Accountability in Public Services Public Disclosure Authorized in South Africa Public Disclosure Authorized Selected Issues Public Disclosure Authorized Public Disclosure Authorized Accountability in Public Services in South Africa

Selected Issues

March 2011 © 2011

Photo credits: Cover and pages 9, 33, and 65 by John Hogg; pages 13, 25, 53, 71, 87, 105, 117, and 123 by Trevor Samson.

Design, editing, and production by Communications Development Incorporated, Washington, DC. Contents

Preface ix Acknowledgments xiii Executive Summary 1 Introduction 9

Chapter 1 Progress in Service Coverage and Quality 13 Setting the stage 13 Challenges to public services 14 Public expenditures on basic services 16 Targeting of government spending 18 Assessments of services 19 The rising tide of service delivery protests 22

Chapter 2 Citizen Voice 25 Elections 26 Political accountability at the local level 26 Broader political accountability 28 Conclusion 31

Chapter 3 The Compact 33 The compact 33 Measuring inputs, outputs, and outcomes 37 Procedures 37 Silos 38 Community Development Workers 40 Enclaves 41 Contracts 42 Decentralization: international experience 43 The system of decentralization in South Africa 46 Performance of local governments in South Africa 49

Chapter 4 Citizen-User Power 53 The Batho Pele principles of consultation and redress 54 Improving information flow 55

Accountability in Public Services in South Africa iii Provider choice and user fees or copayments 60 Participatory local government budgeting and development planning 61 Community-driven development 62

Chapter 5 Participatory Assessment of Service Delivery 65 The municipalities 65 The methodology 66 Results by service 67 Results by municipality 68 Perceptions of communities 69 Perceptions of service providers 69 Conclusion 70

Chapter 6 Education 71 Change and challenges 71 Learning outcomes 71 Public spending 73 Public spending and learning outcomes: an association 74 Funding the schools 74 The enabling—­ and­ disabling—­ environment­ for teachers 76 Accountability at the school level 81 Policy recommendations and research agenda 84

Chapter 7 Water and Sanitation 87 The changing approaches to service provision 88 Infrastructure provision for water services 89 Deteriorating water services 94 Accountability relationships in the water sector 95 Conclusions and policy recommendations 101 Notes 103

Chapter 8 Toward Local and Community‑Driven Development 105 Approaches toward community and local government empowerment 105 Centralization versus decentralization 106 Participation in sectoral programs 107 Social action funds 108 Community-driven development and decentralization 109 Accountability mechanisms 110 Toward a model of coproduction of services 110 The status of and prospects for local and community-driven development 112 Conclusion 114

iv Accountability in Public Services in South Africa Chapter 9 Approaches and Tools for Effecting Change 117 Quality control and lean production in automobile manufacturing 117 World Bank and South African experience 118 Monitoring inputs and outputs 119 Impact evaluation 120

Chapter 10 Conclusions 123 Broad recommendations 125 Sector-specific recommendations 125

Annex 1 Policy Matrix 127

Annex 2 South Africa’s Legal Framework for Community-Driven Development Grant Funding Mechanisms 131

Annex 3 Scaling Up Local and Community‑Driven Development 133

Annex 4 Land Reform in South Africa: Legacies, Limits, and Innovations 139

Annex 5 Project Consolidate: Hands‑On Local Government Support and Engagement Program 145

References 147

Boxes 1.1 the Community Works Program 14 1.2 the Treatment Action Group 16 2.1 elections in South Africa 27 2.2 Improving local development through community-based planning 30 4.1 citizen report cards: Bangalore and Uganda 57 4.2 reporting absenteeism in the health and education sectors 59 5.1 tzaneen politicians’ views of their role 68 6.1 School-based performance awards in Chile 81 6.2 Meta-analysis of school-based management initiatives in the United States 83 7.1 Instances of disease attributable to poor management of water services 94 7.2 examples of current institutional models for water services provision 98 7.3 focus groups in eThekwini Metro 100 7.4 citizen awareness of rights and responsibilities 101 8.1 the Comilla model of rural development 107 8.2 the contest between centralization and decentralization in India, 1948 to present 107

Accountability in Public Services in South Africa v 8.3 the Bamako Initiative in Benin, Mali, and Guinea 108 8.4 Design principles for community-driven development 110 8.5 Six innovations in financial accountability for community project management 111 8.6 linking sector programs, local government, and communities 111 8.7 Proposed program of community development grants 114

Figures 1 Service provider institutions: the short and long routes to accountability and the three relationships 1, 9 2 Accountability relationships and functions 11 1.1 real expenditure on basic public services, 1995/96–2009/10 (2005 prices) 16 1.2 expenditure on basic public services as a share of GDP, 1995/96–2009/10 17 1.3 Major expenditure areas as a share of total budget, 1995/96–2009/10 17 1.4 expenditure on basic public services as a share of total budget, 1995/96– 2009/10 17 1.5 Per capita expenditure on basic public services, 1995/96–2009/10 18 1.6 Social protests, 2004–2010 22 3.1 example of a silo’s functions in a land reform project 39 3.2 Decentralization of services 44 3.3 Decentralizing accountability relationships 44 3.4 transfers by type, 2003/2004–2009/2010 49 3.5 Key interventions proposed by the Local Government Turnaround Strategy 50 4.1 Share of funds reaching Ugandan schools 56 5.1 Scores of accountability, by service and accountability leg 69 5.2 Accountability challenges perceived by service providers (percent) 70 5.3 Accountability challenges perceived by politicians (percent) 70 6.1 Average grade 8 Trends in International Math and Science Study math test score and difference between poorest and richest quintile, with share of difference due to within- and between-school differences 72 6.2 Average grade 5 Progress in International Reading Literacy Study reading test score and difference between poorest and richest quintile, with share of difference due to within- and between-school differences 72 6.3 Difference between poorest and richest quintile in Southern Africa Consortium for Monitoring Educational Quality math and reading test score, with share of difference due to within- and between-school differences 73 6.4 expenditure incidence of primary and secondary schools, public spending over time in South Africa and selected developing countries 73 6.5 efficiency index of public spending on education in each quintile 74 6.6 Premiums in monthly and hourly earnings of public sector teachers, private sector teachers, and other public sector employees, relative to private sector employees (1995 to 2006) 79 6.7 Students attending schools where the principal reports that more than 50 percent of parents participate in three types of school-based activities, by socioeconomic quintile of the student 82 7.1 Access to water and sanitation in South Africa, 2008 89 7.2 user satisfaction with municipal water services, 2005–2009 92 7.3 Average payments for water services, 2004–2009 92 7.4 Key trends in networked water and sanitation 93

vi Accountability in Public Services in South Africa 7.5 civil engineering professionals in local government, 2005 93 7.6 the formal division of responsibility for water services provision 95 7.7 Applying the World Development Report triangle to water servicing in South Africa, 2008 96 8.1 Index of sector decentralization in 19 countries 109 A1 basic institutional requirements 134

Tables 3.1 forms of decentralization 45 3.2 vertical division of revenue, 2003/2004–2009/10 (ZAR million) 49 4.1 Absenteeism rates in developing countries (percent) 57 4.2 teacher absenteeism rates in Indian states (percent) 58 4.3 teachers absent during an unannounced visit to randomly selected schools 58 5.1 characteristics of the six municipalities 65 5.2 Satisfaction with access to services across municipalities (percent) 66 5.3 complaints and compliments about quality of services 67 5.4 total adverse comments across municipalities, by accountability leg 69 6.1 Spending per student by type of spending and school quintile, 2005 (ZAR per student) 74 6.2 Disbursed public spending on school grants that actually reach schools: results from Public Expenditure Tracking Surveys 75 6.3 time devoted to teaching and learning in basic education (percent) 77 6.4 estimates of teacher earnings “premium,” based on analysis of household survey data 78 6.5 estimates of teacher earnings “premium,” based on analysis of individual and household survey data: pooled data from 1995 to 2006 79 6.6 Actual and simulated teacher earnings at different points in the earnings distribution 80

Accountability in Public Services in South Africa vii

Preface

In 1994, the legacy of apartheid presented the Rationale and justification new South Africa with enormous challenges: This study attempts to shed light on service poverty, inequality, and the immense aspira- delivery capacity and quality challenges in South tions for greater access to basic public services. Africa by analyzing the institutions that “pro- The government immediately began address- duce” services. “Institutions” are not defined as ing these challenges by enshrining constitu- “organizations” but as the rules that govern the tional rights to service access, radically reform- interaction between service providers and ben- ing economic and sector policies, and funding eficiaries. This study examines the accountability ambitious service delivery programs. A new relationships among citizen-users, policymakers, governance and administrative framework for and service providers. These relationships can racially diverse provinces and municipalities be seen as the three legs of a triangle: the first had to be built­—­and the new structure had leg represents accountability between citizen- to progress quickly in delivering the improved users and politicians/­policymakers, the second services expected by the people. These chal- between politicians/­policymakers and service lenges have grown in the past 16 years, as providers, and the third between service provid- South Africa’s population has increased with ers and citizen-users. immigration and the HIV/AIDS crisis has This study is less concerned with the organo- spread. gram (organizational chart) of a particular min- Since passing the Constitution, the govern- istry, the skill level of civil servants, the content ment has continued to spend on basic public of the teaching curriculum, or the adequacy of services and on public order and safety. Despite overall budget allocations. Instead, it concen- the enormous effort and money put into new trates on the “rules of the services game.” What policies and institutions for improved service incentives do stakeholders have to deliver ser- delivery, providing coverage and quality ser- vices effectively and efficiently? The fundamen- vices for the rapidly growing, mostly black and tal question posed by this study is whether the poor population remains a colossal challenge. organizations in charge of deciding what ser- The government and civil society are increas- vices to deliver and how those services should be ingly concerned that the huge investment in delivered are accountable to the citizens they are development is having less impact than ex- meant to serve. pected. Inadequate services, combined with A number of hypotheses will be presented unemployment and continued poverty, have to try to explain the shortfall in service deliv- frequently ignited township protests and riots ery. These hypotheses are based on the 2004 (known as “service delivery protests”) and have World Development Report’s “Making Services spurred vigorous policy analysis, debate, and Work for Poor People,” which proposes that the reform proposals. more service providers are held accountable by Creating jobs and improving services have citizens, the more service delivery improves; to emerged as the highest priorities of President improve quality and coverage of public services, Jacob Zuma’s administration, which emphasizes “the key is to enhance the power of poor clients accountability for results in all tiers of govern- in service provision” (World Bank 2003, p. 64). ment, as demonstrated by the establishment of a “When governments do not feel the pressure to new Ministry for Performance Monitoring and respond to citizen demands, when they are in- Evaluation. capable of enforcing basic performance rules

