access for the poor. A similar 6 WORLD DEVELOPMENT REPORT 2004 sale in Tucuman, Argentina, led to riots The framewcork of relationships- in the streets and a reversal of the con- between c63ents, providers, cession. Overall, though, privatizing and policymakers water in Argentina, by expanding access of poor communities to clean water and To help understand the variety of experiences of por comunites toclea wate and with traditional and alternative service deliv- sewage treatment, appears to have had a he service derve chain favorable effect on health. One study ery arrangements, the service delivery chain estimates that it prevented about 375 can be unbundled into three sets of actors, deaths of young children a year." and the relationships between them exam- ined (figure 5). Poor people-as patients in Some societies transfer responsibilty clinics, students in schools, travelers on buses, (for financing, provision, and regula- consumers of water-are the clients of ser- tion) to lower tiers of government, vices. They have a relationship with the front- Again, the record has varied-with yl potentially weaker capacity and greater line providers, with schoolteachers, doctors, potenticall waker cacty lan grelate bus drivers, water companies. Poor people potcal patronage at the local level and have a simlar relationship when they buy somethe mred soptweighforthe re ributi something in the market, such as a sandwich sometimeS outweighing the benefits (or a samosa, a saltena, a shoo-mai). In a from greater local participation. Local- . a ' ' government delivery of infrastructure in 'competitive-market transaction, they get the South Africaimproved service provision "service" because they can hold the provider in a short time.'12 But decentralizing accountable. That is, the consumer pays the social assistance in Romania weakened provider directly; he tan observe whether or the ability and incentives of localcoun- not he has received the sandwich; and if he is cilsthdeablityr cand irancfen s tofthe loal con3 dissatisfied, he has power over the provider cihs tod ram csh tnsfbeingrscentoahepr.' with repeat business or, in the case of fraud, The program is now being recentralized. with legal or social sanctions. Responsibility is sometimes transferred to For the services considered here-such communities-or to the clients them- as health, education, water, electricity, and selves. El Salvador's Community-Man- sanitation-there is no direct accountabil- aged Schools Program (Educo) gives par- ity of the provider to the consumer. Why ents' associations the right to hire and fire not? For various good reasons, society has teachers. That, plus the monthly visits to decided that the service will be provided the schools by the parents' associations, not through a market transaction but has reduced teacher-and student- through the government taking responsibil- absenteeism, improving student perfor- ity (box 2). That is, through the "long mance. route" of accountability-by clients as citi- Still other programs transfer resources zens influencing policymakers, and policy- and responsibility to the household. Mex- makers influencing providers. When the ico's Education, Health, and Nutrition relationships along this long route break Program (Progresa) gives cash to families down, service delivery fails (absentee teach- if their children are enrolled in school and ers, leaking water pipes) and human devel- they regularly visit a clinic. Numerous opment outcomes are poor. evaluations of the program show consis- tently that it increased school enrollment Figure 5 The framework of accountability (eight percentage points for girls and five relationships for boys at the secondary level) and improved children's health (illness among young children fell 20 percent).'4 With all these innovations, of course, comes the challenge of understanding what works-where, how, and why-so that, with enough resources, successful results - _ can be replicated on a broad scale. - - Overview 7 Consider the first of the two relationships is a well-functioning electoral system, poor along the long route-the link between poor people may not be able to influence politi- people and policymakers or politicians (fig- cians about public services: they may not be ure 5). Poor people are citizens. In principle, well informed about the quality of public _ they contribute to defining society's collective services (and politicians know this); they objectives and they try to control public may vote along ethnic or ideological lines, action to achieve those objectives. In practice, placing less weight on public services when this does not always work. Either they are evaluating politicians; or they may not excluded from the formulation of collective believe the candidates who promise better objectives or they cannot influence public public services-because their term in office action because of weaknesses in the electoral is too short to deliver on the promise-and system. Free public services and "no-show" they vote instead for candidates who provide jobs are handed out as political patronage, ready cash and jobs. with poor people rarely the beneficiaries. As a result, public services often become Even if poor people can reach the policy- the currency of political patronage and clien- maker, services will not improve unless the telism. Politicians give "phantom" jobs to policymaker can ensure that the service teachers and doctors. They build free public provider will deliver services to them. In schools and clinics in areas where their sup- Cambodia, policymakers were able to specify porters live. Former Boston mayor James the services required to the NGOs with Curley strengthened his political base by con- whom they contracted. But for many ser- centrating public services in the Irish Catholic vices, such as student learning or curative areas while denying them to the Protestants, care, the policymaker may not be able to who eventually moved to the suburbs.'5 specify the nature of the service, much less In 1989 Mexico introduced PRONASOL impose penalties for underperformance of (Programa Nacional de Solidaridad, or the contract. Teacher and health-worker National Solidarity Program), a poverty alle- absenteeism is often the result. viation program that spent 1.2 percent of Given the weaknesses in the long route of GDP annually on water, electricity, nutrition, accountability, service outcomes can be and education construction in poor commu- improved by strengthening the short route- nities. Assessments of the six-year program by increasing the client's power over found that it reduced poverty by only about providers. School voucher schemes (Colom- 3 percent. Had the budget been distributed bia's PACES) or scholarships (Bangladesh's to maximize its impact on poverty, the Female Secondary School Assistance Pro- expected decline would have been 64 percent gram, in which schools receive a grant based with perfect targeting. It would have been 13 on the number of girls they enroll) enable percent even with an untargeted, universal clients to exert influence over providers proportional transfer to the whole popula- Figure6 ItpaidtovoteforPRI through choice. El Salvador's Educo program tion. The reason becomes apparent when PRONASOL expenditures according to party and Guinea's revolving drug scheme (where one examines the political affiliation of com- in municipal government co-payments inspired villagers to stop theft) munities that received PRONASOL spend- Average expenditures per capita are ways for client participation to improve ing. Municipalities dominated by the Institu- (real 1995 pesos) service provision (see the spotlight on tional Revolutionary Party (PRI), the party 400 Educo). in power, received significantly higher per 300 PRI Turn now to a closer look at the individ- capita transfers than those voting for another PRO ual relationships in the service delivery party (figure 6).16 200 Other chain-why they break down, how they can just as a well-functioning democracy does P be strengthened. not guarantee that poor people will benefit 100 from public services, some one-party states 0 Citizens andpoliticians! get good health and education outcomes- 1989 1990 1991 1992 1993 1994 policymakers-stronger voice even among the poor. Cuba has among the Poor citizens have little clout with politi- best social indicators in Latin America-at a PRD = Party of the Democratic Revolution; PAN cians. In some countries the citizenry has much lower income than its peers, such as = National Action Party. Source: Estovez, Magaloni. and Diaz-Cayeros only a weak hold on politicians. Even if there Chile and Costa Rica. China has reduced (2002). 8 WORLD DEVELOPMENT REPORT 2004 infant mortality dramatically, and achieved dismal, and know that everyone else knows nearly universal primary enrollment. To be that-and yet most do not feel free to com- sure, in China, cases during the earliest phase plain.19 of the outbreak of severe acute respiratory syndrome (SARS) in 2002 were not openly Policymakers and providers- reported, thus making its further spread stronger compacts almost inevitable.'7 And Cubans, who had Strengthening poor people's voice can make high levels of health and education in the policymakers want to improve services for 1950s, remain poor on other dimensions. the poor. But they still may not be able to. The lesson seems to be that the citizen- Well-intentioned policymakers often can- policymaker link is working either when not offer the incentives and do the monitor- citizens can hold policymakers accountable ing to ensure that providers serve the poor. for public services that benefit the poor or The absenteeism of teachers, the rude treat- when the policymaker cares about the ment of patients, the siphoning of pharma- health and education of poor people. These ceuticals are symptoms of the problem. politics are "pro-poor." Even in the private sector, where the What can be done when the politics are incentives presumably are better aligned, not pro-poor? Societies can still introduce performance is not much better-for the various intermediate elements to make same reasons that private marlkets are not public institutions more accountable. Par- the solution to these problems in the first ticipatory budgeting in Porto Allegre, place. Private providers fail to reach the very Brazil, started as a means for the citizens to poor. Weak regulation leads to poor-quality participate in budget formulation and then health services in India's private sector. to hold the municipal government account- Ineffectively privatizing water incites riots able for executing the budget. in the streets of Cochabamba. Perhaps the most powerful means of In the former Soviet Union, state and increasing the voice of poor citizens in poli- party control over providers ensured com- cymaking is better information. When the pliance with delivery norms for free ser- government of Uganda learned that only 13 vices. Services worked, and levels of health percent of recurrent spending for primary status, particularly for the poorer Central education was arriving in primary schools, Asian republics, were much higher than it launched a monthly newspaper campaign other countries at their level of income. But on the transfer of funds. That campaign the breakup of the Soviet Union weakened galvanized the populace, inducing the gov- state control over providers, and health and ernment to increase the share going to pri- education services collapsed. mary schools (now over 80 percent) and Solving the problem requires mentally, compelling school principals to post the and sometimes physically, separating the pol- entire budget on the schoolroom door. icymaker from the provider-and thinking The media can do much to disseminate of the relationship between the two as a com- information about public services. Higher pact. The provider agrees to deliver a service, newspaper circulation in Indian districts is in return for being rewarded or penalized associated with better local-government depending on performance. The compact performance in distributing food and may be an explicit contract with a private or drought relief.'8 The more people who can nonprofit organization-or between tiers of read, the stronger the influence of the governrment, as in Johannesburg, South media. In Kerala, India, this led to a virtu- Africa (see the spotlight on Johannesburg). ous cycle of literacy leading to better public Or it could be implicit, as in the employment services, which raised literacy even more. agreements of civil servants. But information is not enough. People Separating the policymaker from the must also have the legal, political, and eco- provider is not easy, for those who benefit nomic means to press demands against the from the lack of separation may resist it. government. Most citizens in Uttar Pradesh, Teachers' unions in Uttar Pradesh, India, India, know that government services are blocked an attempt to put teacher hiring, Overview 9 firing, and attendance under the control of having that information reach the policy- the village panchayat. On the other hand, maker. New technologies, including e-gov- health professionals in Brazil participated ernment, can make this easier.24 in a national coalition that prepared the So can some ingenious methods using plan for health reforms and municipal human beings. When Ceara, Brazil, hired a health councils.20 The separation usually cadre of district health workers, the govern- happens because of a fiscal crisis (Johannes- ment sent their names to the applicants burg), a major political change (decentral- who were not selected, inviting them to ization in Latin America), or a legacy of his- report any problems with service in the tory (public regulation of water providers health clinics. More fundamentally, these in the Netherlands). output-based incentive schemes require rig- Even with a separation of policymaker orous program evaluation, so that the poli- and provider, the compacts cannot be too cymaker knows and understands what is explicit. It is difficult to specify precisely working and what isn't. Evaluation-based what the schoolteacher should do at every information, important not only for moni- point in the day. Too much specificity can toring providers, also enables the rest of the lead to inflexibility. Parisian taxi drivers, to world to learn about service delivery. make a point about excessive regulations, sometimes miieticulously follow the rules in Clients and providers-more the Code de la route-slowing traffic in the choices, more participation French capital to a snail's pace.2' Given the difficulties in strengthening the Since the contract cannot be fully speci- long route of accountability, improving the fied, policymakers look to other means of short route-the client-provider relation- eliciting pro-poor services from providers. ship-deserves more consideration. There is One way is to choose providers who have an no question that this relationship is broken intrinsic motivation to serve the poor. A for hundreds of millions of poor people. study of faith-based health care providers in Voices of the Poor and other surveys point to Uganda estimates that they work for 28 per- the helplessness that poor people feel before cent less than government and private for- providers-nurses hitting mothers during profit staff, and yet provide a significantly childbirth, doctors refusing to treat patients of higher quality of care than the public a lower caste.25 Unlike most private providers, sector.22 Another way is to increase incen- public water companies funded through bud- tives to serve the poor or work in under- getary transfers often ignore their customers. served areas. But one study of Indonesia These are but symptoms of the larger prob- shows that it would require multiples of cur- lem: many service delivery arrangements rent pay levels to get doctors to live in West neglect the role of clients, especially poor Papua, for instance (where the vacancy rate clients, in making services work better. is 60 percent).23 A third way is to solicit bids Clients can play two roles in strengthening for services and use the competition in the service delivery. First, for many services, bidding process to monitor and discipline clients can help tailor the service to their providers. Many water concessions are man- needs, since the actual mix cannot be specified aged this way. A recent innovation in Mad- in advance. In some parts of Pakistan, girls are hya Pradesh, India, allows NGOs to compete more likely to attend school if there is a female for concessions to primary schools, with teacher. The construction of separate latrines payments conditional on higher test scores for girls has had a strong effect on girls' enroll- based on independent measurement. ment in many primary schools. When the As with the citizen-politician relation- opening hours of health clinics are more con- ship, a critical element in the policymaker- venient for farmers, visits increase. Second, provider relationship is information. The clients can be effective monitors of providers, policymaker can specify a contract based since they are at the point of service delivery. only on what he can observe-on what The major benefit of Educo came from the information is available. There has to be a weekly visits of the community education method for monitoring providers and for association to schools. Each additional visit 10 WORLD DEVELOPMENT REPORT 2004 reduced student absenteeisrm (due to teacher low aquifers-for drinking water. Unfortu- absenteeism) by 3 percent.26 nately, no one arranged for the monitoring How can the role of clients in revealing of water quality-a public good-so the demand and monitoring providers be arsenic in the water went undetected. If the strengthened? By increasing poor people's stakes are high enough, communities tackle choice and participation in service delivery. the problem. When the Zambian govern- When clients are given a choice among service ment introduced a road fund financed by a providers, they reveal their demand by "voting charge on trucks, truck drivers took turns with their feet." Female patients who feel more policing a bridge crossing to make sure that comfortable with female doctors can go to overloaded trucks did not cross. Of course one. The competition created by client choice such co-payments or user fees reduce also disciplines providers. A doctor may refuse demand-and so should not be used when to treat lower-caste patients, but if he is paid the demand effects outweigh the increase in by the number of patients seen, he will be supply, as in primary education. But for concerned when the waiting room is empty. water, electricity, and other services whose Reimbursing schools based on the number of benefits are enjoyed mainly by the user, students (or female students) they enroll cre- charging for them has the added benefit of ates implicit competition among schools for increasing the consumer's incentive to moni- students, increasing students' choice. tor the provider. Farmers in Andhra Pradesh, School voucher programs-as in India, are finding that, when they pay for Bangladesh, Chile, Colombia, C6te d'Ivoire, their water, the irrigation department and the Czech Republic-are explicitly aimed becomes more accountable to them. In the at improving education quality by increasing words of one farmer, "We will never allow parents' choices. The evidence on these the government to again give us free water."'30 schemes is mixed, however. They seem to have improved student performance among some Donors and recipients- groups. But the effects on the poor are strengthening accountability, ambiguous because universal voucher schemes not undermining it tend to increase sorting-with richer students Improving service outcomes for poor peo- concentrating in the private schools.27 When ple requires strengthening the three rela- the voucher is restricted to poor or disadvan- tionships in the chain-between client and taged groups, the effects are better.2" The provider, between citizen and policymaker, Colombian program showed lower repetition and between policymaker and provider. In rates and higher performance on standard- their zeal to get services to the poor, donors ized tests for students participating in the often bypass one or more of these relation- scheme-with the effect for girls higher than ships. The typical mode of delivering aid- that for boys.29 Even in network systems such a project-is often implemented by a sepa- as urban water provision, it is possible to give rate unit outside the compact, bypassing the poor communities choice-by allowing the relationship between policymakers and poor to approach independent providers, providers. The project is typically financed introducing flexibility in service standards by earmarked funds subject to donor-man- such as lifeline rates, and so on. dated fiduciary requirements. It and other When there is no choice of providers, donor initiatives, including global "funds," increasing poor people's participation in ser- bypass the citizen-policymaker relationship vice provision-giving them the ability to where the budget is concerned. To be sure, monitor and discipline the provider, for when the existing relationship is dysfunc- example-can achieve similar results. Clients tional, it may be necessary to go around it. can play the role of monitors since they are But the cases where the benefits outweigh present at the point of service. But they need the costs are probably fewer than imagined. to have an incentive to monitor. Recognizing the gap between ends and In Bangladesh, thanks to reduced import means, some donors and recipients try to tariffs, households were able to purchase use foreign aid to strengthen, not weaken, tubewells that tapped ground sources-shal- the links in the service delivery chain. One Overview approach is to incorporate donor assistance been improving." The reforms detailed in in the recipient's budget, shifting to the this Report (aimed at recipient countries recipient's accountability system. In Uganda and aid agencies) can make aid even more assistance from Germany, Ireland, the productive. When policies and institutions Netherlands, Norway, the United Kingdom, are improving, aid should increase, not and the World Bank is all part of the coun- decrease, to realize the mutually shared try's budget, the outcome of a coordinated objective of poverty reduction, as specified and participatory process. in the Millennium Development Goals. Another approach is for donors to pool At the same time, simply increasing pub- their assistance in a single "pot" and to har- lic spending-without seeking improve- monize their fiduciary standards around ments in the efficiency of that spending-is that of the rest of the government. The sec- unlikely to reap substantial benefits. The torwide approach (SWAp) to health, educa- productivity of public spending varies tion, transport, and other sectors is a step in enormously across countries. Ethiopia and this direction. Possibly the biggest payoff Malawi spend roughly the same amount per comes when donors help generate knowl- person on primary education-with very edge-as when donor-financed impact eval- different outcomes. Peru and Thailand uation studies reveal what works and what spend vastly different amounts-with simi- doesn't in service delivery, or when donors lar outcomes. pool technical assistance resources at the On average, the relationship between retail level, as in the multidonor Water and public spending on health and education Sanitation Program. Knowledge is essential and the outcomes is weak or nonexistent. A to scaling up service delivery. Although it simple scatter plot of spending and out- emerges locally, it is a global public good- comes shows a clear line with a significant precisely what aid is designed to finance. slope-because richer countries spend more on health and education and have What not to do better outcomes. But controlling for the The picture painted so far of the difficulties effect of per capita income, the relationship in government-led service delivery may lead between public spending on health and some to conclude that government should under-five mortality rates is not statistically give up and leave everything to the private significant (figure 7). That is not surprising: sector. That would be wrong. If individuals most public spending on health and educa- are left to their own devices, they will not tion goes to the non-poor, much of it fails provide levels of education and health that they collectively desire (box 2). Not only is Figure 7 Increased public spending is not enough this true in theory, but in practice no coun- try has achieved significant improvement in Under-five mortality rate, 2000 child mortality and primary education 1 without government involvement. Further- 100 * # * more, as mentioned earlier, private sector or * VI* * NGO participation in health, education, and infrastructure is not without problems- 0 + . especially