Accountability in Public Services in South Africa ix on those directly responsible for services deliv- as much consideration and respect as they ery, and when citizens have no control or choice would like for themselves. The public service over service providers, typically service will fail” is committed to honest, transparent commu- (Fiszbein 2005, p. 4). nication with citizens. This theory, which is finding increasing -em • Providing of information. Information on pirical support, lays at the center of several suc- services should be available at the point of cessful approaches to development, such as care- delivery, but other arrangements will be fully managed decentralization, citizen report needed for users who are far from the point cards, and community-driven development. of delivery. • Openness and transparency. The public should Batho Pele know how national, provincial, and local In 2002, recognizing the service delivery chal- government institutions operate, how well lenge, the government launched Batho Pele­—­ they use resources, and who is in charge. It Sesotho for “people first”­—­an initiative that is anticipated that the public will take advan- aims to improve access to public services through tage of this principle by suggesting improve- increased transparency, accountability, and citi- ments to service delivery mechanisms and zen involvement in public service planning and raising queries with government employees operations. The Batho Pele message resonates to hold them accountable and responsible. strongly with that of the 2004 World Develop- • Redress. Emphasis is needed on quickly and ment Report: services can improve by empower- accurately identifying when services are fall- ing poor people to monitor and discipline service ing below the promised standard and on providers by raising their voices in policymaking having procedures to remedy the situation. and strengthening incentives for service provid- • Value for money. Many improvements the ers to serve the poor. public would like to see require no addi- The Batho Pele policy actions to improve tional resources and can sometimes even re- public service delivery include: duce costs. Failure to give a member of the • Consultation and choice. Citizen-users should public a simple, satisfactory explanation to be consulted in a number of ways: customer an inquiry may, for example, result in an in- surveys; interviews; group consultations; correctly completed application form, which meetings with consumer representative bod- will take time to fix. ies, nongovernmental organizations, and Note that all the Batho Pele principles at- community-based organizations; and In- tempt to strengthen accountability between tegrated Development Plans. Consumers citizens and service providers. Objectives of should also be given real choices. this report include assessing how well the Batho • Precise and measurable service standards. In Pele principles are being practiced and suggest- consultation with users, the government ing possible improvements and areas for further should benchmark citizens’ satisfaction with research. the services or products they receive from departments. To make South Africa glob- Outline of the report ally competitive, standards should be bench- The introduction discusses the accountability marked (where applicable) against those used framework developed by the World Bank that internationally, taking into account South will be used for the report. Africa’s current level of development. Chapter 1 sets the stage for the rest of the re- • Increased access. The government should rec- port by broadly assessing South Africa’s progress tify the apartheid inequalities in service dis- on service access and quality, summarizing re- tribution. Access to information and services cent policy initiatives, and highlighting salient empowers citizens, creates value for money, considerations about South Africa: its excellent improves service quality, and reduces unnec- Constitution, which enshrines universal rights essary citizen spending. to basic services; its well-functioning democ- • Ensured courtesy. Service providers should racy; and its substantial increase in pro-poor empathize with citizens and treat them with spending since 1994 on vital public services such

x Accountability in Public Services in South Africa as education, health, social grants, and housing. chapter will show how scaling up community- It then summarizes the most important govern- driven development can be effective for location- ment programs, offers objective assessments of and beneficiary-specific public services that are public service delivery, and identifies where this not characterized by economies of scale. A South study can add value to the debate. African community-based planning approach Chapters 2, 3, and 4 describe the international currently being scaled up will be reviewed, along and South African experience with mechanisms with a proposal by the Ministry of Cooperative that seek to improve accountability­—­public sec- Development and Traditional Affairs about how tor reform, citizen report cards, and others­—­and to scale up community-driven development. posits hypotheses to be explored in the following Chapter 9 explains the continuing “learn- chapters. ing” practices pioneered in the manufacturing Chapter 5 applies the World Bank’s account- sector and addresses how these practices might ability framework to a participatory assessment be used by the South African government to ef- of services in six municipalities in South Africa. fect change. Some suggestions to improve moni- Chapters 6 and 7 apply the framework to the toring and evaluation in South Africa are also education and water and sanitation sectors. presented. Chapter 8 explains why community-driven Chapter 10 summarizes the conclusions, development does not factor in any main South translates these into main hypotheses to be African development programs, though simi- tested in future work, and formulates a number lar approaches were successful before 1994. The of policy recommendations for public debate.

Accountability in Public Services in South Africa xi

Acknowledgments

This study was identified in the joint Country Dave DeGroot, Michael Kihato, Ntombini Partnership Strategy 2008–2011 between the Marrengane, Zimkhitha Mhlanga, and Rogier Government of South Africa and the World van den Brink; and Hans Binswanger, Ronald Bank under the theme “capacity building for Henderson, Zweli Mbhele, and Kay Muir- public services.” The work was undertaken in Leresche. Bruce Ross-Larson and Rob Elson partnership with the National Treasury, Inter- of Communications Development Incorporated governmental Relations division, led by Deputy- (Washington, DC) edited the manuscript. Director General Lungisa Fuzile, who provided overall guidance. The report’s goal is to initiate Supporting research discussion within and outside government on The report relies heavily on several recent studies how to improve public services. by teams of World Bank staff and local research- This report was written by a team led by Rogier ers.1 The research has led to a number of rea- van den Brink that included Hans Binswanger, sonable hypotheses, which should be tested by Sam Chimbuya, Dave DeGroot, Kathy Eales, stakeholders in future research. Deon Filmer, Steven Friedman, Ian Goldman, Vincent Hungwe, Rod MacKinnon, Ntombini Participatory service assessment in six municipali- Marrengane, Zani Naudé, Laïla Smith, Servaas ties in South Africa. Participatory assessments of van der Berg, Ahmedi Vawda, Michele Zini, services access and quality in five services (water and Phindile Ngwenya. Peer reviewers were supply and sanitation, housing, health, edu- Daniel Cotlear and Brian Levy. The respon- cation, and social grants) were undertaken by sible technical manager was John Panzer of the the Khanya-African Institute for Community- Poverty Reduction and Economic Management Driven Development (Khanya-aicdd), a local unit for Southern Africa. The study was initi- nongovernmental organization. The six munici- ated under the guidance of Ritva Reinikka, when palities were selected to represent a wide variety she was country director for Botswana, Lesotho, of situations. Madagascar, Mauritius, Namibia, South Africa, and Swaziland. It was updated in June 2010 by Tshwane citizen report card. The Human Sci- Hans Binswanger, Deon Filmer, Kathy Eales, ences Research Council of South Africa and the Laila Smith, and Rogier van den Brink under World Bank (led by Craig Schwabe and Caro- the leadership of the current country director lyn Winter) piloted a citizen report card project Ruth Kagia. The update benefited from the com- in South Africa’s capital. Tshwane, with a very ments of Ronette Engela, Martin Gustaffson, detailed service delivery compact, is obliged to Christopher Loewald, Neva Makgetla, and monitor its activities regularly, but it has no sys- Themba Zulu at a June 15, 2010, workshop on tem to assess outcomes. The exercise produced a the report. Sandeep Mahajan guided the re- user evaluation of the main services provided by port through the update. Background papers the municipality, ranked according to their cus- were prepared by Doreen Atkinson and Lolette tomers’ satisfaction rating. Kritzinger-van Niekerk; David Everatt and Lulu Gwagwa; Craig Schwabe and Carolyn Winter; Education. Servaas van der Berg (University of Servaas van der Berg and Deon Filmer; Somik Stellenbosch) and Deon Filmer reviewed lit- Lall, Rogier van den Brink, Kay Muir Leresche, erature on the education sector, conducted an and Basab Dasgupta; Stephen Berrisford, expenditure incidence analysis, and tested a set

Accountability in Public Services in South Africa xiii of survey instruments. This work served as the illustrate the difficulties that such projects face, foundation for a full-scale school survey under- the study focused on one community, composed taken in 2008–2009 under the auspices of the largely of laid-off farm workers that more than Department of Education and financed by the a decade ago wanted to buy their own farm in Directorate General of International Coopera- a peri-urban area west of Johannesburg. Their tion of the Netherlands. A qualitative assess- dream was to establish a mixed-use settlement. ment using focus groups in rural communities They called the village Ethembalethu­—­“Our informed the work program’s broader analysis Hope.” About 250 families started their own of the relationship among schools, parents, and association and savings scheme to establish the communities. It also informed the development village. By 1997, they had saved enough money­ of the school survey instruments. —ZAR125,000­ (about US$18,000)—­ to­ make their first purchase offer. The study tracks the Responsibility for services. Doreen Atkinson and numerous challenges faced by the community Lolette Kritzinger-van Niekerk explored a range over the years in making their dream a real- of aspects of the existing responsibility assign- ity. The report surfaced as a World Bank work- ment for services. They argued that the discrep- ing paper (Berrisford and others 2008) and also ancy between inputs and service outputs and formed the basic script of a video case study outcomes could be explained partly by the evo- produced by Black Earth Communications and lution of South Africa’s system for intergovern- published on YouTube.3 mental relations, which has resulted in systemic policy and program challenges for services, * * * particularly for municipalities. Service difficul- The above studies, combined with a literature ties are exacerbated by the weak accountability review, were used to create hypotheses across of municipal politicians, service providers, and themes and sectors. After a World Bank review, citizens. But stronger top-down regulatory, a concept note for the current report was written fiduciary, and other state oversight systems have in accord with the National Treasury. An initial strengthened the compact among national, pro- version of the report was completed in June 2008 vincial, and local governments, partially com- and circulated in the government, where it was pensating for the lack of direct accountability read widely. Formal discussion and dissemina- between citizens and service providers. tion of the report were postponed until after the 2009 election­—­until the new government had Subsidized housing and access to land in South Afri- time to review policies and implementation with can cities. Introduced as part of the Reconstruc- respect to services. The present report, com- tion and Development Programme, the gov- pleted in July 2010, has benefited from a work- ernment has focused on providing subsidized shop with government officials and a subsequent housing built by construction companies con- update. tracted by provincial departments of housing.2 Using new household survey data, Somik Lall, Notes Rogier van den Brink, Kay Muir Leresche, and 1. Local partners included the Human Sci- Basab Dasgupta conducted an initial impact ences Research Council, the University of evaluation of the Reconstruction and Develop- Stellenbosch, Khanya-aicdd, and several ment Programme’s housing initiative. local consultants. 2. “Subsidized Housing and Access to Land in In Search of Land and Housing in the New South South African Cities” by Somik V. Lall, Ro- Africa: the Case of Ethembalethu. This case study gier van den Brink, Kay Muir Leresche, and was undertaken at the request of Mogale City Basab Dasgupta. for World Bank technical assistance on inte- 3. http://www.youtube.com/watch?v= grated housing and agriculture projects. To RZxXsG6mO3g.

xiv Accountability in Public Services in South Africa Executive Summary

This study is concerned primarily with incen- democracy, internal state capacity and account- tives and accountability. What incentives do the ability, and citizen participation (World Bank stakeholders involved in providing services have 2003; Fiszbein 2005). to deliver services effectively and efficiently? The fundamental question imposed by this study is South Africa’s achievements whether the organizations in charge of deciding South Africa has already attained several impor- what services to deliver and how those services tant achievements on which it can build to should be delivered are accountable to the citi- develop an effective public service delivery sys- zens they are meant to serve. tem, including: • Democracy. Accountability relationships in public • The Constitution, including the constitu- services: conceptual framework tional mandates for basic services. The World Development Report’s 2004 frame- • The decentralization framework. work identifies three different accountabil- • The intergovernmental fiscal system. ity relationships among citizens, politicians/­ • The Batho Pele (Sesotho for “people first”) policymakers, and service providers. These principles. relationships form two routes to accountability, • Sector-specific legal and policy frameworks. a long one and a short one (figure 1): • A vibrant civil society. • The political process citizens try to use to in- But stakeholders agree that implementation fluence politicians is called “citizen voice”­—­ has fallen short of citizen aspirations and the the long route’s first “leg.” government’s goals. • The relationship between politicians/policy- This report begins by analyzing trends in makers and service providers is the second public spending. Overall spending on public ser- leg of the long route­—­“the compact.” Many vices has increased dramatically since 1994, but of the government’s current initiatives to im- Figure 1 Service provider institutions: the short and long routes to prove accountability focus on this leg. accountability and the three relationships • The short route—­ the­ third leg—­ considers­ how much the citizens, now acting as clients The state of public services, can directly pressure the Politicians Policymakers service providers to ensure efficient delivery­ —­we call this “client power.” Compacts Contracts Accountability is strongest when both the Long route Citizen voice long and short route to accountability work. This is much easier said than done. A growing body Citizens/clients Providers of research demonstrates that “getting these ac- countability relationships right” is the key to im- Coalitions/inclusion Client power Management proving public services for the poor. Getting it Nonpoor Poor Frontline Organizations Short route right means that accountability relationships are all working properly, reinforcing each other as Services a public service delivery system. This method- ology emphasizes strengthening both the short Source: World Bank 2003. and long route to accountability­—­strengthening