in reaching poor people. The * * ' + * extreme position is clearly not desirable. -50* Some aid donors take a variant of the * S 1 "leave-everything-to-the-private-sector" posi- tion. If government services are performing -150 so badly, they say, why give more aid to -150 -100 -50 0 50 100 150 those governments? That would be equally Per capita public spending on health, 1990s average* wrong. There is now substantial research *Public spending and child mortality are given as the percent deviation from rate predicted by GDP per capita. showing that aid is productive in countries Note: For the regression line shown, the coefficient is -0.148 and with good policies and institutions, and the t-statisiic is 1.45. Source: GDP per capita and public spending data, World those policies and institutions have recently Development Indicators database; under-five mortality, UNICEF 12 WORLD DEVELOPMENT REPORT 2004 to reach the frontline service provider, and even be counterproductive-if there are seri- service providers face weak incentives to ous problems elsewhere in the service deliv- deliver services effectively. ery chain. In water or curative health care, Linked to the "simply increase public tightening the policymaker-provider link spending" approach is one that advocates could make providers respond more to the for more foreign aid without accompanying demands of their superiors-and less to measures to improve the productivity of their poor clients. Relying on user groups, foreign aid. This can be just as misleading- often generously funded by donors, may and not just for the same reasons that sim- inhibit the development of genuinely demo- ply increasing public expenditure is mis- cratic local governments. Finally, countries, leading. Sometimes the modes of delivering and regions within countries, vary enor- foreign aid, by undermining rather than mously in the conditions that make service strengthening service delivery in the recipi- innovations work. A failed state mired in ent country, can reduce the productivity of conflict will be overstretched in resources public spending in the medium run. and institutional capacity, and able to man- Finally, when faced with disappointing age only certain interventions. Countries health and education outcomes, especially with high prevalence of HIV/AlDS will for poor people, it is tempting to recom- require short- and long-term adaptations of mend a technical solution that addresses the service delivery systems. proximate cause of the problem. Why not Does this mean there are no general give vitamin A supplements, de-worm lessons about making services work for schoolchildren, and train teachers better? poor people? No. The experience with ser- Why not develop a "minimum package" of vice delivery, viewed through the lens of this health interventions for everybody? Report, suggests a constellation of solutions, Although each intervention is valuable, rec- each matching various characteristics of the ommending them alone will not address service and the country or region. While no the fundamental institutional problems one size fits all, perhaps eight sizes do. Even that precluded their adoption in the first eight may be too few, which is wvhy some of place.32 Lack of knowledge about the right the "sizes" are adjustable, like waistbands. technical solution is probably not the bind- The eight sizes can be arrived at by ing constraint. What is needed is a set of answering a series of questions. institutional arrangements that will give policymakers, providers, and citizens the Pro-poor or clientelist politics? yes incentives to adopt the solution and adapt it How much is the political system in the to local conditions. country geared toward pro-poor public ser- vices-and how much does it suffer from clientelist politics and corruption? This is the Pro-poor? The varied experience with traditional and most difficult dimension for an outside actor, innovative modes of service delivery clearly such as a donor, to address: the recipient of shows that no single solution fits all services the advice may also be the source of the prob- in all countries. The framework of account- lem. And politics do not change overnight. no ability relationships explains why. In differ- Even so, at least three sets of policy ent sectors and countries, different relation- instruments can be deployed where the pol- ships need strengthening. In education the itics are more clientelist than pro-poor. biggest payoff may come from strengthen- ing the client-provider link, as with vouch- First is choosing the level of government ers in Colombia or scholarships for girls in responsible for the service. Countries Bangladesh. But that may not be so in differ in the patronage politics and capa- immunization campaigns. bilities of different tiers of govern- Furthermore, poor people are often ment-and this should inform the ser- trapped in a system of dysfunctional service- vice delivery arrangement. delivery relationships. Making just one link o Second, if politicians are likely to capture more effective may not be enough-it may the rents from free public services and Overview 13 distribute them to their clients, an extreme case-when individual prefer- arrangement that reduces the rents may ences matter-the appropriate solution leave the poor better off. This might will involve individual choices of service include transparent and publicly known (if there is the possibility of competition) rules for allocation, such as per-student and such interventions as cash transfers, grants to schools, or conditional transfers vouchers, or capitation payments to schools to households, as in Progresa. In some or medical providers. If there are shared cases it may include fees to reduce the preferences, as in education, or free-rider value of the politicians' distribution deci- problems, as in sanitation, the community sions. India's power sector was nationally is the correct locus of decisionmaking. The owned and run because it was a network appropriate policy will then involve local- (and therefore not amenable to head-to- government decisions in a decentralized head competition). But the huge rents setting-or depending on political realities, from providing subsidized electricity community decisions (as for social invest- have been diverted to people who are not ment funds) and user groups (such as par- poor-all within a parliamentary ents in school committees). democracy. Reducing those rents by rais- ing power tariffs or having the private Easy or hard to monitor? sector provide electricity, even if it vio- Services can be distinguished by the diffi- lates the principles of equity-they are culty of monitoring service outputs. The already violated in the existing system- difficulty depends on the service and on the may be the only way of improving elec- institutional capacity of government to do tricity services to the poor. the monitoring. At one extreme are the ser- Third, better information-that makes vices of teachers in a classroom or doctors citizens more aware of the money allo- in a clinic: Both transactions allow much cated to their services, the actual condi- discretion by the provider that cannot be tions of services, and the behavior of observed easily. A doctor has much more policymakers and providers-can be a discretion in treating a patient than an elec- powerful force in overcoming clientelist trician switching on a power grid. And it is politics. The role of a free and vibrant difficult to know when high-quality teach- press and improving the level of public ing or health care is being provided. It may discourse cannot be overstated. be possible to test students. But test scores tell very little about the teacher's ability or yes Homogeneous or heterogeneous effort, since they depend at least as much on clients? students' socioeconomic status or parental Homogeneous? The answer to this question depends on the involvement. More easily monitored are service. Students with disabilities have spe- immunizations and clean latrines-all cial needs for quality education but not for measurable by a quantitative, observable yes immunization. Heterogeneity is also defined indicator. by regional or community preferences. Of course it depends on who is doing Whether a girl goes to school may depend the monitoring. Parents can observe - - on whether there are separate latrines for whether the teacher is in attendance, and Pro-poor? boys and girls. If that depends on local pref- what their children are learning, more eas- erences, the village should have a say in ily than some central education authority. design. Previously homogeneous societies, Better management information systems such as Sweden and Norway, are changing and e-government can make certain ser- no with increased immigration. They are giving vices easier to monitor. And monitoring yes more discretion to local communities in tai- costs can be reduced by judicious choice of loring the education system to suit the lin- providers-such as some NGOs, which --- guistic abilities of their members. may be trustworthy without formal moni- i Homogeneous?] The more that people differ in their toring. In short, the difficulty of monitor- desires, the greater the benefits from ing is not fixed: it can vary over time and no decentralizing the decision. In the most with policies. 14 WORLD DEVELOPMENT REPORT 2004 FigureB Eightsizesfitall? in Cambodia. Infrastructure services could be managed by a national utility or provided by the private sector with regulatory over- yes nit r. I 4,7L) sr i g:1ht. Easy sig to Note that the particular configuration in monitor? no which this arrangement will work is special. yes ( 01`s`r In some developed countries there is much discussion of a set of reforms, started in Homogeneous ? New Zealand, that involve greater use of no | Drfl U iur,a1h fidt | explicit contracts-either from the govern- yes 'lIiUiil! ment to the private sector, or from central yes Easy ministries to the ministries responsible for mntor? monitor? n specific services. The New Zealand reforms are justified by a well-established public sector ethos, reasonable management infor- Pro-poor? mation systems, and supporting institu- | 7j-!itlWriAu ., .@s ,-J. .,1, 1 tions, including legal systems, to allow con- yes | tract enforcement. These features increase Etoy the "monitorability" of certain services by no monitor? reducing the gap between contracted and yes n ll @iTh....jx realized outcomes. '41['i. ,*il, itF,,un.. These preconditions do not exist in Homogeneous ? many developing countries, so the template of these reforms cannot be used mechani- Eas yes ____,________________ cally.3 If there is no good legal system and Easy the civil service is subject to bribes (a form monitor? of clientelist politics), private sector con- no tracts might be a major source of corrup- tion. In these countries, government should perhaps be even more output-oriented- not as a means of tweaking a well-function- Eight sizesfit all ... ing system but as a way of getting the sys- Now examine different combinations of tem to provide much greater improvements these characteristics, to see which service in services and generating new information. delivery arrangement would be a good fit- and which would be a misfit (figure 8). To Central government provision (2). When be sure, none of the characteristics can be the service is difficult to monitor-explicit easily divided into such clean categories, contracts are difficult to write or enforce- because countries and services lie on a con- but the politics are pro-poor and clients tinuum. Even so, by dividing the salient homogeneous, the traditional, centralized characteristics, and looking at various com- public sector is the appropriate delivery sys- binations, the "eight sizes fit all" approach tem. The French education system, which can be applied to the considerations spelled administers a uniform service centrally, is out earlier. one of the best examples.34 But too many countries fall into the trap of thinking that Central government financing with con- just because the service is difficult to moni- tracting (1). In a favorable political con- tor, it must be delivered by the government. text, with agreement on what government When students are heterogeneous, when should do, an easy-to-monitor service such the politics of the country are not geared as immunization could be delivered by the toward poor people, government control of public sector, or financed by the public sec- the education system-with no participa- tor and contracted out to the private or non- tion by students, parents, or local commu- profit sector, as with primary health centers nities-can leave the poor worse off. Overview 15 Local government financing with contract- health services with externalities, such as ing (3). With heterogeneous preferences, immunization. In curative health care, the local governments should be involved in asymmetry of information between client and services. When local politics are pro-poor provider makes strengthening client power (but national politics aren't), local govern- problematic. Better ilnformation on preven- ments could be more reliable financiers of tive care or on how to choose medical services, and vice versa. Easily monitored providers (possibly disseminated by nonprofit services such as water or electricity can be organizations) can ameliorate the problem. In contracted out to public or private utilities, extreme cases, it may be that only community as in Johannesburg. groups or altruistic nonprofits can effectively provide these services to poor people.35 Local government provision (4). For diffi- These service delivery arrangemenits rep- cult-to-monitor services, such as education resent efforts to balance problems with the (for quality), management responsibility long route of accountability (clientelist pol- might be ceded to parent groups when the itics, hard-to-monitor services) with the politics are conducive, as in the Educo pro- short route. The reason societies choose the gram. Giving clients a choice through vouch- long route is that there are market failures ers enables them to express their heteroge- or concerns with equity that make the tra- neous preferences. And the competition ditional short route-consumers' power created by clients having a choice may improve over providers-inadequate. But the "gov- service quality-as with water vouchers in ernment failures" associated with the long Chile or sanitation vouchers in Bangladesh. route may be so severe that, in some cases, the market solution may actually leave poor Client power (5, 6, 7, 8). When publicly people better off. financed services are subject to capture-the politics are not pro-poor-the best thing to ... with adjustable waistbands do is to strengthen the client's power as The foregoing simplified scheme captures much as possible. But that can be difficult. only part of the story. At least two features are Even means-tested voucher schemes or sub- left out. sidies could be diverted to the non-poor. Transparent, rule-based programs, such as Failed states. Countries where the state is Progresa in Mexico, are needed to make it failing (often countries in conflict) need difficult to hide middle-class capture. service delivery arrangements different from In services such as water and electricity, those where the state is fairly strong. Primary governments intervene to regulate monop- school completion rates in Senegal and the oly providers and protect the poor-and Democratic Republic of Congo are about 40 not because there are significant externali- percent. In Senegal-a stable democracy- ties. So separating the policymaker from the the reforms in education, including those that provider, and making the provider account- strengthen client-provider links, would go able to the client through prices, can through the government (to strengthen the strengthen client power and lead to better policymaker-provider links as well). In the results. Poor people can be protected from Democratic Republic of Congo-where con- high prices if charges rise with use (with an flict has significantly weakened the state- initial free amount). Allowing small, inde- ways should be found to empower communi- pendent water providers to compete with ties to improve education services-even if it the local monopoly can also discipline pro- means bypassing government ministries in vision and keep prices down. the short to medium term. Social funds and But prices-without accompanying subsi- community-driven development are exam- dies or transfers to poor people-cannot be ples. They can be effective in improving ser- used to strengthen client power in education vice outcomes, but concerns about their sus- because of the externalities in primary educa- tainability and scalability-and whether they tion. A market-based allocation would not be crowd out the growth of local government in society's interest. The same applies to capacity-should not be overlooked. 16 WORLD DEVELOPMENT REPORT 2004 History. The country's history can also tures, and disseminate information about have a bearing on which service delivery home-based health and nutrition practices. arrangements are likely to succeed. Until the Each of these activities is different, yet they 19th century, the education systems of are often provided by the same arrangement, Britain and France were private and the such as a central government public health church was the dominant provider. The gov- system. They should be differentiated. ernment had an incentive to develop anl Information about hand washing, exclu- oversight mechanism to ensure that the sive breastfeeding, and nutrition can be schools taught more than just religion. That delivered (and even financed) by NGOs proved valuable when education was nation- and other groups, delivery that works alized in these countries: the systems contin- best when reinforced by the community. ued to run with strong regulatory oversight. Outreach services, such as immuniza- Water providers in the Netherlands . c started as private companies, making the tons, can be contracted out but should concept of water as an economic good, and be publicly financed. charging for it, acceptable. When the system least able to monitor, but the case in was shifted to municipal ownership, pricing which government failures might swamp remained. Even if the Dutch never introduce private participation in water, they have market failures. Where the politics are achieved the separation between policymaker extremely pro-rich, even public financ- and provider. In sum, a country's history can ing of these services (with private provi- generate the incentives for certain institu- sion) can be counterproductive for poor tions to develop-and those institutions can people. The non-poor can capture this make the difference in whether a particular financing, leaving no curative services service arrangement succeeds or fails. for the poor-and no room in the bud- get for public health services. Strength- Sectoral service reforms ening client power, through either What do these conclusions tell us about the demand-side subsidies or co-payments, reform agenda in individual sectors? In edu- can improve matters for poor people, cation there is a tradeoff between the need even if there is asymmetric information for greater central authority to capture soci- between client and provider. etywide benefits, such as social cohesion, In the infrastructure sectors-such as and the need for greater local influence water, sanitation, transport, and energy- because student learning is difficult to the rationale for government intervention is monitor at the central level. The tradeoff is different from that in education and health, sharper when the concern is the quality of and so should be the policy responses. The education rather than the quantity. In main reason for government involvement Indonesia centralized public delivery of in water and energy provision is that those education has enrolled children in schools, services are provided through networks, so but it has been less successful in teaching direct competition is not possible. Govern- them valuable skills. To increase the quality ments also intervene to ensure access by of education, therefore, reforms should poor people to these services. So the role of concentrate on increasing the voice and government is to regulate and in some cases participation of clients-but not neglect subsidize production and distribution. the importance of central government over- There are few advantages to the govern- sight. In practical terms, this would call for ment's providing the service itself, which more community management of schools explains why the past decade has seen many and demand-side subsidies to poor people, privatizations, concessions, and the like in but with continuing stress on nationally water and energy. determined curricula and certification. Whether delivered by a private or public Governments intervene in health to con- company, the service needs to be regulated. trol communicable diseases, protect poor Who that regulator is will determine service people from impoverishing health expendi- outcomes. At the very least, when the com- Overview 17 pany is public, the regulator should be sepa- reform. The surveys have been been repli- rate from the provider (when the policy- cated in 24 Indian states. The public expen- maker and provider are indistin-guislhable, diture tracking survey in Uganda is another making this separation is all the more diffi- example, as is the Probe report on India's cult). The situation is worse when water or education system. energy is subsidized, because the sizable rents Beyond surveys, the widespread and sys- from this subsidy-the benefits of below- tematic evaluation of service delivery can market-rate services-can be captured by have a profound effect on progress toward politicians, who use them to curry favor with the Millennium Development Goals. Evalua- their rich clients rather than the poor. tions based on random assignments, such as Sanitation is different because individu- Mexico's Progresa, or other rigorous evalua- als can offload their refuse onto their neigh- tions, give confidence to policymakers and bors. So subsidies to individual households the public that what they are seeing is real. will not solve the collective action problem. Governments are constantly trying new Instead, using community-level subsidies, approaches to service delivery. Some of them and giving communities the authority to work. But unless there is some systematic allocate them, puts the locus of authority evaluation of these programs, there is no cer- where the external effects of individual tainty that they worked because of the pro- behavior can be contained. gram or for other reasons. Based on the sys- tematic evaluations of Progresa, the Scaling up government has scaled up the program to How can all these reforms be scaled up so encompass 20 percent of the Mexican people. that developing countries will have a chance The benefits of systematic program eval- of meeting the Millennium Development uation go beyond the program and the Goals? First, as noted at the beginning of country. These evaluations tell policymak- this Report, additional resources-external ers in other countries what works and what and internal-will be needed to capitalize doesn't. They are global public goods- on these reforms. Second, these reforms which might explain why they are so must be embedded in a public sector scarce.36 If these evaluations are global pub- responsible for ensuring poor people's lic goods, the international community access to basic services. This means that the should finance them. One possibility would sectoral reforms must be linked to ongoing be to protect the 1.5 percent of World Bank (or nascent) public sector reforms in such loans that is supposed to be used for evalua- areas as budget management, decentraliza- tion (but rarely is), so that this sum-about tion, and public administration reform. It $300 million a year-could be used to also means that a well-functioning public administer rigorous evaluations of projects sector is a crucial underpinning of service and disseminate the results worldwide. delivery reform. In the same vein, there In addition to creating and disseminat- should be reform in donor practices-such ing information, other reforms to improve as harmonizing procedures and making service delivery will require careful consid- more use of budget support-to strengthen eration of the particular setting. There is no recipient countries' efforts to improve ser- silver bullet to improve service delivery. It vice outputs. may be known how to educate a child or Third, a recurring theme in this Report stop an infant from dying. But institutions is what information can do-as a stimulant are needed that will educate a generation of for public action, as a catalyst for change, children or reduce infant mortality by two- and as an input for making other reforms thirds. These do not crop up overnight. Nor work. Even in the most resistant societies, will a single institutional arrangement gen- the creation and dissemination of informa- erate the desired results. Everything from tion can be accelerated. Surveys of the qual- publicly financed central government pro- ity of service delivery conducted by the vision to user-financed community provi- Public Affairs Centre in Bangalore, India sion can work (or fail to work) in different have increased public demand for service circumstances. 