Accountability in Public Services in South Africa 1 weak per capita spending on health and education has “delivery” agencies of the national sphere rather accountability increased only 20–25 percent over 15 years­—­ than as representative institutions accountable systems underpin less than usually assumed­—­while real per capita to citizens. To strengthen political account- spending on housing has fallen 12 percent since ability, communication with all citizens must the failure of 1990. be improved and institutionalized, particularly widespread The “double whammy” of population in- among the poor and those living in informal effective service crease and smaller household size presents settlements. unique challenges for South Africa. Population South Africa’s sectoral legal and policy delivery increased 20 percent, from 40.5 million in 1996 frameworks place a high priority on citizen- to 48.5 million in 2007, and the average house- user participation in public administration hold size decreased from 4.48 people per house- and service provision­—­a view enshrined in hold to 3.88. Thus, the number of households the Constitution. But participation is often increased 38 percent. Except for social security, missing in delivery. The PSC found that “part welfare, and public order and safety, spending of the implementation weaknesses are the on most basic public services as a share of GDP non-involvement of beneficiary communities has remained constant or declined. in the planning and implementation of pro- The study then reviews the South African grams . . . departments generally do not adopt government’s self-assessment of service provi- project management approaches that allow for sion. As noted by the Public Service Commis- effective beneficiary participation and align- sion (PSC)­—­an independent and impartial ment of projects with local development plans” body created by the Constitution to investi- (PSC 2008a, p. 40). gate and evaluate the performance of the public The government developed ambitious deliv- administration­—­in several recent assessments, ery targets to overcome the backlog in services most sectors have long had a coherent and well- and provided ample financial resources to meet crafted legal and policy framework to support these targets. But national, provincial, and local the mending of apartheid inequities. Yet it finds institutions for service delivery were still weak, that implementing these frameworks to improve along with technical and managerial capac- citizen well-being remains a serious challenge. ity. Many sectors stressed community and user The Presidency’s Towards a Ten Year Review, the participation initially, but participation systems African Peer Review Mechanism, and the sec- were time-consuming and hard to scale; policy- tor chapters in this report conclude that weak makers feared this would slow implementation. accountability systems underpin the failure of Consequently, the sectors relied on centralized widespread effective service delivery. mechanisms focused on outputs, rather than on more participatory approaches focused on out- Batho Pele: missing in “delivery” comes. The sector programs invariably became In principle, South Africa’s parliamentary “silos”­—­supply-driven and focused on the deliv- democracy, combined with a free press, allows ery of sector outputs­—­and in the process took citizens to influence policymakers and express power from citizens. Intense delivery pressures their voice. But a number of issues weaken citi- and a focus on quantitative output and input tar- zen voice, or political accountability in the broad gets were therefore partly, albeit inadvertently, sense. These issues do not seem to be rooted in responsible for undermining the intentions of the national, provincial, and local electoral sys- the Constitution, the Batho Pele principles, and tems or processes. More fundamental issues are the original sector strategies and policies. at play. The South African experience is not unusual. The South African development paradigm Losing focus on citizen participation in favor of has focused on service delivery, and meeting centralization and delivery silos is not unique to national infrastructure development targets has South Africa. The same has happened in many undermined citizens’ choices of what services places, including India, the Philippines, most are needed and their participation in how to African countries, and Latin America. Many deliver them. Centralized planning has tended countries today face the same service challenges to use provincial and local governments as as South Africa.

2 Accountability in Public Services in South Africa Another unintended consequence of the sup- municipalities studied do not. Subject to further citizens feel ply-driven approach in South Africa is that citi- research, this suggests that reforms would prob- their loss of zens now expect the government to assume full ably require initiatives to change the mindset of participation and responsibility for service delivery, including reg- important stakeholders. ular maintenance for even private assets such as The participatory assessment found that accountability household toilets. Service problems are increas- water and sanitation (a municipal responsibility acutely ingly blamed on the government. Discontent has and the most decentralized service) consistently exploded into many episodes of unrest as ambi- received the best evaluation, despite considerable tious government targets for development and sustainability problems. Healthcare services, increased citizen expectations overwhelm capac- mandated to be managed at the provincial level ity, especially for municipally or provincially de- without any local community oversight, received livered services. the worst. As background for this study, the partici- patory assessment of service delivery in six mu- Education nicipalities held focus groups with politicians, Education rated high in access but not in qual- service providers, and citizen-users. The study ity. Education service delivery is also managed has led to several hypotheses about where in the at the provincial level, but it benefits from an long and short routes to accountability main important community oversight institution­—­ challenges stand in the way of better service School Governing Bodies. Student education outcomes: achievements (outcomes) differ significantly by • Politicians, service providers, and citizen-us- how effective these bodies are at guiding school ers agreed that accountability problems be- decisions. tween citizen-users and politicians were few. An expenditure incidence analysis shows a Similarly, few problems were identified with modest improvement in reducing the small re- the policy frameworks in each sector. maining public spending inequality between • Despite well-designed policies, politicians, rich and poor communities between 2000 and service providers, and citizen-users all see 2005. While the spending imbalance of the many problems in the second leg of account- apartheid era has been corrected­—­an important ability: the compact between politicians and achievement­—­the benefit in terms of the learn- service providers. ing outcomes of poor students lags. • Significant disagreement exists between pol- What could explain this? We suggest a num- iticians/policymakers and service providers ber of reasons: on one hand and citizen-users on the other • Teacher pay structures provide the wrong on how well accountability between service incentives. While better educated and more providers and citizen-users is working. Nei- experienced teachers are underpaid, teachers ther politicians nor service providers see this with less attractive characteristics or endow- accountability relationship as a serious prob- ments earn more than employees in other lem, so few recent initiatives to improve ac- sectors, providing the perverse incentives countability focus on this leg. But citizens of attracting and keeping lower performing give accountability problems in this relation- teachers in the profession. ship weight equal to those between politi- • “Time on task”­—­the classroom time that cians and service providers. teachers actually devote to teaching and It appears that citizens feel their loss of par- learning activities—­ is­ low. ticipation and accountability acutely. Weak par- • Deficiencies in school budget management. ticipation and short-route accountability could Schools that receive a “paper budget” must partly explain the limited impact that public rely on provincial procurement for access to spending has had on the quality and availability funds, goods, and services, which do not al- of public services. ways arrive on time or in the quantities in- The PSC and many senior policymakers and tended. Schools that ostensibly have author- researchers cited in this report share this view, ity over their budget face many restrictions but politicians and service providers in the six on how to use it.

Accountability in Public Services in South Africa 3 better • Most schools have little decisionmaking au- and sanitation problems have become critical in understanding thority, and parents have little information many municipalities. of actual school on, and minimal control over, budgets and The approach to service delivery in water teacher performance, resulting in a lack of and sanitation is similar to many others: state- financial local accountability. centric, top-down, target-driven, supply-led, operations would What could improve the benefit incidence? generously funded, and often fully subsidized. help identify how The short route to accountability is weak in most This approach has skewed infrastructure in- South African schools­—­especially among those vestment decisions toward highly capital- and to further the serving the poorest communities. It is important skills-intensive solutions, aggravating the pre- reform process that information on education outcomes be dis- existing skills shortage. Since most municipali- seminated so that progress can be tracked and ties have opted for in-house municipal provision parents can know how schools are performing. rather than contracting, the situation has wors- Ensuring that parents are empowered to hold ened further, raising the risk of service failure, schools accountable would strengthen citizen increasing recurrent cost liabilities, and under- voice. mining infrastructure sustainability. Increasing school decisionmaking respon- The approach has also usurped power from sibility is also critical. Carefully crafted school citizens. And there are many ways to improve decentralization­—­providing increased auton- the accountability relationships among the dif- omy in decisionmaking on financial operations, ferent spheres of government. Some specific rec- on procurement of goods and services, and on ommendations, including strengthening regu- hiring and firing educators­—­coupled with more latory oversight in the water sector, are offered information on local school performance, can in the report. But much more attention needs to empower schools and make them more account- focus on the relationship between citizens and able to children and parents. municipal service providers. To strengthen the compact between gov- The primary responsibility of holding local, ernment and frontline providers, the operation municipal service providers accountable should of the paper budget in non-Article 21 schools not lay with a remote regulator. Citizens need could be reviewed. Likewise, restrictions to direct, short-route accountability mechanisms to funds flowing to Article 21 schools could be re- reroot service providers and local political repre- viewed for their usefulness. Reported delays in sentatives. They should have recourse to a regu- procurement, and uncertainties about remaining lator, but the primary engagement and response budgets, create an environment in which schools must be local because that is usually where the cannot plan effectively. problem is felt most acutely and where immedi- A better understanding of actual school fi- ate intervention is required. nancial operations would help identify how to Four measures would make a substantial further the reform process. Teacher pay struc- impact: tures should also be reviewed to create more in- • Dedicate significant resources in each mu- centives for good teaching and to retain effec- nicipality to building public understanding tive, experienced teachers. of how water and sanitation services work and help citizens understand their rights, Water and sanitation roles, and shared responsibilities. This will In water and sanitation, a municipal responsi- build mutual accountability relationships for bility, service deficiencies are most pervasive in service provision. the former “homeland” areas, where munici- • Experiment with local and community- palities were established for the first time in driven water services approaches where 2000. Political commitment to meet ambitious appropriate­—­with vouchers for toilets, for goals for universal service coverage is strong, instance. and expansion in infrastructure coverage has • Strengthen public accountability in select- been extraordinary. But the quality, mainte- ing service providers, preferably through re- nance, and sustainability of these achievements vising the Municipal Systems Act to permit are in danger of being compromised. Water more integrated and inclusive assessment

4 Accountability in Public Services in South Africa of provider options such as service partner- spheres of government, service providers, com- municipal capacity ships with local businesses and/or commu- munities, and citizen-users in a combined “co- varies enormously nity-based organizations, particularly in production” approach to service delivery. Lack- and many struggle rural areas where municipal service providers ing are joint operational manuals between silos struggle to provide the support and manage- and spheres that describe the responsibilities of to attract people ment required. different coproducers and how they should be with technical and • Strengthen the compact and the regulatory carried out, are lacking. And many financing managerial skills framework by using results-based perfor- mechanisms now fragmented into separate silos mance management, revising the conditions and spheres need reform. attached to infrastructure grants, and pro- Decentralization can encourage citizen-user viding incentives for the proper maintenance participation in the design, implementation, and of new and existing infrastructure. monitoring of services, but this potential has not been fulfilled. Our research suggests that Decentralization participation is often confined to consultations Has decentralization helped or hindered service with citizens, including in preparing South Af- provision? South Africa has made tremendous rica’s Integrated Development Plan, intended to progress establishing a constitutional framework be a main instrument for participatory develop- for decentralization, creating racially diverse ment planning and budgeting. Policymakers and municipalities and provinces to replace an service providers often minimize citizen par- incredibly complex set of race-based local juris- ticipation because it risks slowing delivery and dictions. Decentralization also benefits from missing infrastructure development targets. But an excellent intergovernmental fiscal system, reviewing international findings on local and but municipal capacity varies enormously and community-driven development—­ the­ decen- many struggle to attract people with technical tralization of services along with the resources and managerial skills. Meanwhile, municipali- to plan and implement services­—­shows that ties bear the brunt of service discontent, even for participation can be scaled up cost-effectively those whose services they are not responsible. and incorporated into massive delivery programs This study hypothesizes that decentralizing without slowing them. This would require com- service responsibilities is often incomplete, even plex training and facilitation logistics and inno- where it has the potential to improve services. vative financial accountability mechanisms. In many sectors, decentralization in South Af- Fortunately, financial policies, planning ap- rica is merely a kind of delegation, with lim- proaches, and implementation mechanisms for ited devolution of responsibilities and resources some municipal devolution of development and to the local level. Each ministry develops pro- service functions are already established and grams in which responsibilities for services are have shown promise in application. The Ministry delegated to provinces and municipalities. Pro- of Cooperative Governance and Traditional Af- grams reach the final provider or end-users in fairs has issued a proposal to widen these mech- fragmented “silos,” which are difficult to coor- anisms. This study’s findings support that these dinate and integrate into a coherent whole. Con- proposals be fully debated and implemented. sider the following example: to establish a farm, In addition, community works programs­—­also a South African that has received land under a built on participation, community empower- reform program must apply for separate fund- ment, and decentralization­—­should be reviewed ing for planning, land acquisition, agricultural for their potential in other communities. investments, advisory services, and housing­—­an Governments sometimes respond to poor almost insurmountably complex set of processes service delivery by recentralizing implementa- and silos for both the service providers and the tion and control­—­but this is risky. If decentral- beneficiaries. ization is incomplete and client participation is The debate centers on dividing roles and re- weak, recentralized service delivery will worsen sponsibilities between national and provincial service provider accountability to citizens. In- governments. Policy should instead focus on ternational evidence suggests that services that defining the specific responsibilities of different are difficult to monitor and require discretionary