18 WORLD DEVELOPMENT REPORT 2004 vice delivery teaches us the importance of B O X 3 Why is this WDR differentfrom all other WDRs? implementation. Singapore and Nigeria This World Development Report builds on also complements the 1994 WDR on infra- (both former British colonies) have simi- previous WDRs while venturing into some structure-which presciently focused on larly designed education systems. But in new areas. Starting from the decennial private-sector participation in the sector- implementation, the outcomes, especially WDRs on poverty, it applies the 2000 WDR's by addressing the politics of infrastructure for poor people, could not be more differ- theme of empowerment to the 1990 WDR's provision for poor people.lt expands on ent. Governments and donors often over- emphasis on health and education as one aspect of the 1997 WDR on the role of important ways of escaping poverty. In so the state-the state's responsibility for look implementation difficulties when doing, it extends the 1993 WDR on basic services. And it builds on the two designing policies. There may be benefits to health-which prescribed a technical solu- most recent WDRs-on institutions for having the central government administer tion (cost-effectiveness analysis) to improv- markets and institutions for collective schools (such as social cohesion). But the ing outcomes for poor people-by examin- action-by identifying the incentives for ing the institutional factors that may give reforming institutions to make services problems with central provision of a hard- rise to the correct technical solutions. It work for poor people. to-monitor activity such as primary educa- ______ _____ _____ ______ _____ _____ _____ ______ _____ _____ ____ tion are so great, especially am ong hetero- geneous populations, that the government should rethink its position of centrally con- Rather than prescribe policies or design trolled schools. Selectivity is not just about the optimal institution, this Report choosing from the available design describes the incentives that will give rise to options-it is about choosing with an eye the appropriate institution in a given con- toward options that can be implemented. text (box 3). Decentralization may not be That there is no silver bullet, that we the optimal institutional design. But it may should be looking for incentives that give give local governments the incentives to rise to appropriate institutions, that we build regulatory capacity that, in turn, need to be more realistic about implemen- could make water and energy services work tation in choosing among options-all better for poor people. NGO service provi- imply that these reforms will take time. sion might be effective in the medium run, Even if we know what is to be done, it may as it has been in education in Bangladesh. be difficult to get it done. Despite the urgent But the incentives it creates for the public needs of the world's poor people, and the sector to stay out of education make it many ways services have failed them, quick much harder to scale up or improve qual- results will be hard to come by. Many of the ity-as Bangladesh is discovering today. changes involve fundamental shifts in Many of these institutions cut across the power-something that cannot happen public sector-budgetary institutions, overnight. Making services work for poor intergovernmental relations, the civil ser- people requires patience. But that does not vice-which reinforces the notion that ser- mean we should be complacent. Hubert vice delivery reform should be embedded in Lyautey, the French marshal, once asked his the context of public sector reform. gardener how long a tree would take to In addition to looking for incentives to reach maturity. When the gardener an- generate the appropriate institutions, gov- swered that it would take 100 years, Marshal ernments should be more selective in what Lyautey replied, "In that case, plant it this they choose to do. The experience with ser- afternoon." Cambodia Contracts to improve health services-quickly Camtibodia begani experimeniting with differenttforms of contracting to improve health services in 1998. The lesson- thanks to good evaluation-is that contracting can lhelp increase the coverage of some key services in a short time. i aorethan25yearsofconflictleft facility surveys were conducted in 1997 significantly higher salaries to providers than It/I Cambodia with little health before the experiment. No district had in the other types of districts. 1.. V . iinfrastructure. In the late 1990s more than 20 percent of its planned health The pattern of increases is similar across its health indicato's were among the worst facilities functioning. All had very poor a variety of service and coverage indicators health service coverage. And all were com- (figure 2). The contracted-out districts in Southeast Asia. Average life expectancy at birth was less than 55 years. Infant miiortal- parable in their socioeconomic status. often outperformed contracted-in districts, itv was 95 per l 000 live births. And imater- Annual per capita recurrent spending by which outperformed control districts. But nal mortalit was 437 per100,000live donors and government was higher in the not all indicators were as responsive. The nil .rthithe pbi hea care syse coiitracted districts: $2.80 in the con- share of deliveries assisted changed by only births.'iTned pubientalth aerae fatyustei tracted-in district, $4.50 in contracted-ouit a small amount in all three districts. And remained rudicentary: average facility use districts, coimipared with $2.90 in control there was no difference between con- patients complained of erylow qpralitye districts.? These differences are large and tracted-in and contracted-out districts in represent less than 20 percent of the health the increase in vitamin A coverage. The tracted with nongovernmental entities to expenditures (inicluding private anid level of immunization in contracted dis- provide health services in several districts, excluding capital investments from the gov- tricts also remained quite modest, peaking Thvie ontractinghinea sedaccsce s tosevel hlths erlmient) in all of the districts. at only 40 percent. The contracting Increased access to health Oto-oktepniue nhat * t . ~~~~~~~~~~~~~~~~~~~Out-of-pocket expenditures on health services-and not at the expenise of equity. Contracting for better results care services fell dramatically in the con- Contracting primary health care All districts imiproved service coverage in a tracted-out districts but increased slightly services(in and out) short time. After only 2.5 years of the four- in contracted-in and control districts. The year experiment, all districts had achieved reduction was especially marked among the ntrventandonlyselectedroral aitri,eascnsiste o their contractual obligations for most of the poor ($35 a year, or 70 percent), indicating randomly seeteualdstievaluation indicators.4 The use of health ser- better targeting and more efficient transfers 100,000 to 200,000 people.2 Contractors vices among the poorest half of the populace of subsidies. were chosen through a competitive process iIncreased by nearly 30 percentage points in Even though the health ministry based on the quality of their technical pro- the contracted-out district (figure 1). One cncouraged all districts to implement offi- posal and their price. Three approaches possible explaniation is that the contracted- cial user fees, only one contracted-in dis- were used. out districts did not charge official user fees; trict established a formal user fee system * Contracting out. Contractors had full they also discouraged health care workers and used the receipts from the system to responsibility for the delivery of speci- from taking "'ulnofficial" user fees by paying reward health care workers with monthly fied services in the district, directly performance and punctuality bonuses. employed their staff, and had full man- Figure 1 Use of district health services That could accotint for slightly higher agement control (two districts), increased most in districts that contracted out spending for this type of district.5 Percentage of illnesses treated at a health There are several possible reasons for * Contractinig in. Contractors provided facility for people in the poorest half of the Thes e p sever o ossibte contr only management support to civil ser- populace these pro-poor outcomes in the contracted vice health staff, and recurrent operat- Percent ing costs were provided by the govern- 35 * The regular availabilitv of drugs and qual- ment through normal government 30 ified staff strengthelced service provision channels (three districts). 25 at health centers in the villages, where * Corttrol areas. The usual government 20 2001 most poor people are concenitrated. provision was retained (four districts). 15 1997 \ The contracted nonigovernmiiental orga- A budget supplement was provided to con- 10 \ nizations (NGOs) used a market-based tracted-in and control districts to make 5 wage and benefits package to attract and recurrent expenditures roughly equal in 0 MA retain health care providers. these districts. Control Contracted Contracted * A reduction in the private out-of-pocket Performance indicators were measured in out cost of services and a more predictable for all the districts by household, and health Source: Bhushan, Keller, and Schwartz (2002). and transparent fee structure increased 20 Spotlight on Camiibodia Figure 2 Coverage for many services increased in In the contracted-in districts, the NG;Os contracted areas with variation across services supplemented provider salaries with their Percentage point increase in coverage of selected health funds and, in one district, alocated a indicators between 1997 and 2001 in control and contracted fu districts of Cambodia larger share of user-fee income. The con- Percent trol districts, left to their own devices, al- 60 Control lowed workers to pursue private income- Control maximizing behavior through unofficial 50 Contracted in fees and private practice, to the detriment lContracted out fe ldpiaepatc,t h ermn of the public health care services for the 30 poorest of the poor. 20 * * L l l l ] ' Transparent and predictable fee struc- tures are important in improving access to 10 health services. Official user charges were 0 introduced in only one contracted-in dis- 10 I trict, in consultation with communiities, to Antenatal Tetanus Assisted Full Vitamin A provide incentives to health workers. To care toxoid deliveries immunization coverage remove ambiguity about charges, a sched- Source: Bhushan (2003). ule of user fees was prominently displayed in all health facilities. This discouraged private practice and helped bring "under- the-table" payments formally into the sys- the demand for health care services by health indicators were agreed on-for poor tem. Out-of-pocket spending on health fell the poor. people--progress toward achieving them in that district. No user fees were intro- * The availability of hcalth services in vil- was measured through independent house- duced in the other two contracted-iii dis- lages reduced travel expenditures to hold sLrveys and spot checks by government tricts, or in the control disrricts, where seek health care, and NGOs enforced staff Payments were linked to achieving tar- ouit-of-pocket spending did not come rules against informal payments by gets, with bonuses for better-than-agreed-on down. patients. performance. (Contracting health services to NGOs lmproving health services for the poor can be difficult for policymakers to accept. requires that health workers be adeqtiately But the Camiibodian experience shows that Agreements on deliverables- compensated and effectively supervised and it can be effective and equitable. It helped and enforceable contracts supported. The NGOs working in con- convince policymakers that the model Contracting health services to NGOs can tracted-out districts revised the salaries of could bc adopted on a larger scale. They are expand the coverage for poor people. In health care providers, bringing them in line extending contracting to II poor and Cambodia it took agreements on deliver- with average salaries in the private sector. In remote districts, where the public provision ables and an enforceable contract, which in return, the NGOs required the provider, to of services is dismal. turn required an independent performance work full time in health facilities and to verification system. Once targets for 13 key have no private practice. Educo Educaci6n con Participaci6n de la Comunidad en El Salvador By contracting directly with communities, El Salvador dranmatically increased the primary school enrollment of children in poor and remote areas-without reducing the quality of learning. E I Salvador was wracked by civil war cmmunity schools, bearing the cost them- in 1992 to 555,000 in 1995-with over 75 throughout the 1980s. Some 80,000 selves and paying teacherswhen they could. percent of the new students enrolled in people died-in a total population of The government seized on this model of Educo schools (figure 1). By 2001 there roughly 5 million-and many more were community-based schooling as the basis for were almost 260,000 students enrolled in wounded and disabled. Income per capita fell a formal program that would be financially Educo primary schools, 41 percent of all almost 40 percent between 1978 and 1983.1 and administratively suLpported by the mnin- students enrolled in rural schools-and In 1989 the conservative Republican Alliance istry: Educaci6n con Participaci6n de la more than 100,000 children enrolled in Party won a majority in the national assem- Comunidad, or Educo, with the goal of en- Educo preschools, 57 percent of all children bly, with Alfredo Cristiani as president. couraging the establishment of preschools in preschool. Despite contentious negotiations, a peace and primary schools, or classrooms in Even as enrollments increased rapidly, accord was signed in January 1992.2 existing schools. there is little evidence that learning quality The war had severely damaged the edu- Begun in 1991, Educo targeted 78 of suffered. A survey of 30 Educo primary cational system. Communication between the country's poorest rural municipalities schools and 101 traditional schools in 1996 the central ministry and schools broke (of 221 urban and rural municipalities), found no significant differences in average down, supervision collapsed, and many By 1993 the program was expanding to all math and language test scores among third teachers, viewed by some as government rural areas, including many areas formerly graders in the two types of schools.8 A fol- "agents" and by others as agents of social under opposition control. But not all of low-up study in 1998 found that grade pro- opposition, abandoned their posts. By 1988 the "popular schools" established during motion anid repetition were similar across the more than a third of the country's primary the war were incorporated into Educo. two types of schools as well.9 As the innova- schools had closed.3 And by the end of the Some observers claimed there was selective tion matured, the institutional arrangements war some 1 million children were not in inclusion based on political favoritism; that it introduced took hold and ensured school.4 others saw not incorporating popular rapid expansion of school places and enroll- schools into a government program as a ment of poor children, seemingly without a way of sustaining spontaneous commu- substantial cost in quality.'0 Establishing Educo-Education nity-based education.6 with the Participation Each Educo school (or section withini a Parent visits to classrooms made much of Communities traditional school) is operated by a Com- of the difference The Ministry of Education quickly identi- munity Education Association (ACE)-an That Educo schools served the poorest of El fied expanding access to basic education elected committee made up primarily of Salvador's students, in the poorest areas, and raising its quality as central goals- students' parents-that enters into a one- makes these results all the more astonish- both to rebuild national unity and to pro- year renewable agreement with thle min- ing. How did they do it? Using retrospective mote long-term economic development. istry. The agreement outlines rights, Minister of Education Cecilia Gallardo de responsibilities, and finiancial tr-ansfers. The Figure 1 Enrollment in Educo schools has been Cano, a reform proponent from the "mod- Ministry of Education oversees basic policy increasing rapidly ernizing" wing of the Republican Alliance and technical design. Using the monev Number of students enrolled in traditional rural and in Educo primar classes Party, was intent on lessening the distrust directly transferred to them, ACEs select, a Ec between former combatants. hire, monitor, and retain or dismiss teach- Thousands of students But skepticism was high. The Ministry ers. Teachers at Educo schools are hired on 700 600 of Education was distrusted in many parts one-year renewable contracts. Parents are ' of the country and by organized groups taught about school management and how 500 Traditionall such as the National Association of Teach- to assist their children at home.7 400 ers. Expansion of the traditional education 300 system was viewed suspiciously as a covert Three-quarters of new enrollments 200 means of reasserting national control and Educo succeeded in many respects. From a 100 building political support in opposition- pilot phase of six ACEs in three depart- 0 Educo i-t-- AJ dominated areas.5 ments, it scaled up nationally to all of the 1990 1992 1994 1996 1998 2000 2002 During the war many communities had country's departments by 1993. Rural pri- Note:Figuresfor2tO2 are estimates. recruited local teachers and established mary enrollments increased from 476,000 Source:ElSalvadorMinistryofEducation. 22 Spotlight o01 Educo data that allow controls for child, house- earnings, which suggests that parent associ- university places fueled by opening higher hold, teacher, and school characteristics- ations used compensation to motivate education to the private sector). and statistically adjusting for the fact that greater effort among teachers.'3 Offering or These factors suggest that the Educo unobserved abilities of children might sys- withholding future employment itself was model might not be directly replicable in a tematically differ betwveen the two types of an incentive, and one that ACEs used. different setting. But some lessons are gen- schools-researchers found that commu- Turnover among Educo teachers was high, eral. First, with political will it is possible to nity involvement explains much of Educo's which suggests that job loss was not an idle change the relationships between the actors success. threat. in basic education. Second, schools can be Parents are more active in Educo transformed to work in ways that promote schools (figure 2). And their involvement Converging with traditional enrollment, participation, and learning- affects learning. Each classroom visit by schools even for children from the poorest house- parents was associated with significantly Educo's administration has become embed holds. Third, getting parents to participate higher math and language test scores ded in the Ministry of Education, and effectively in managing schools can help regardless of the type of school. Parents Educo has developed into a major schooling overcome some of the potential pitfalls in were more active informally as well: they model in the country. Aspects of traditional the provision of education services-espe- were more likely to meet with teachers or to and Educo schools have been converging cially monitoring schooling in remote assist teachers in monitoring attendance or Traditional schools now have more ge areas. Fourth, it is possible to scale up small maintaining schoolfurniture. participation in school governance and innovations to have a significant impact on How did Educo and parent involvement management and are more autonomous national outcomes. affect test scores? At least part of the story is with supporting block financing. Similarly, that teachers were less likely to be absent in the pay packages of teachers are more simi- Educo schools (averaging 1.2 days of lar: Educo teachers receive the same salaries absence a month rather than 1.4 days). Stu- and benefits as teachers in traditional dents in Educo schools were also absenlt less schools. Even so, a key distinction remains: (three fewer days a month) than students in Educo teachers are hired (and potentiallv traditional schools.i2 In addition, Educo's fired) by parent committees while those in more flexible compensation scheme traditional schools are not. resulted in greater variability in teacher Is the Educo model applicable Figure 2 Educo promoted parent involvement elsewhere? Parent visits to Increase in test scores Educo's achievements might appear idio- classrooms in previous associated with a visit syncratic. The end of a bloody civil war that month had thrown the traditional education sys- Number of visits Percent increase 6 565 6 . 5.71 tem into chaos opened up a unique oppor- tunity to change the way schools were man- 5 5 ~~ - aged. Based in part on coping strategies 4 4 378 during the civil war, El Salvador had a his- cia - l; 3 - . - . l tory of community involvement in school management. Indeed, the community asso- ciations appear to have worked better in l ~ - ~- ~= - cl 1 places that had experience in community 0 Tradit'onalEducoorganization.'4 In addition, in the after- Traditional Educo Math Language math of the war there was an unusually schools schools large pool of educated people without jobs Source:Adapted from Jimenez and Sawada 119991. (coinciding with the rapid expansion of ~'l Johannesburg Accountability in city services In 1999 the Transformation Lekgotla, the political body directed to address thefinancial and institutional crisis of Johan- nesburg, South Africa, appointed a new city management team. The team's task was clear: not to fix street lights but to fix the institutions thatfix street lights.' The solution was a three-year plan-iGoli 2002-to reconfigure city services.2 B y most developing world standards They slashed capital and operating budgets, independent, single-purpose entities to Johannesburg is not a poor city. But and even expenditures needed to maintain overhaul larger municipal services. it faces serious development and minimum service levels. They froze posts, These operating entities were the major service delivery challenges. Apartheid made causing huge increases in workloads as innovation of iGoli 2002. sure that exclusive white suburbs were well despairing officials began to drift away. And Three utilities were established for user- serviced, forcing black residents into they began to explore public-private part- charge-based services-water and sani- sprawling underdeveloped slums. Poverty, nerships. tation, electricity, and waste manage- unemployment, and homelessness are all worsened by the deeper problem of ment. inequality. The city of gold-iGoli 2002 Two agencies were established-for The Johannesburg Metropolitan Munici- The new city management team realized parks and cemeteries, and for roads and pality was democratically elected in 1995 to that Johannesburg needed a new system of storm water-where expenditure would address the service imbalances. It quickly accountability for service delivery within a still have to be covered by tax revenue. found itself in a fiscal and institutional crisis. dramatically different institutionial archi- Smaller corporatized units were set up Johannesburg was not one institution tecture. To address fragmentation and the for facilities like the zoo and the civic but five, with an overarching Metropolitan severe moral hazard, the city had to be theater. Council and four primary-level councils. reunified. Political debate focused on two Each could decide its priorities and approve models of metropolitan coordination: All were established as new companies, with the council as sole shareholder. its budget. But responsibilities for key ser- Defining more clearly the rules of bud- Two key units would guide and oversee vices were split between the two levels, and geting, fiscal transfers, and service deliv- the new entities: a corporate planning the operating budgets of the councils had to ery between the metropolitan and unit to do citywide strategic planning, and balance only in aggregate. That meant each municipal tiers, strengthening both a contract management unit to regulate assumption that its shortfalls would be off Creating a one-tier metropolitan gov- the operating utilities through a range asetbysurpluses is anorthr w ernment. 