Accountability in Public Services in South Africa 5 Governments decisionmaking by the provider (health and then passed to departments and provincial and sometimes education, for instance) should be progressively local governments. These government tiers could respond to poor decentralized, not recentralized, to improve im- learn from the problem solving systems used in pact. Centralized planning should be realigned the automobile industry and implement con- service delivery with its areas of comparative advantage: form- tinuing learning and feedback to deal with valid by recentralizing ing and analyzing policy, setting and monitoring complaints and problems. implementation standards, and implementing easy-to-monitor activities that are subject to large economies of Accountability issues and control—­ ­ scale and are skills-intensive. The following recurring issues associated with but this is risky the incomplete implementation of the Batho Performance monitoring and evaluation­ Pele principles (the short route to accountability) —­learning from mistakes emerge: A new Presidential Department of Performance • Participation is often diluted to mere com- Monitoring and Evaluation has been established munity consultations and treated as some- in South Africa. After an intensive study, the thing to check off on the official’s checklist government developed 12 critical outcome goals of “his” project. associated with measurable outputs that will • Little or no choice of providers by users, guide performance agreements with ministers combined with low or absent user fees and and interministerial coordination arrangements. copayments. Data collection is being improved to monitor • Underdeveloped or absent user supervi- outcomes and measurable output, with progress sory committees and power over service reviewed by the President bimonthly. Impact providers. evaluation is being strengthened to support • No systematic and institutionalized use of policymaking. client feedback mechanisms such as citizen Departmental monitoring and evaluation report cards. must also be reformed. All departments have • No systematic and institutionalized report- monitoring and evaluation units, but they are ing by service providers and local govern- almost always separate from the implementa- ments to the citizens about coverage and tion hierarchy to help ensure their indepen- quality of services and use of resources. dence. This separation often makes it difficult • Insufficient use of local languages in infor- to collect basic implementation data on inputs mation provision and feedback mechanisms. and outputs, let alone outcomes. Findings are • Ineffective complaints, redress, and sanc- often late, incomplete, and not automatically tioning/reward mechanisms. fed back into improving operations. Many de- Despite President Zuma’s considerable ef- partments know little about which policies forts to strengthen the long route to accountabil- and programs are working, which are not, and ity, the underdevelopment of the short route is why. Responses to failed programs often focus further aggravated by the following problems in on establishing new programs and subdepart- the long route: ments, worsening coordination and increasing • Compacts focused on inputs and outputs, fragmentation. rather than on outcomes. International businesses and other organiza- • Services organized as centrally driven sec- tions have incorporated “learning by doing” or tor silos, which are difficult to coordinate, as “learning from failure” approaches. Quality con- demonstrated internationally and in South trol strategies now used widely in the automotive Africa. industry empower all workers to stop the assem- • Complex procedures, some still stemming bly line when they detect a mistake, allowing from apartheid, developed in central offices expert teams to resolve the problem before pro- rather than in the field through participa- duction resumes. No such problem solving teams tory methods, and after field testing, prior to exist in South African development. The Presi- adoption. dency has instituted a hotline where citizens • Lack of practical and comprehensive opera- can complain about services, and complaints are tion manuals for implementation.

6 Accountability in Public Services in South Africa • Inadequate monitoring, evaluation, feed- South African government. While this may services are back, and learning processes, which treat seem daunting and overly ambitious, the task organized typically monitoring and evaluation as ancillary rather does not involve introducing any new legal or as supply-driven than integral to service delivery. policy frameworks. Needed is consistent and • Financial management and information sys- effective implementation of the existing, well- silos delivering tems not integrated with the data on inputs designed constitutional decentralization and the outputs without and outputs, resulting in dispersed, inac- fiscal and sector policy frameworks already in citizen-user cessible, poor, and untimely basic data on place. services. Reforms are being led by the Presidency participation • Underdeveloped feedback mechanisms with with support from the Department of Coopera- no systematic approach for correcting errors tive Governance and Traditional Affairs and the quickly. National Treasury. This study concludes with a • Lack of follow-up and sanctioning­—­both proposed list of important components and roles exacerbated by poor data­—­when manager to make service providers as accountable to citi- performance is unsatisfactory, including in zens as they are to politicians/policymakers. financial management. Typically, when a new program is initiated, • Incomplete decentralization due to a lack central units of national departments formulate of consistent application of the subsidiarity implementation “policies” without involving the principle. operations managers, staff, or end-users. Once • Overlapping mandates and responsibilities the program is rolled out, these policies are often among the center, provinces, districts, mu- not formalized into operation manuals, which nicipalities, and service providers, and a lack should be continually updated as implemen- of coproduction agreements that define the tation experience accumulates. Programs and functions and responsibilities of the actors. projects would benefit from placing more effort • Excessively complex Integrated Devel- on systematically building on implementation opment Plan process, interpreted differ- experiences. ently by politicians/policymakers than by Finally, services are organized typically as citizen-users. supply-driven silos delivering outputs without • Municipalities prefering to deliver services citizen-user participation. Successful service themselves rather than through public-pri- outcomes depend on the government working vate or other partnerships, due to the oner- with citizens to design, implement, and evalu- ous requirements for contracting service ate services­—­with each party holding the other provision. accountable. Progress will come not from add- ing more resources and more delivery capacity Effecting change but from creating systems that give incentives to Multiple breakdowns in both the short and long all parties­—­politicians, service providers, and routes to accountability mean that reforms must citizens­—­to achieve better outcomes and con- be broader than those initiated by the current tinually learn from experience.

Accountability in Public Services in South Africa 7

Introduction

What analytical framework could assist us in communication between citizens and politicians thinking through successes and failures in public beyond the electoral process. When citizens service for the poor? The 2004 World Develop- express their priorities and preferences for par- ment Report used an approach that looked at the ticular policies to politicians, we call the politi- politician and service provider incentives to pro- cians “policymakers” receiving “policy feedback.” vide quality services to citizens, including the: Policymakers include ministry staffs, provincial • “Political voice” that citizens have with poli- and local governments, and regulators. For sim- ticians to shape priorities and policy design. plicity, we assume that together, politicians and • “Compacts” and “contracts” between policy- policymakers constitute the central state and are makers and service providers to provide ser- responsible for setting policies and allocating vices effectively and efficiently. budgets. • “Client power” of citizens over service pro- Policymakers reach agreements with service viders to ensure delivery. providers­—­public, private, or nongovernmental­ The goal of policy should be to create a sys- —­to deliver the services citizens demand. This tem that gets these accountability relationships Figure 1 Service provider institutions: the short and long routes to right. Getting it right does not mean a one-size- accountability and the three relationships fits-all approach: local context, history, and is- sues specific to each institution affect reforms. The state A successful system provides opportunities for Politicians Policymakers learning, rewards good outcomes while learning from poor ones, and adapts as it progresses. Compacts Contracts Long route Accountability relationships in public Citizen voice services: conceptual framework Citizens/clients Providers See the triangle of relationships in figure ­—1 ­ the accountability relationships between citi- Coalitions/inclusion Client power Management zens, politicians/policymakers, and service Nonpoor Poor Frontline Organizations Short route providers, and two routes to accountability: a long one and a short one. The political process Services through which citizens try to influence politi- cians is called “voice.” This is the long route’s Source: World Bank 2003. first “leg,” which represents both voting and

Accountability in Public Services in South Africa 9 The triangle of second leg of the long route to accountability community-driven development will affect all relationships in involves the relationship between politicians/ three legs of accountability. public services is policymakers and the frontline service provid- The accountability model covers several ers. This leg is about the “vertical” account- functions: delegating, financing, performing, more complicated ability relationships in the state to deliver pub- informing, and enforcing (figure 2). than a market lic services. If this agreement is solely between • Delegating implies that responsibility for the transaction agents in the public service, we call it a compact. delivery of a service is assigned by one actor If the government hires a private or nongovern- to another, who will be held accountable between a buyer mental provider to deliver the services, we call for it. and a seller the agreement a contract. The compact relates • Financing is the payment for the costs of de- to state capacity; think of this as the “command livering a service. and control” part of service delivery within the • Performing is the actual supply of a service. government bureaucracy (but including govern- • Informing is gathering and communicating ment outsourcing of certain activities to the pri- relevant information on whether and how vate sector). well a service has been delivered. The short route­—­the third leg­—­is about how • Enforcing is the sanctioning mechanisms much influence citizens, now acting in their role (positive and negative) to reward good and as the clients of public services, can directly exert bad performance. on service providers to ensure efficient delivery; These functions are equally useful in think of this as client power. The two routes are: analyzing simple market transactions, such as • Long route to accountability: buying a kilogram of mealie meal, or complex • Leg 1 is about citizen voice­—­political actions, such as running a municipal democ- accountability. racy. Asking for the mealie meal is delega- • Leg 2 is about the compact­—­state capac- tion, paying for it is finance, selling it is per- ity and accountability. formance, eating it generates information, and • Short route to accountability, leg 3, is about enforcing it comes from the buyer either going client power. back to the same seller for more or getting it The long route represents how accountability from another seller. Similarly, citizens choosing usually manifests in the modern state: citizens a local politician to manage their municipality elect and influence their political representa- is delegation, paying taxes allowing politicians tives, who in turn form and lead a government to make budget decisions is finance, and service to deliver a variety of public services. If it does so providers producing services is performance. poorly, the citizens will vote the politicians out These services are assessed by citizens based on of power, so in theory politicians have a strong their experience and information, and citizens incentive to hold public servants accountable for ultimately act to control the local politician services. Given this feedback mechanism, ser- politically or legally, which is enforceability vices are bound to continually improve. (Reinikka 2007). The short route focuses on the direct inter- The triangle of relationships in public ser- action between citizens and service providers. vices is more complicated than a market transac- When services are delivered, citizens should be tion between a buyer and a seller. Take educa- able to exercise various degrees of influence over tion: a child does not “buy” education directly the service providers­—­we call this “client power. from a public school teacher, and if the child While analytical separation of accountability feels service is bad, he or she typically cannot into the long route and the short route is helpful, change teachers. Similarly, families do not di- in practice there will be cases in which a spe- rectly affect the budget allocation to education cific mechanism­—­such as participatory develop- through their own purchasing power. Without ment planning and budgeting­—­can strengthen relevant information, benchmarks, and compar- both the long and short routes. Policymakers ators, families have little on which to base their will see it as an expression of voice, but citi- opinions on the quality of schooling. zens will try to use it to increase their power Whether a teacher does a good job depends over providers. Similarly, decentralization and on complex, interrelated factors: a political

10 Accountability in Public Services in South Africa Figure system that influences policies and budgets; an 2 Accountability relationships and functions administrative system that assigns roles and re- sponsibilities, allocates the budget to schools, establishes curricula, and defines salary scales Delegating and bonuses, among other things; and a gover- Actors Financing nance system at the school level, where the in- Accountable actors (principals) volvement of some parents in school boards as- (agents) including including clients, Performing sists teachers in finding the best way to impart policymakers and citizens, and knowledge to pupils. providers policymakers Informing The analytical framework in this report looks at each of these categories of interactions Enforcing and tries to identify what incentives and other accountability mechanisms the various actors Source: World Bank 2003. in the public service systems face. A grow- ing body of research demonstrates that “get- ting these accountability relationships right” The report reviews the international and is critical to improving public services to the South African experiences with the various poor. Getting it right means that account- mechanisms to strengthen service accountabil- ability relationships are all working properly, ity. South Africa’s challenges are similar to other reinforcing each other as a public service de- countries, but others have typically had more livery system. This methodology emphasizes time to devise solutions, whereas South Africa strengthening both the short and long routes is still young. South Africa has a progressive, to accountability—­ strengthening­ democracy, innovative government that has already learned internal state capacity and accountability, and substantial lessons about what works and what citizen participation (World Bank 2003; Fisz- does not. Though South Africa is unique, much bein 2005). can be learned from international experience.

Accountability in Public Services in South Africa 11

1 Progress in Service Coverage and Quality

This chapter describes the broad trends in South equitable share linked to population and poverty African public service provision and quality. and through conditional grants. In addition, real The overall assessment, shared by a wide range estate property taxes provide a source of locally of researchers and observers, is that the South collected own revenues for municipalities, and African government has a pro-poor approach to subnational sovereign borrowing is an option for governance. creditworthy municipalities. But South Africa faces many challenges. Setting the stage Progress in improving poverty and income dis- South Africa has an excellent Constitution that tribution indicators over the past decade has enshrines universal rights to basic services. Since been mixed. The percentage of people living the fall of apartheid, the country has become a below the poverty line rose from 28 to 33 per- well-functioning democracy that has increased cent over 1995–2000,1 and inequality also in- public spending on basic public services tar- creased.2 Poverty fell over 2000–2004, though geted to the previously disadvantaged majority. researchers disagree on both the extent of the The economy recovered from slow growth before drop and how much the rapid increase in social 1994 and, until the global economic crisis, had grant spending contributed.3 been growing at about 4 to 5 percent annually. The “Second Economy Strategy Process” Exemplary fiscal management and revenue col- (TIPS 2009) led to a Second Economy Strategy lection led to a dramatic fiscal turnaround, approved by the Cabinet in 2009. The strategy reducing public debt. The country is now reap- uses the terms “first” and “second” economy to ing the benefits and has been expanding pub- describe the conditions still prevailing in South lic spending from the country’s own savings Africa, with wealth and resources concentrated rather than borrowing. The strong fiscal posi- in the first economy and poverty and disadvan- tion has also enabled South Africa to withstand tage concentrated in the second. Certain legacies the recent global economic crisis far better than of apartheid make this inequality deeply struc- many Organisation for Economic Co-operation tural, particularly the: and Development countries. • Centralized monopoly structure of South A sound decentralization framework has Africa’s core economy, including legacy. been established in South Africa, with an ad- • Labor market laws and the highly skewed vanced intergovernmental fiscal framework. It distribution of assets such as land and capital. allocates funds from the national to the pro- • Spatial legacy of Bantustans and apartheid vincial spheres of government, based on an cities.

Accountability in Public Services in South Africa 13 • Legacies of deep inequality in the develop- huge employer of unskilled and semiskilled ment of human capital. workers in other contexts­—­have barely increased The Second Economy Strategy identifies because the currency exchange rate is not com- these structural issues as the main causes of pov- petitive due to the commodity boom, high inter- erty, alienation, and mixed economic and fiscal est rates, and low savings. performance, which has not reduced the huge The Second Economy Strategy is thus pes- structural unemployment problem. simistic about the performance of the core South In 2007, unemployment had settled at about African economy as well as the second econ- 25 percent, based on data for people having ac- omy’s ability to generate jobs. So it focuses on tively looked for work in the last week, or slightly human capital, the expansion of the social wage less than 40 percent if including all people able (grants and basic services), and a significant ex- and wanting a job. The economic structure in pansion of public employment, including the South Africa, inherited from apartheid, does new Community Works Program that is both not absorb large numbers of unskilled workers. participatory and decentralized (box 1.1). Manufacturing, mining, and agriculture have Finally, important demographic phenomena, all shed unskilled workers, while the labor force such as the “double whammy” of population in- has increased dramatically as new entrants­—­ creases and smaller households, pose challenges typically young African women­—­have entered for service coverage. The South African popu- the marketplace. Stats SA estimates that 5 to 6 lation increased 20 percent over 1996–2007, or million people joined the workforce over 1995– from 40.5 million to 48.5 million, and house- 2005. Labor-intensive manufactured exports­—­a hold size fell from 4.48 people to 3.88. Hence, the number of households increased 38 percent. Box 1.1 The Community Works Program Similarly, urbanization has added many The Community Works Program (CWP) uses many of the implementation and ac- households to towns and cities. According to countability mechanisms advocated in this study: decentralization, community par- the United Nations Population Division, South ticipation, contracting, and accountability mechanisms to beneficiaries in the com- Africa’s urban population accounted for about munities it serves. It embodies approaches and lessons from India’s National Rural 60 percent of the total population in 2005 (up Employment Guarantee Program. For each site, the program guarantees regular, from 55 percent in 2000), growing almost 3 per- predictable employment, such as two days per week, to about 1,000 unemployed cent a year over 1970–1990 and almost 2.5 per- or underemployed people of working age­—­people that do not receive social trans- cent a year since 1990. Cities have been unable fers. The CWP uses community participation to identify useful work and priorities in the sites or communities where it operates. It requires formal government and to add the necessary infrastructure capacity to uses ward committees, thereby avoiding the creation of parallel systems, which deal with this influx. strengthens local government. But centrally appointed implementation agents are employed to avoid placing extra burden on local government. These agents have Challenges to public services to account locally for the usefulness of the work and the quality of the assets and What does this mean for public services? South services delivered. Africa’s development model has focused on reor- In practice, some common priorities have anchored the programs: ganizing the state to move from an undemo- • A strong focus on food security. cratic state delivering quality services to the • Home-based care, mainly for HIV- and TB-affected households. • Care of orphans and vulnerable children in households headed by children. white minority to a democratic state delivering • Social programs to tackle alcohol abuse, violence, and crime. the same standard to the black majority. • Development of recreation spaces and sporting facilities for youth. First, the development model for South Af- • Environmental rehabilitation, maintenance, and upgrading of informal rica has sought to reverse the authoritarian and settlements. nontransparent workings of the apartheid state • Zibambele-type road maintenance. machinery by adopting mechanisms that intro- After a period of intensive piloting, the program was mainstreamed as a Coop- duce and strengthen “procedural” democracy. erative Governance and Traditional Affairs program in April 2010. It has reached at For instance, citizen participation and consul- least 55,582 people, having developed a scaling strategy that fans out from exist- ing worksites to cover entire municipalities. CWP hopes to reach the most marginal tations with business and labor­—­often together municipalities first and all municipalities soon after. referred to as “stakeholder participation”­—­are formally required in drafting legislation and in Source: TIPS 2009; Philip 2009. all subsequent policy formulation and review. Also, several independent “watchdog” agencies

14 Accountability in Public Services in South Africa accountable to Parliament were established by of the poorest quintile is attributable to state the challenge is to the Constitution to provide checks and bal- transfers” (PCAS 2003, p. 18). ensure that public ances on the executive, including the Pub- Fortunately, due to the fiscal space created transfers result in lic Service Commission, the Human Rights and the successful targeting of fiscal resources Commission, and the Financial and Fiscal to the poor, South Africa finds itself in an envi- better outcomes­ Commission. able position: fiscal resources have not been the Second, to ensure the transformation in binding constraint to improved service coverage coverage, the development model enacted con- and quality. Instead, the challenge is to ensure stitutional guarantees for basic services. The that public transfers result in better outcomes­ Constitution of the Republic of South Africa —­current outcomes are disappointing. The best (1996), in its Bill of Rights (chapter 2), defines measured examples come from the education the following rights: access to adequate housing, sector, where South Africa participates in several healthcare services, sufficient food and water, international education evaluations.4 South Af- social security, and education. It dictates that the rican student performance in math, science, and state must take reasonable measures, legislative reading is disappointing compared with most and otherwise, within its available resources, to countries, except a few very poor countries in progress toward realization of these rights. Africa. Results from South Africa’s Department To realize these rights, enormous energy of Education evaluations confirm this underper- first had to be applied to merging the previously formance. And despite great shifts of resources separate “white” and “black” administrative sys- to poorer schools since the transition to democ- tems. This task, often underappreciated, has racy, inequalities in education outcomes do not now been completed. A highly fragmented and appear to have narrowed. unrepresentative system has now been consoli- Fourth, the development model must ad- dated into a single representative state, includ- dress the double whammy: population increases ing a public service that reflects the demography have been accompanied by smaller households, of the country. But this process “has also led to a problem that has created service backlogs and the loss of experience and institutional memory” substantially increased the challenge of improv- (PCAS 2003, p. 106). And the massive logis- ing service coverage and quality. Declines in tical exercise of amalgamating the previously fertility associated with higher incomes do not white and black bureaucracies shifted emphasis alone explain the smaller households. The rea- to organizational matters at the expense of more sons are more complex: the Presidency, for in- important institutional matters­—­the new “rules stance, suggests that an element of opportunistic of the game” of a developmental state. The lat- household behavior is involved as families “un- ter is a more difficult problem, particularly for bundle” into several households, anticipating a state that inherited apartheid institutions, be- that doing so would increase their probability of cause “institutions are not machines that can be accessing services (PCAS 2003). reprogrammed by a new ‘organogram’” (Pieterse Fifth, the South African development model and others 2008, p. 8). has accorded a leading role to the state and to The main challenge now is to ensure that this state-owned companies­—­and will continue to strong state and its strong procedural democracy do so. No leading roles in public services are en- achieve a more fundamental transformation to- visaged for civil society organizations and the ward a state that is responsive and accountable private sector. The focus on formal procedural to all its citizens, including those who have yet democracy driven by a strong state has, perhaps to benefit from its public services (Pieterse and unintentionally, reduced the role of communities others 2008). and civil society in development, with few ex- Third, to reduce poverty and undo the in- ceptions, such as the Treatment Action Group equality and high unemployment inherited from (box 1.2). apartheid, South Africa’s development model Sixth, HIV/AIDS puts enormous strains has made redistribution­—­transfers to poor on both the supply of, and the demand for, ser- households­—­its primary priority. According to vices.5 South Africa’s development model must the government, “fully two-thirds of the income target this problem. For instance, in 2005,