3 of new instruments, including licensing set by surpluses in another. ~~~~~agreements and annual service level The arrangement was a recipe for disas- Johannesburg chose a hybrid. It central- agreementsa ter. Each municipality went on a spending ized political authority, treasury manage- spree, and ambitious infrastructure plans ment, and spatial planning under one met- One size does notfit all were rolled out without the necessary ropolitan government. But it organized Since the operating entities are not bound finance. Deteriorating revenues-due to a service delivery through decentralized by overarching administrative rules, they service-payment boycott culture left over structures. This meanit merging five sepa- have scope to differentiate. Each could set from anti-apartheid struggles, poverty, and rate councils into one overarching munici- up different management structures, poor credit control-made the situation pality, creating integrated service delivery reporting lines, delegations, job descrip- worse. The city was forced to delve into its structures with new incentives. tions, performance management systems, reserves, but these could go only so far, and by late 1997 major creditors could no longer Accountability in service delivery and opeirating procedures. Each could con- be paid. At the peak of the crisis, the city had Under one metropolitan council, iGoli 2002 figue iseiner a lity to suit a specific service delivery environment. an operating deficit of R314 million. split the institution for policy formulation Three examples. Johannesburg was in serious trouble. and regulation from the institutions for Having decentralized responsibilities, the implementation. On one side, a core admin- The water and sanitation departments national government followed the intergov- istration remained responsible for strategic were merged into one department and ernmental rules and would not bail the city planning, contract administration, and such under the Company's Law converted out. So Johannesburg had to dig itself out corporate services as finance, planning, and into a city-owned utility with a board of of its own crisis. communication. On the other, two sets of directors. The assets and workers of the Two years of harsh cutbacks followed. operating entities were established: 11 new departmients were transferred to the Blaming officials for the crisis, politicians regional administrations for libraries, utility, which was put under a five-year took a much tighter rein over day-to-day health, recreation, and other community management contract with a private decisions, ending management discretion. services; and financially ring-fenced, semi- company. 24 Spotliglht o07 Johannesburg e The roads department wvas cooiverted into Figur 1 Getting back tc an operating markets xvent frotii R300 million to well a citv- owned agency with a professional surplus-thanks to iGoli 2002 over a billion in two years. board and dividcd into two depart Operating surplus/deficit ments- -for planning and for contracts. Miliaons of rands Risks and prospects The contracts department operated \\Risk an prsu mata against specific outcomes set by the plan- 10; bWil policym ak ing poideprsandr- ningdepartment,wviththethreatthatfail- 50 beLswvcji pciticymakuig, provieers, and reg- ningdto epetbenthtarkscouldeatdhtfaio- 50 pulators? '[he roles of client ard contractor n0 t m b c are stilt evolving. Some implementatioii contracting tasks out to the private sector. -1 0* capacity remains within the core adminis- 'I'he gas company was sold to the private * tration. /As in the past, managers occasion sector. -203 ally get hauled into councilors' offices to Teer gv0i esNplain ibeir actions. There are also unre The reforms gave operating entities -300 iw solved e overnance debates, with the council management independence. For example, -350 salaries have been adjusted to attract top- 1995 1996 1997 1998 1999 2000 2001 2002 arguing for a greater councilor representa flight skills, and new systems have been pro- Srrte: Allan, Gotz, and Jnseph (2ti1) tion on the boards of oberatinca entities. cured for everything from human resource the comritment to the principles of inh management to remote water-pressure t02om meLering-increasing productivity and ser- 0 vice efficiency. And they have introduced inF, witlh dramatit imiiprovem-nent in both Keeping the monitoring and regulatory innovative staff development programs and operating and capital budgets (iigure 1). ulliis of the opei ating entities within the performance-linked pay schemes. city administration; they are not legally The entities operate at arm's length from Engaging other stakeholders and administratively independent. the council, but accountability has been Labor: Despite protracted negotiaiioiis 'viaintaining the contract management strengthened because the primary mecha- wilh organi/ed labor, i(Goli 21002 did not get unit's operational autonomv and capac- nism is no longer the impossible-to-digest its endorsement According to labor groups, ity---and thus the indeperdence of the committee report on everyday operational the city's crisis was not a result of a failure operating entities. matters. Now councillors focus on strategiL of' institutionial designi. iistead it v:as a oversight, and officials are responsible for resultof"alackofskillsand experience, Ald aBendsmarkitng service delvery stan- outcomes clearly defined in service-level maniagemenit's unwvillingness to [establishl dards, monitoring these over time, and agreements. Reporting goes through struc ftimctional organizations and . . . financially making the information available. tured channels, either to the contract man- unsound decisions."4 Ensuring ihat fisLal and financial deccn- agement unit or to company boards of tralization remains binding. Municipali- directors, which include external specialists National Government,: [he team ties relying primarily on their own rev- capable of probing service results. nc,otiated a R500 mlillioi restructuring enue sources to fulfill their democratic The operating entities have also set up gr,-oit with the National Treasury to support duties without national guarantees are user forums allowing communities to com- iGoli 2002 in exchangc for a commitment more likely to be accountable to their municate needs, raise complaints, and even to timely and steadfast implementationi oi' ciii.'ns. '[he current intergovernmental participate actively in service provision. its key elemeiits. It is a key accountability system has devolved authority and Officials are much more sensitive to ever mechanism between the national and cii3 accountability to the cities; this needs to changing service delivery challenges. governments and has became an incentixv remain. These management improvements are scheme to catalyze citywide restructuiing o Both counLinors and officia's consisteintly already translating into better service deliv- throughout the counii y. adhering to a clear, courageous, and far- ery. WVaste collection has been extended to sightecl strategy. Sustaining momentum poorer neighborhoods for the first time. Capital Markets: On the strength of the will rcquire greater citizen voice at all lev- Fleets of new buses now serve outlying reforms, management sought a newv kredii els. '5 he decentralized operating cntities communities. In addition, expenditure on raling, aiming to win hack the confidence and the administrative regions have water infrastructure has increased and of the city's banking community. As th city)' mechanisms for engaging citizens. Using water services have expanded. Results are slifted from a laigc deficit to a balanced them will be critical for sustaining iGoli also apparent in the city's financial stand- budLget, capital ecxpenditure financed by the 2002. Overview 25 Endnotes Overview 33. Schick (1998). 34. Realizing that the central education system has led to underrepresen- 1.f tracking teaheng world a hole hidesrthe foact ttu a nAi s tation of students, from low-income families, one of the prestigious French off track in reaching the income poverty goal. 2. Walker, Schwarlander, and Bryce (2002). grandes ecoles, L'institut des etudes politiques ("Sciences Po") has begun to 3* Devarajan, Miller, and Swanson (2002). use separate admissions criteria for students from poor neighborhoods. 3. Devarajan, Miller, and Swvanson (2002).35Lenr(20) 4. Peters and others (2002), p. 218. 3.Loir 20) 4. Petersa and othersso (2002),p.28. 36. Another reason is that most project managers are not interested in 5. Reinikka and Svensson (2001). 6. ChaLudhury anld Hammiiiier (2003). investing in knowledge that might show their program to have been a failure. 7. Jaffre, Olivier, and de Sardan (2002). 8. Probe Team in association with Centre for Development Economics, spotlight on Cambodia 1998; Rosskam (2003),. 1. Cambodia National Institute of Statistics and ORC Macro (2001) 9. Analysis of Demographic and Health Survey Data (see table 1. I of the 2. The operations research was funded by the Asian Development Bank. Report). See also U.K. Department for International Development (2002). 3. Bhushan, Keller, and Schwartz (2002) 10. Bhushan, Keller, and Schwartz (2002). 4. There is only one instance, that of vitamin A coverage, in which one I1. Galiani, Gertler, and Schargrodsky (2002). district had not increased coverage at the time of the midterm evaluation. 12. Ahmad (1999). 5. Soeters and Griffiths (2003). 13. ATlorld Bank (2002b). 14. Behrman and Hoddinott (2001), Gertler and Boyce (2002). spotlight on Educo 15. Glaeser and Shleifer (2002). 16. Diaz-Cayeros and Magaloni (2002). 1. World Bank (2002c). 17. World Health Or anization 2003 ~2. Eriksson, Kreimer, and Arnold (2000). 17. World Health Organization (2003) 3. This assessment of the position of teachers and school closings is 18. Besley and Burgess (forthcoming). 19. When asked why he did not complain, one villager replied, "I could from Reimers (1997). ,with an accident on the road. I could be put in the brick kiln oven. My 4d Action learning program on participatory processes for Poverty meet ~~~~~~~~~~~~~~~~~Reduction Strategy Papers (2003). bones could be broken.'' 5. Action learning program on participatory processes for Poverty 20. International Labour Office (2002).. Reduction Strategy Papers (2003). 21. Scott (1998). 6. For an example of this critique see Davies (2000) and a discussion in 22. Reinikka and Svensson (2002). Reiimers (1997). 23. Chomitz and others (1998). 7. Initial studies suggested that fewv of these "Parent School" programs 24. Computerization of land registration in Karnataka, India, reduced took hold. But thev were made an official program-with financial sup- the transaction time to 30 minutes and eliminated the payment of bribes, which hd rise to 25to 50 imes te regitratioi fee.port-in the past five years, and they appear to have expanded since then. which had risen to 25 to50i trgrine 8. Jimenez and Sawada (I 999). 25. Koeng, Foo, and Joshi (2000). 9. Jimenez and Sawada (2002). 26. simeneh and Urawadla (19). 10. Indeed, one early assessment based on a survey of 140 schools in 1993 27. Garie and Urauiola (2003). found little difference between different types of schools (Reimers 1997). 28. Angrist and Vawdas (2003). 11. El Salvador Evaluation Team (1997). 29 Anrs anltes 20) 2. Jimenez and Sawada (1999). 30. Interview by John Briscoe. 31. World Bank (1998) and Vorld Bank (2002a). 13. Sawada (1999). 32. Even a recommendation to apply interventions that pass a social 14. Reimers (1997). benefit-cost analysis test will not be eniough. Social benefit-cost analysis is concerned with valuing an intervention's outputs and inputs at the right set spotlight on Johannesburg of shadow prices (Bell and Devarajan 1987, Dreze and Sterni 1987). Yet the 1. Allan, Gotz, and Joseph (2001). problem is that the inputs often do not translate to the desired output 2. iGolinmeans"cityofgold." because of weak incentives. The same point applies to recommendations of 3. Ahniad (1996). using "cost-effective" interventions in health (World Bank 1993). 4. Allan, Gotz, and Joseph (2001). References The word "processed" describes informally reproduced works that may not be commonly available through libraries. Ahmad, Junaid. 