Accountability in Public Services in South Africa 15 Box 1.2 The Treatment Action Group Framework­—­has created the fiscal space to sub- The Treatment Action Group (TAG) has successfully challenged the lackluster im- stantially increase pro-poor spending on vital plementation of the government’s HIV/AIDS strategy, with a particular emphasis public services such as education, health, and so- on treatment and the prevention of mother-to-child transmission. It has done so cial grants. Further, analyzing expenditures on through advocacy, demonstrations, and court challenges of particular government such basic services shows that the government approaches by, for example, forcing itself into leg 1 of the accountability frame- has been effective in increasing pro-poor orien- work between citizen-users and politicians. Its success has led to a revamping of tation of social spending. Resource shifts have the HIV/AIDS strategy and programs in which TAG participation was encouraged by then Vice President Zuma. In addition, it has inserted itself into the third leg been large and rapid so that few areas are left of accountability, between citizen-users and service providers, by participating in in which the poorer deciles are not obtaining at HIV/AIDS treatment programs. It has even developed treatment programs of its own least their share of public expenditure. for particularly disadvantaged groups­—­a rare example of community-driven devel- In 1996, the South African government con- opment. In these ways, the TAG has helped lay the foundations for the turnaround stitutionally enshrined the right for everyone to HIV/AIDS strategy of the current government. have access to housing, health care, food, water, social security, and education. Since passing the Constitution, the government has significantly 10 percent of public servants may have been increased nominal expenditures on basic pub- affected by HIV/AIDS, a quarter of whom lic services. On education alone, public spend- could die by 2012 (PSC 2008a). In 2004, the ing reached ZAR148 billion (US$19 billion), HIV/AIDS prevalence rate among educators in or 6.1 percent of GDP, in 2009/10. Welfare public schools was 12.7 percent (PSC 2008a). transfers amounted to about ZAR118 billion, Children may drop out of school because they while spending on free or low-income housing have to take care of the sick or become head of amounted to ZAR17.5 billion. an orphaned household. And clean water and Even in real terms, the increases have been proper sanitation become vital­—­literally. This substantial. Measured in 2005 prices, education pandemic affects service delivery in many ways. was allocated ZAR109.7 billion in 2009/10, a 55 percent real increase since 1995/96. Over the Public expenditures on basic services same period, health spending grew 110 percent So how has this democratic, pro-poor, and state- in real terms, social welfare spending grew 186 led model, with its strong constitutional guaran- percent, and public order and safety spending tees for basic services, allocated the fiscal space grew 95 percent (figure 1.1). Real expenditure created by sound fiscal management? on housing and community services increased Exemplary fiscal management­—­based on 355 percent over the same 14-year period. a strong revenue effort and a three-year rolling The GDP share of allocations has increased and results-based Medium Term Expenditure for each major social sector spending category since apartheid (figure 1.2). The share of re- Figure 1.1 Real expenditure on basic public services, 1995/96–2009/10 (2005 prices) sources devoted to education as a percentage of total GDP declined initially, falling from 6.1 ZAR (billions) percent in 1995/96 to 5.1 percent in 2007/08, 120 Education before recovering strongly to 6.1 percent by 2009/10. Health spending increased from 2.9 100 Social security and welfare percent of GDP in 2001/02 to 3.9 percent in 80 2009/10, recovering to the levels seen in 1996– 1998. Resources allocated to housing and com- 60 Public order and safety Health munity services have also grown sharply, from

40 0.9 percent of GDP in 1995/96 to 2.7 percent in Housing and 2009/10. community services 20 Social security and welfare saw its share in GDP rise from 2.6 percent in 1995/96 to almost 0 1995/96 1997/98 1999/2000 2001/02 2003/04 2005/06 2007/08 2009/10 5 percent in 2009/10. Over the same period, re-

Source: Own calculations from Ministry of Finance data. sources allocated to public order and safety rose from 2.5 percent of GDP to 3.2 percent.

16 Accountability in Public Services in South Africa Figure 1.2 Expenditure on basic public services as a share of GDP, The composition of public spending has 1995/96–2009/10 also changed in favor of the social sector (figure

1.3). Defense spending, which absorbed about Percent 15 percent of the total budget in 1983/84 and 7 8 percent in 1995/96, makes up only 4 percent Education 6 today. In figure 1.3, defense spending has been grouped with expenditures on mining, manu- 5 Social security facturing, construction, agriculture, and fishing. 4 and welfare Health Collectively, these other sectors saw their share of the total government budget decline from 3 Public order and safety about 7 percent in the early 1980s to 2.2 percent 2 Housing and in 2009/10. community services The share of the budget spent on interest 1 payments has also fallen significantly, from 19.8 0 1995/96 1997/98 1999/2000 2001/02 2003/04 2005/06 2007/08 2009/10 percent in 1995/96 to 7.3 percent in 2009/10, further generating the fiscal space needed to Source: Own calculations from Ministry of Finance data. ramp up social sector spending. The prudent fis- cal policy adopted by the successive democrati- Figure 1.3 Major expenditure areas as a share of total budget, 1995/96–2009/10 cally elected governments, which sought to re- verse the unsustainable stance of apartheid, has Percent helped create this fiscal space. 60 With this added fiscal space, and to its Health, education, housing, and social security (Rural Housing Service) credit, the government has increased the share 50 of the public budget dedicated to basic public 40 services (health, education, housing and com- munity services, and social transfers) from 44.8 30 percent in 1995/96 to 48.1 percent in 2009/10. Interest on debt 20 Public order Interestingly, there has been a steady increase in and safety the share of this broad spending category since 10 Defense and primary sectors 2001/02, following a noticeable inverse-V pat- tern before that. 0 Housing and community services 1995/96 1997/98 1999/2000 2001/02 2003/04 2005/06 2007/08 2009/10 For basic public services, social security and welfare saw the biggest jump in its share of the Source: Own calculations from Ministry of Finance data. total budget, increasing from 9.5 percent in 1995/96 to 15 percent in 2009/10 (figure 1.4). Figure 1.4 Expenditure on basic public services as a share of total budget, 1995/96–2009/10 Education spending fell from 22.1 percent in 1995/96 (its peak level since the 1980s) to 18.8 Percent percent in 2009/10, while health spending rose 25 from 10.4 percent to 12 percent. Housing and community welfare spending saw a two-and- 20 Education half-fold increase, from 3.4 percent to 8.8 per- Social security cent. The share of social security and welfare 15 and welfare grew more than 50 percent over this period, Health from 9.5 percent to 15 percent. 10 South Africa’s real GDP has grown signifi- Housing and cantly, but so has its population, increasing the 5 community services scale and complexity of providing public ser- vices. Thus, it is of interest to separately con- 0 sider real per capita expenditure on basic pub- 1995/96 1997/98 1999/2000 2001/02 2003/04 2005/06 2007/08 2009/10 lic services, trends that are significantly more Source: Own calculations from Ministry of Finance data. muted (figure 1.5). Real per capita expenditure

Accountability in Public Services in South Africa 17 Figure 1.5 Per capita expenditure on basic public services, 1995/96–2009/10 basic public services has been surprisingly well targeted post-apartheid (Van der Berg 2006a). A Constant 2005 ZAR large proportion of the poor now receives grants 2,500 through various social security and welfare pro- Education grams. Studies show that spending on health 2,000 and education, the two other major social pro- grams, are also reasonably well targeted to the 1,500 Health poorer segments of the population. The poorest 40 percent of the population receives more than 1,000 their share of overall social spending on these

Public order and safety Social security items. 500 and welfare Housing and Expenditure on social security and welfare community services programs­—­the budget item that has increased 0 1995/96 1997/98 1999/2000 2001/02 2003/04 2005/06 2007/08 2009/10 the most in the past decade­—­appears especially well targeted. A means test has ensured target- Source: Own calculations from Ministry of Finance data. ing of all the grants, including social old-age pensions, disability grants, and child support on education increased 23 percent over 1995/96– grants. Some errors of inclusion have occurred 2009/10, health spending increased 67 percent, (ineligible, nonpoor people receiving grants), but and spending on housing and community ser- errors of exclusion (eligible, poor people not re- vices increased 12 percent. Real per capita ex- ceiving grants) seem fairly minor. penditure on public order and safety also in- Since the poorer segments of the population creased significantly: in 1995/96 it was ZAR761 have more children, they now obtain, relative to but in 2009/10 it was ZAR1,172, an increase of their population size, slightly more than their 55 percent, with the bulk of the increase since share of public spending on primary and second- 2006. ary education. In addition, pupil-teacher ratios Social security and welfare transfers have have largely been equalized. Note, however, that seen an explosive increase: in 1995/96, the aver- urban and affluent schools still retain more -ex age South African received about ZAR279 per pensive, better qualified, and more experienced year (in 2005 prices) in social transfers, which teachers, and thus still receive more public fiscal had increased almost four-fold to ZAR1,003 resources per school child (Crouch 1996; Gus- by 2009/10. While some studies (for example, tafsson and Patel 2006; Motala 2006; Van der Ardington and others 2005) have shown an in- Berg 2006b; Wildeman 2004). crease in poverty and inequality into the late The major exception in education spending 1990s, Aguero, Carter, and May (2008) argue appears to be in tertiary education, where the that these trends reversed partially after 1998. poor benefit far less than proportionate to their Van der Berg (2005) argues that this rever- numbers. This could reflect three factors: sal is explained by social security and welfare • Pro-poor spending might be underestimated transfers that have reduced inequality and the as surveys typically do not capture tertiary spread of the expenditure distribution by lift- students living in their parental home, and ing up the lowest part of the distribution. For students from poor households may be re- instance, both the Child Support Grant (which corded as belonging to single but richer has substantially increased the number of grant households. receivers and the resources of poor households, • Poor students may be constrained from ad- particularly in rural areas) and greater access to vancing to tertiary education because of the services are likely to have improved other aspects weak quality of education in poor schools of well-being. and the poor socioeconomic background of students. Targeting of government spending • Access to tertiary education is more con- How well have these expenditures been tar- strained by financial factors than is primary geted? It turns out that public expenditure on education.

18 Accountability in Public Services in South Africa In health, resource shifts have increased ac- the rich and poor remain large. Several stud- poor outcomes cess to primary healthcare services (mainly clin- ies show that improved targeting of spending in public service ics), even in deep rural areas (Barron and oth- for the poor for both health and education has delivery cannot ers 2006; Blecher and Harrison 2006; Burger delivered only limited benefits in terms of out- and Swanepoel 2006; Collins and others 2000; comes. (For health, see Burger 2007; Burger and be explained by Grobler and Stuart 2007; McIntyre, Muirhead, Swanepoel 2006; Grobler and Stuart 2007. For inadequate fiscal and Gilson 2002). According to the 2006 Gen- education, see Patel and Crouch 2008; Gustafs- allocations eral Household Survey, 83 percent of respon- son and Patel 2006; Van der Berg 2005, 2006a, dents who had suffered an illness or injury in the 2006b; Van der Berg and Burger 2002, 2005.) preceding month had consulted a health worker. Ajama and Aron (2008) confirm this finding, Distance (as a proxy for access) was only the stressing that poor outcomes in public service third most prevalent reason for not consulting delivery cannot be explained by inadequate fis- a health worker: only 8.6 percent of those who cal allocations. did not consult a health worker cited distance as Then what could explain this? We will turn their reason for not doing so. This means that first to a number of authoritative assessments of only 1.5 percent of those not ill or injured did state performance, relating particularly to service not consult a health worker because of distance. delivery. We then delve deeper to suggest plau- But perceptions about low-quality public sible explanations for the inadequate outcomes. health care have resulted in health care largely reflecting the characteristics of an “inferior Assessments of services good” (Palmer 1999; Palmer and others 2002). What are the most important assessments of This is also reflected in our participatory assess- public services in South Africa? What do they ment of service delivery in six municipalities say about overall performance? We have selected (Khanya-aicdd 2008), where health received the two main assessments to discuss in this sec- lowest scores for access and the most complaints. tion: a 2003 assessment led by The Presidency Long waiting times were extremely prevalent in of South Africa, Towards A Ten Year Review, and public health care: 43 percent of patients visiting the 2007 assessment made by the African Union public clinics in the 2006 General Household under its African Peer Review Mechanism. The Survey complained of long waiting times. In results of other assessments, such as Van Donk addition, 15 percent found the supply of drugs and others (2008), will be referenced throughout lacking, and 9 percent complained about rude the report. staff. Thus, especially for more serious ailments, Towards a Ten Year Review even the poor “vote with their feet” and use Concluded in 2003, Towards a Ten Year Review more expensive private health services. Not sur- was an internal review of the government’s per- prisingly, more affluent members of society are formance, overseen by a steering group of min- even more likely to pay for private health ser- isters. The study commissioned 35 overview vices. The concentration curve for public health and 25 “gap analysis” papers from officials and services reflects that the more affluent benefit researchers. Overall, the Presidency concludes little from such services while the poor benefit that the government is making progress on its more than proportionally. The Department of stated objectives, which it reaffirms are the cor- Health (2007) also acknowledges the quality rect ones. The objectives originate in the Recon- imperative. struction and Development Programme and In conclusion, public spending on vital pub- include meeting basic needs, building the econ- lic services is substantial, and in most instances omy, democratizing the state and society, devel- public spending is growing and generally well oping human resources, and building the nation targeted to the poor. But does this translate to (PCAS 2003). good outcomes? Unfortunately, in many areas But the report also concludes that weak- increased spending has not translated to suffi- nesses reside “at the coalface of interaction with ciently improved outcomes to meet the expec- the public and in the supervision and manage- tations of the population­—­the gaps between ment of implementation” (PCAS 2003, p. 102).

Accountability in Public Services in South Africa 19 education The report warns that continuing on the current Subjective assessments complete the picture. concerns persist, trajectory would “. . . precipitate a vicious cycle of Based on The Institute for Democracy in South and opinions decline in all spheres” (PCAS 2003, p. 102). The Africa’s Afrobarometer (covering 1994–2002) findings above are consistent with the findings and Markinor (1996 and 2003), a consistent pic- on health care in the following chapters, including those of the ture of improving government performance in are consistently participatory assessment in six municipalities. South Africa emerges (PCAS 2003). But edu- negative and Towards a Ten Year Review used a number cation concerns persist, and opinions on health of composite outcome indices to measure how care are consistently negative and getting worse. getting worse much the “average” province in South Africa has Public approval for housing, however, improved caught up with the “best” province since 1994. substantially (PCAS 2003). The best province was rated 100 and the worst Towards a Ten Year Review reaffirms the role province rated 0. (Note that these indices do not of the state, saying that successes were often measure progress in absolute levels since 1994.) achieved in areas where government had sig- From 1995 to 2002, the: nificant control, though acknowledging that this • Infrastructure Services Index—­ measuring­ finding is not consistent (PCAS 2003). It pos- access to piped water, toilets, rubbish re- its that “the architecture of the new democratic moval, telecommunications, and electricity­ State is in place, yet in many areas of service de- —­increased from 46 to 60. livery the performance of the Public Service re- • Quality of Life Index—­ measuring­ access to quires much improvement” (PCAS 2003, p. 75). medical facilities, adult functional literacy, And it identifies an area of contestation of roles use of electrical stoves as opposed to the en- and “competing” legitimacy claims between vironmentally bad use of wood and coal fires­ elected bodies, nongovernmental organizations, —­increased from 52 to 67. and community-based organizations (PCAS • Political Participation Index—­ measuring­ 2003, p. 76). voter turnout for local government elections, The report also acknowledges that govern- a political approval rating from Afrobarom- ment coordination could have been better, and eter, and union participation among formally that less progress was made in areas that require employed workers­—­increased from 37 to 47. partnership with others­—­the private sector and • Social Inclusion Index—­ measuring­ the pro- civil society, for example. By contrast, it notes portion of households in which the house- the positive use of faith-based organizations to hold head has a partner, home ownership, increase the number of child support grant re- and participation in cultural organizations­—­ cipients (PCAS 2003). increased from 27 to 36. The report states that “involving social part- • Safety and Security Index­—­measured as seri- ners and the broader civil society must go fur- ous crime, court prosecution, and resolution ther than the mere articulation of visions and rates­—­decreased from 53 to 49. frameworks. . .” (PCAS 2003, p. 105). It calls • Economic Participation Index­—­which uses for a social compact, not just to better integrate the narrow definition of unemployment government programs into society but also to and average earnings received by workers­ harness the existing capacity in civil society —­decreased from 63 to 60 (PCAS 2003, for national development (PCAS 2003). It also pp. 88–89).6 calls for greater participation and interaction of The report also records changes in inequal- people with the state. From ward committees ity as measured by the Gini coefficient between to the Integrated Development Plan process to 1997 and 2000, the only years for which the workplace forums, actual participation has been relevant data were available. As mentioned, the limited mainly to special interest groups or hin- role of social transfers in reducing inequality is dered by short-term self-interest (PCAS 2003). substantial. Before transfers, the Gini coefficient The policy recommendation is “to improve ac- was between 0.68 to 0.59; after social transfers, countability to, and contact with, the electorate it decreased from 0.44 to 0.35 (PCAS 2003)­—­a by all levels of government . . . [and to] develop comparative increase in wealth equality. and maintain partnerships with civil society with emphasis on practical programs” (PCAS

20 Accountability in Public Services in South Africa 2003, p. 114). This report emphasizes the same sustaining democracy and political governance Lack of capacity message. through bottom-up decision-making processes” is identified as a The government’s report also cites the suc- (APRM 2007, p. 8). The machinery should en- major constraint cessful recentralization of the social grants sys- sure grassroots, local participation in the design tem into a National Social Security Agency, and implementation of public services. on the quality suggesting that if capacities do not exist at lower This APRM recommendation closely mir- of governance spheres of government, and are unlikely to be rors the findings of our study. In general, the in all spheres of created, the government should consider “asym- APRM finds that public consultation, educa- metric allocation of responsibility where capac- tion, and feedback in policymaking and imple- government ity exists to undertake functions beyond the pro- mentation are inadequate and recommends im- visions of the Constitution . . .” (PCAS 2003, proving public participation in these processes p. 108). (APRM 2007). The APRM warns that the centralist, top- African Peer Review Mechanism down approach taken by the South African The African Peer Review Mechanism (APRM) government to deliver services and build local is a self-monitoring body created under the New capacity raises serious concerns. It suggests a Partnership for Africa’s Development of the more bottom-up approach that vests legal au- African Union. Open voluntarily to all African thority and local governance responsibilities in Union members, its goal is to ensure that “the local communities. As is the government’s stated policies and practices of participating coun- policy, the APRM recommends that local gov- tries conform to the values, principles, codes, ernments be progressively empowered to make and standards enshrined in the Declaration on democracy truly effective at the grassroots level Democracy, Political, Economic, and Corporate and engender development from below (APRM Governance” (APRM 2007, p. 1). The report 2007). This report takes up this theme in chap- lauds the unprecedented success achieved in ter 8. meeting the basic needs of the people since the Lack of capacity is identified as a major first democratic elections of 1994. It also praises constraint on the quality of governance in all the Constitution as the major achievement of spheres of government. The existence of three post-apartheid South Africa, formalizing a pro- spheres of government seems to prevent a dy- cess of contestation within a spirit of cooperative namic use of scarce skills across the spheres, be- government to resolve conflicts and a system of cause each sphere protects its own staff (APRM intergovernmental checks and balances. 2007). Further, a centrally driven initiative to But the APRM also notes imbalances and build capacity in poorly performing municipali- paradoxes on which the government should ties (Project Consolidate) may perpetuate, rather focus more effort, including the continued than resolve, the syndrome of dependency on high levels of poverty and inequality, the severe the center (APRM 2007; see annex 5 for a de- HIV/AIDS pandemic, the neglect of rural areas, scription of Project Consolidate). The skills con- and the asymmetry between legislation and im- straint is further aggravated by the large emigra- plementation, between rights and realities, and tion of skilled professionals; the APRM urges between policy formulation and putting such the government to encourage the return of these policies into operation (APRM 2007). This dis- professionals, especially to local government in connect between policy and implementation is municipalities where implementation capacity is also a recurrent theme in this report. weakest (APRM 2007). The APRM also worries about the extent Increases in education spending are wel- to which democratic ideals are being put into come, but the results are poor: the system does practice, pointing out that ordinary South Af- not provide those leaving school with the skills ricans have insufficient contact with their politi- the economy needs. The APRM also highlights cal representatives. It urges the African National that more effort should be taken on land reform, Congress (ANC), as the dominant political arguing that it is the single most important fac- party, to “. . . put in place the necessary political tor to social, political, and economic empower- machinery that will nurture citizens capable of ment. It recommends that “. . . greater flexibility

Accountability in Public Services in South Africa 21 in land acquisition options and decentralized grievances; as many protests were recorded in community-driven initiatives” (APRM 2007, 2009 as in the previous five years combined, p. 22) be supported. Again, this is fully consis- with disturbingly high levels already recorded in tent with the findings in this study. early 2010 (figure 1.6). The APRM warns of the emergence of a It is inadequate to describe these merely as dependency syndrome: “There is also a need to service delivery protests, as the issues voiced address enduring attitudes of dependency on range far beyond service provision. Quot- government handouts. If not checked, creeping ing Martin Luther King Jr., Sicelo Shiceka­ dependency tends to put incremental pressure on —­the former minister of the Department of service delivery, thereby further straining capac- Cooperative Government and Traditional ity for service delivery, given the scarcity of re- Affairs, responsible for local government­—­has sources. Inability to deliver may lead to dissatis- described such protests as “the language of the faction and, ultimately, to disaffection with the unheard” (Shiceka 2010). It seems that the pro- governance system and the spirit of reconcilia- tests have disturbed the complacency of the de- tion” (APRM 2007, p. 79). partment, which has responded to these protests The APRM report underscored the issue with unprecedented vigor and with frank and of corruption, citing the results of stakeholder comprehensive assessment of local government consultations and survey data. For example, in a performance. Gauteng provincial survey, 63 percent of respon- dents agreed with the statement that “there is Notes lots of corruption in the Departments” (APRM 1. The informal poverty line is set at about 2007, p. 111). Finally, the APRM was alarmed ZAR300 per month. Among researchers, at the levels of reported violence against women serious debate about the quality of poverty and children. Again, the legal and policy frame- data in South Africa continues. See Policy works are excellent, but implementation lags. Co-ordination and Advisory Services, So- In conclusion, both the government’s assess- cial Sector, The Presidency, South Africa. ment and that of the APRM raise the gamut of 2006. “A Nation in the Making: A Discus- accountability issues outlined in the analytical sion Document on Macro-social Trends in framework of this report. South Africa.” 2. See Bhorat and Kanbur (2005). The rising tide of service delivery 3. See van der Berg and others (2005). Authors protests estimate a reduction in the number of peo- Protests soared in South Africa in 2009 in ple living in poverty by more than 3 million response to poor services, unresponsive coun- (from 18.5 million to 15.4 million). Also see cilors, corruption, ongoing job losses, and other Meth (2006). The author argues that a fall by 1.5 million is more likely. Figure 1.6 Social protests, 2004–2010 4. The international evaluations include the

120 Progress in International Reading Literacy Study (PIRLS), the Southern African Con- 100 sortium for Monitoring Educational Qual- ity (SACMEQ ), and the Trends in Inter- 80 national Mathematics and Science Study

60 (TIMSS), though South Africa opted not to participate in the 2007 round of TIMSS. 40 5. South Africa’s steady decline in the Human Development Index (a composite index of 20 education, health, and GDP per capita) is 0 caused entirely by the rapid decline in life 2004 2005 2006 2007 2008 2009 January–March 2010 expectancy due to HIV/AIDS. By 2010, it is

Source: Municipal IQ 2010. estimated that AIDS will have reduced life expectancy at birth in South Africa by 20

22 Accountability in Public Services in South Africa years: without AIDS, it would have been 68 that had the best measure of that compo- years, with AIDS, it will be 48, or perhaps nent in either of the two years received a even less (Seekings 2007). score of 1. . . . After scaling all the constitu- 6. These indices are not absolute performance ent parts of the index from 0 to 1, they were indicators but relative measures. “The scale then averaged to reveal an index value for the adopted was based on a comparison of pro- two periods in question.” This implies, for vincial rates of each index in the two time instance, that even if all provinces achieve periods. The worst ranked province at ei- lower scores from one year to the other, but ther of the two time-points received a score the distribution stays the same­—­the overall of zero for that component. The province performance index would not change.