1996. "The Structure of Urban Governance in South Behrman, Jere R., and John Hoddinott. 2001. "An Evaluation of the Impact African Cities."International Tax and PublicFirnance 3(2):193-213. of PROGRESA on Preschool Child Height." IFPRI FCND Discussion - 1999. "Decentralizing Borrowing Powers."World Bank PREM Note Paper 104. Washinigton, D.C. Available on line at www.ifpri.org. 15. Washington, D.C. Bell, Clive, and Shantayanan Devarajan. 1987. "Intertemporally Consistent Allan, Kevin, Graeme Gotz, and Carmel Joseph. 2001. Johantnesburg: An Shadow Prices in an Open Economy: Estimates for Cyprus." Journal of African City in Chatige. Johannesburg: Zebra Press. Public Econottics 32(3):263-85. Angrist, Joshua, Eric Bettinger, Erik Bloom, Elizabeth M. King, and Michael Besley, Timothy, and Robin Burgess. Forthcoming. "The Political Economy Kremer. 2002. "Vouchers for Private Schooling in Colombia: Evidence of Government Responsiveness: Theory and Evidence from India." from a Randomized Natural Experiment." American Economic Reviewv Quarterly Journal ofEconomtics. 92(5):1535-58. 26 WORLD DEVELOPMENT REPORT 2004 Bhushan, Indu, Sheryl Keller, and Brad Schwartz. 2002. "Achieving the Twin Jaffre, Y, J.-P. Olivier, and Olivier de Sardan, eds. 2002. Les Dysfonction- Objectives of Efficiency and Equity: Contracting for Health Services in nements des SysrWmes de Soins. Rapport du Volet Socio-anthropologique. Cambodia." Asian Development Bank, Policy Brief Series 6. Manila. Enquetes sur l'Acces aux Soins dans 5 Capitales d'Afrique de l'Ouest. Mar- Bruns, Barbara, Alain Mingat and Ramahatra Rakatomalala. 2003. "A seille: Unicef-Cooperation Francaise. Chance for Every Child: Achieving Universal Primary Education by Jimenez, Emmanuel, and Yasuyuki Sawada. 1999. "Do Community-Man- 2015." World Bank. Washington, D.C. aged Schools Work? An Evaluation of El Salvador's Educo Program." Bhushan, Indu. 2003. "Contracting Health Services in Cambodia." Back- World Bank Economic Review 13(3):415-41. ground note for the WDR 2004. .2002. "Does Community Management Help Keep Kids in Schools? Cambodia National Institute of Statistics, Ministry of Health, and ORC Evidence Using Panel Data from El Salvador's EDUCO Program." World Macro. 2001. "Cambodia Demographic and Health Survey 2000." Bank Economic Review 13(3): 415-441. Phnom Penh, Cambodia. Koenig, Michael A., Gillian H. C. Foo, and Ketan Joshi. 2000. "Quality of Chaudhury, Nazmul, and Jeffrey Hammer. 2003. "Ghost Doctors: Absenteeism Care within the Indian Family Welfare Programme: A Review of Recent in Bangladeshi Health Facilities.' Background paper for the WDR 2004. Evidence.' Studies in Family Planning 31(l):1-18. Chomitz, Kenneth, Gunawan Setiadi, Azrul Azwar, Nusye Ismael, and Widi- Leonard, Kenneth L. 2002. "When Both States and Markets Fail: Asymmet- yarti. 1998. "What Do Doctors Want? Developing Incentives for Doctors ric Information and the Role of NGOs in African Health Care." Interna- to Serve in Indonesia's Rural and Remote Areas." World Bank Policy tional Review of Law and Economics 22:61-80. Research Working Paper 1888. Washington, D.C. Peters, David H., Abdo S. Yazbeck, Adam Wagstaff, G. N. V. Ramana, Lant H. Davies, Ian. 2000. "The Debate over Popular Education in El Salvador." Pritchett, and Rashmi R. Sharma. 2002. Better Health Systems for India's Edgewood College, Madison, Wisconsin: LASA, Miami, Florida, March. Poor: Findings, Analysis, and Options. Washington, D.C.: World Bank. Available on line at http://forlang.edgewood.edu/ian/Education.htm. Reimers, Fernando. 1997. "The Role of the Community in Expanding Edu- Processed. cational Opportunities: The EDUCO Schools in El Salvador." In James Devarajan, Shantayanan. 2002."Growth Is Not Enough"World Bank. Processed. Lynch, Celia Modgil, and Sohan Modgil, eds., Education and Develop- Devarajan, Shantayanan, Margaret J. Miller, and Eric L. Swanson. 2002. ment: Tradition and Innovation: Volume Two-Equity and Excellence in "Goals for Development: History, Prospects, and Costs." World Bank Education for Development. London: Cassel. Policy Research Working Paper 2819. Washington, D.C. Reinikka, Ritva, and Jakob Svensson. 2001. "Explaining Leakage in Public Diaz-Cayeros, Alberto, and Beatriz Magaloni. 2002. "Public Services Medi- Funds." World Bank Policy Research Working Paper 2709. Washington, ated by the Political Process." Stanford University. Processed. D.C. Dreze, Jean, and Nicholas Stern. 1987. "The Theory of Cost-Benefits Analy- Rosskam, Ellen. 2003. "No Pills, No Bandages, Nothing." International sis." In A. J. Auerbach and M. Feldstein, eds., Handbook in Public Eco- Labour Office, Geneva. nomics, Vol. 2. Amsterdam, New York: North-Holland. .2002. "Working for God?" World Bank, Washington, D.C. Processed. El Salvador Evaluation Team. 1997. "El Salvador's EDUCO Program: A First Sawada, Yasuyuki. 1999. "Community Participation, Teacher Effort, and Report on Parents' Participation in School-Based Management." World Educational Outcome: The Case of El Salvador's EDUCO Program." Bank Working Paper Series on Impact Evaluation of Education Reforms William Davidson Institute Working Papers 307. University of Michi- 4. Washington, D.C. Available on line at http://www.worldbank.org/ gan. Available on line at http://eres.bus.umich.edu/web/dwpauth.html. research/projects/impact. htm. Schick, Allen. 1998. "Why Most Developing Countries Should Not Try New Eriksson, Joh, Alcira Kreimer, and Margaret Arnold. 2000. "El Salvador: Zealand's Reforms." World Bank Research Observer 13(l):23-31. Post-Conflict Reconstruction." World Bank Evaluation Department, Scott, James. 1998. Seeing Like a State: How Certain Schemes to Improve the Country Case Study Series: Washington, D.C. Human Condition Have Failed. New Haven,Conn.: Yale University Press. Estevez, Federico, Beatriz Magaloni, and Alberto Diaz-Cayeros. 2002. "The Soeters, Robert, and Fred Griffiths. 2003. "Improving Government Health Erosion of One-Party Rule: Clientelism, Portfolio Diversification, and Services through Contract Management: A Case from Cambodia." Electoral Strategy." Paper presented at the Annual Meeting of the Amer- Health Policy and Planning 18(l):74-83. ican Political Science Association Conference, Boston. U.K. Department for International Development. 2002. "Transport's Role in Galiani, Sebastian, Paul Gertler and Ernesto Schargrodsky. 2002. "Water for Achieving the Millennium Development Goals." Transport Resource Life: The Impact of the Privatization of Water Services on Child Mortal- Centre. ity." Stanford University, Cneter for Research on Economic Develop- Walker, N., B. Schwarlander, and J. Bryce. 2002. "Meeting International ment and Policy Reform, Working Paper 154, Stanford. Goals in Child Survival and HIV/AIDS.' The Lancet 360(9329):284-9. Gauri, Varun, and Ayesha Vawda. 2003. "Vouchers for Basic Education in World Bank. 1993. World DevelopmentReport 1993: Investing in Health. New Developing Countries: A Principal-Agent Perspective." Background York: Oxford University Press. paper for the WDR 2004. . 1998.AssessingAid: What Works, WhatDoesn't, and Why. New York: Gertler, Paul, and Simone Boyce. 2002. "An Experiment in Incentive-Based Oxford University Press. Welfare: The Impact of PROGRESA on Health in Mexico." University of .2002a. A Casefor Aid: Building a Consensus for Development Assis- California, Berkeley. Available online at http://faculty.haas.berkeley.edu/ tance. Washington, D.C.: World Bank. gertler/. Processed. _____. 2002b. "Romania: Local Services Delivery Study." World Bank, Glaeser, Edward, and Andrei Shleifer. 2002. "The Curley Effect'" Harvard Washington, D.C. University and National Bureau of Economic Research. Processed. . 2003. "Global Economic Prospects and the Developing Countries: Hsieh, Chang, and Miguel Urquiola. 2003. "When Schools Compete, How Investing to Unlock Global Opportunities."World Bank,Washington, D.C. Do They Compete? An Assessment of Chile's Nationwide School _ _.2002c. World Development Indicators. Washington, D.C.: World Bank. Voucher Program." Princeton University and World Bank. Processed. International Labour Office. 2002. "Social Dialogue in the Health Setvices: World Health Organization. 2003. http://www.who.int/csr/media/sarss_wha. Institutions, Capacity and Effectiveness," Sectoral Activities Program. pdf Geneva. World Development Reporl 2004: Making Services Work for Poor People Poor people suffer from a severe lack of basic services. is the key to effective reform. The report serves to guide However, strong examples where services work show that governments, development agencies, and service governments, service providers, and citizens can do providers on how they can improve the delivery of basic better. The World Development Report 2004 indicates services and accelerate progress toward the Millennium that putting poor people at the center of service provision Development Goals. ...................................................................................................................................... Title Stock # Price Oty. Total US$ World Development Report 2004: Making Services Work for Poor People A Copublication of the World Bank and Oxford University Press. Paperback. September 2003. (ISBN 0- 8213-5468-X). D15468 $26 Hardcover. September 2003. (ISBN 0- 8213-5537-6). D15537 $50 Shipping and Handling charges are $.o00 per order If a purchase order is used, actual shipping will be charged. Subtotal For air mail outside the U.s.. charges are US$7 00 per order plus US$6. 00 per item. PAYMENT METHOD Shipping and Handling Orders from Individuals must be accompanied by payment or credit card information. Total US$ Credit cards are accepted only for orders addressed to the World Bank. Check with your o a local distributor about acceptable credit cards. Please do not send cash. L YES, please send my copy of World Development Report 2004 PAYMENT METHOD (Please Print) Orders from individuals must be accompanied by payment or credit card information. Credit cards accepted only for orders addressed to the Name World Bank. Please do not send cash. Title O Charge $ to my: Oa American Express O Mastercard O Visa Organization__________________________ O Card no. Address Expiration date / City Name State as it appears on the card Sigaur_____________________________ Zip/Postal Code Signature_required for all credit card charges Ot Check no._ Country __________________________ in the amount of $_ _ _ is enclosed. When ordering Phone directly from the World Bank, make check payable in U.S. funds drawn on Fax a U.S. bank to: The World Bank. Please send your check with your order Institutional customers in the U.S. only: E-mail Bill me. Please include purchase order Mail order to: World Bank Publications PO. Box 960, Hemdon, VA 20172-0960, USA, or Fax to 703-661-1501. Order by phone: 703-661-1580 or 800-645-7247. Order online: www. worldbank. org/publications Questions? E-mail us at books@worldbank.org DCWR4 fi16k WORLD BANK Visit our website at Publications www. worldbank. org/publications * ., '- _2- world development reportt Making Services Work for Poor People "' Too often, services fail poor people - in access, in quality, and in affordability. But the fact that there are striking examples where basic services such as water, sanitation, health, education, and electricity do work for poor-people mneans that governments and citizens can do a better job of providing them. Learning from success and understanding the sources of failure, this year's WMorld Development Report - Making Services Workfor Poor People- argues that services can be improved by putting poor people at the center af snervice provi- sion. How? By enabling the poor to monitor and discipline service pr6vidiers, by amplifying their voice in policymaking, and by strengthening the incentives for providers to serve the poor. Freedom from illness and freedom from illiteracy are two of thc miiost impor- tant ways poor people can escape from poverty. To achieve these goals, eco- nomic growth and financial resources are of course necessary, but they are not enough. The World Development Report provides a practical framework for making the services that contribute to human development work for poor people. With this framework, citizens, governments, arid donors can takge action and accelerate progress toward the common objective of poverty reduc- tion, as specified in the Millennium Development Goals. g 2'.. 9 780821 ,56~7 1 ) ~~~~~~~~~~~~~~~9 78I32I 356371 .Q ' .~ E n .-^ ., ..:IA