Accountability in Public Services in South Africa 23

2

Citizen Voice

In the next three chapters this study addresses vaccinations versus curative health care). The international and South African experiences in sector-specific chapters will address these design improving service accountability. Examples of questions in greater detail. institutional choices and accountability mecha- Recall that the accountability framework nisms that have worked for specific services in has two routes. The long route covers the rela- specific contexts, in South Africa and in other tionship from citizens to politicians, from poli- countries, are provided. Also presented are ticians to service providers, and then back to hypotheses to be applied and tested in the fol- citizens. Services are provided through govern- lowing chapters. ment responsibility rather than market-based What are the key messages emerging transactions. from these chapters? First, the best results are Improving the long route has often been the achieved when all three legs of accountability main recommendation for delivering better ser- are improved. We hardly find countries that vices. Under democracy, citizens can hold politi- have improved services by focusing on just one cians accountable through voting and voice; in mechanism­—­elections, or a citizen report card, turn, politicians will hold service providers ac- for example. So we should guard against the countable through compacts and contracts, set- temptation to emphasize any particular tool and ting the appropriate incentives for improved ser- to ignore local context. vices. Only successful delivery to the majority of The second message is that accountability the population keeps politicians in power. Thus, mechanisms­—­as well as the merit of the long the accountability chain is internally consistent route to accountability versus the short route­—­ Citizen voice depend on the characteristics or “technology” of the service and the environment in which it is provided. Services differ radically in all Politicians/policymakers of these dimensions, so the appropriate level of decentralization and the specifics of the ac- The compact countability mechanisms for each service will also differ. Citizen voice There are no “silver bullets” and no universal solutions for designing accountability systems. Citizens/clients Client power Providers Systems must fit not only the sector in ques- tion but also the service provided (for example,

Accountability in Public Services in South Africa 25 The direct and should provide the appropriate incentives for good service delivery depends. Conversely, as relationship all the actors. democratic processes become more rooted in between citizens Unfortunately, reality often differs from the- society, these same innovations can become ef- ory. Take Latin America: a wave of democrati- fective tools for improving service delivery when and politicians/­ zation in the 1980s and 1990s transformed most adapted to local conditions. policymakers is of its countries from dictatorships to democra- Though investigating how to improve elec- the first leg of cies. In its wake, spending on basic public ser- tions is a valid area for research, this study will vices grew considerably and services expanded. only briefly summarize the views of several re- the long route to But in the final analysis, increased democracy searchers on the role of the electoral system in accountability and spending in Latin America led only to making services work for the poor. We focus mixed success. particularly on local government. The local political context may explain the limited success of reforms and spending. Most Political accountability at the local level Latin American countries share a history of in- How does political accountability work in prac- equality between an elite European descent and tice in local government, where local politi- the rest of the often indigenous population. Poli- cal representation should interact most directly tics are often deeply rooted in patron-client sys- with the citizens? (Note that many services are tems where the elite control access to the state delivered not by local government but by other and dispense favors to the poor. In many coun- spheres of government, so investigating political tries, these political characteristics were unaf- accountability at the local level cannot alone be fected by the change to democratic rule and sufficient.) continued to distort public service delivery, even Local government is designed to allow pub- after significant improvements to the long route lic participation, which will benefit development (Fiszbein 2005). and enhance democracy. Chapter 10 of the Con- This chapter describes how to improve the stitution of the Republic of South Africa (1996) first leg of the long route to accountability­—­ ensures that citizens have direct access to elected citizen voice. local government representatives and that they have the opportunity to participate in local gov- Elections ernance. The White Paper on Local Govern- The direct relationship between citizens and ment (DPLG 1998) mandates active citizen politicians/­policymakers is the first leg of the participation to ensure that ward councilors, bu- long route to accountability. Elections are per- reaucrats, and the community all form part of haps the most powerful of several mechanisms the local government in South Africa’s 284 mu- to hold politicians and policymakers account- nicipalities (GCIS 2005). able. But elections are periodic and rare; many The local government electoral system is de- other voice mechanisms are needed. Electoral signed to facilitate a direct link between coun- systems can be established quickly, but creating cilor and constituency. Voters have more than a truly democratic and inclusive society takes one vote: they vote first for a political party using time. a proportional representation system and second Some recent innovations in service deliv- for a ward councilor using the first-past-the-post ery, such as citizen report cards and participa- system (IEC 2008; box 2.1). The assumption is tory budgeting, can nudge society along on this that candidates elected through the proportional process (the “strategic incrementalism” recom- representation system will be accountable to mended by the World Development Report). their political parties, while ward councilors will But these innovations cannot compensate for be accountable directly to the electorate. the absence of fundamental democratic institu- In practice, ward councilors are faced with tions (for example, the separation of powers, its double accountability: to the electorate and to checks and balances, and the rule of law). Merely their political party. Some observers have ar- holding elections or introducing citizen report gued that accountability to the political party cards will not fundamentally alter the sociopo- wins out in most cases (McKay 2001). But the litical and institutional environment on which loyalty of the opposition party councilor cannot

26 Accountability in Public Services in South Africa Box be guaranteed. During the 2007 floor-crossing 2.1 Elections in South Africa window period in South Africa, 250 councilors South Africa holds two main types of elections: one for metro councils and one for crossed the floor, mainly to the African National local councils (which includes district council elections). Congress (ANC), affecting 128 municipalities (GCIS 2008). The African Peer Review Mecha- Metropolitan and local councils: Each voter receives two ballots. A first ballot for nism (APRM) criticizes this phenomenon as their ward will have the names of the ward candidates. The person receiving the most votes in a ward will win that seat. Ward candidates may represent parties or having the potential to undermine democracy be independent. Each voter will also receive a second ballot on which they can vote (APRM 2007). Finally, political parties use na- for a political party. This is the proportional representation ballot. The parties will tional manifestos to contest local government then be given seats according to the percentage of votes they receive in the whole elections, but without a local manifesto it is dif- metropolitan area. Each party has a list of candidates from which the councilors ficult to hold a political party accountable at the are drawn. local government level. Ward committees­—­made up of about 10 District councils: Each voter in a local municipality will vote for the district council people elected by ward residents and chaired by pertaining to their local area. The ballot contains the party names. Of the district council seats, 40 percent will be allocated according to the percentage of votes the ward councilor­—­should facilitate a link be- parties gained in the whole district municipal area­—­that is, through proportional tween councils and communities in between elec- representation. The remaining 60 percent will be allocated to the local councils in tions. Each council has its own regulations gov- that area. Each local council will be given a number of seats and must send coun- erning the election of members. Some observers cilors from their ranks to fill those seats. The seats should be filled according to the argue that these election procedures need to be support that parties have in a specific council. For example, if a local municipality legislated and more exacting. Currently, ward is given five seats on the district council and the African National Congress (ANC) committees are often elected in a public meeting gained 60 percent of the seats on the local council, the ANC councilors should fill without formal rules for a secret ballot or a quo- three of the five seats. The other two should be allocated to other parties according to the number of votes received. rum. Once elected, a ward committee is supposed The Constitution says that the council must reflect overall proportionality of all to make recommendations to the ward councilor votes cast, leading to a complicated method of allocating seats. The method for and council (Cloete and Thornhill 2005). doing this is set out in the Municipal Structures Act: a threshold of votes needed Interpretations within the South African to gain a seat is determined by adding the total number of votes cast in that elec- government of the status and role of ward com- tion for proportional representation and party ward candidates and then dividing mittees differ widely. Some argue that they are it by the number of seats on the council. All votes cast for a party on the propor- part of the system of democratic government­—­ tional representation ballot and for that party’s candidates on ward ballots are then legal entities accountable to all residents of the added. Each party’s total is then divided by the quota to see how many seats they are entitled to receive. The number of ward seats already won by that party is then ward. Others argue that they have no formal au- subtracted from the total number of seats allocated to the party. The remainder of tonomous powers, do not carry a mandate from seats the party is entitled to is then allocated to the same number of people on the a constituency, and are an advisory body insti- party’s proportional representation list. tutionally subordinated to ward councilors. Still others argue that ward committees should be Example: A total of 100,000 votes were cast in Makana Municipality, where there is seen as the “voice of civil society.” But in South a 10-person council. The Inkatha Freedom Party (IFP) wins 60,000 of the combined Africa, civil society is interpreted as nongovern- proportional representation and ward votes in Makana, the ANC wins 32,000, and mental, a very odd way to think of legally man- the Democratic Alliance (DA) wins 8,000. The quota for each seat is 10,000 votes. Therefore the IFP is entitled to six seats, the ANC to three, and the DA to one. The dated institutions. IFP wins four of the five wards and therefore already has four of the six seats they The lack of legislation specific to ward com- are entitled to. The DA wins one ward and is entitled only to one seat. The IFP is mittees may explain why many councilors do not given another two seats to be filled by people on their proportional representation have them. Surprisingly, poorer, rural areas have list. The ANC won no ward seats and therefore gets three proportional representa- had more success establishing ward committees tion seats. The DA gets no proportional representation seats. than have urban areas. In urban areas, however, 61 percent of people in informal settlements had heard of ward committees, compared with only Parker 2005). Communities sometimes perceive 35 percent in formal areas (Buccus and others that ward committees are “owned” by the ward 2008). councilor. Animosity among political parties Preliminary research findings suggest that often hampers participation, and civil society ward committees are failing to enhance partici- is often deliberately excluded from governance patory local governance (Barichievy, Piper, and (Habib and Kotzé 2003).

Accountability in Public Services in South Africa 27 citizens do Some argue that the functioning of ward and in alienation from the institutions of local not perceive committees is affected by the lack of capacity democracy” (Booysen 2007, p. 24). Ward coun- that the formal of elected members (Masango 2002). Others cilors have often been the target of frustration, posit that the communication links among the sometimes violently (Bekker and Leildé 2003, democratic councilor, the ward committee, and the commu- p. 154). However, the protests worked in the system serves nity are insufficient. Ward councilors are legis- sense that they triggered the attention of absen- them in the way latively mandated to consult with communities tee councilors, mayors, and municipal managers. by hosting quarterly ward committee meetings, President Mbeki even visited some of the af- they expect for instance, but the public in many parts of fected areas, and the ANC replaced several in- the country is not aware of this mandate so the cumbent councilors. President Zuma is continu- councilor is not held accountable when meet- ing the tradition of ward visits. In other words, ings do not occur. Greater public understanding South African communities have incorporated of ward councilor roles and responsibilities is a “the brick” (protests) in their accountability critical missing link in this public accountability mechanisms (Booysen 2007). mechanism. The ANC has a voter loyalty of 95 percent Without mechanisms for effective partici- (CSIR 2006). After the protests, voters returned patory governance, communities find alterna- the ANC to power by even greater margins than tive, often undemocratic methods to express before: in national elections for local govern- dissatisfaction. The apartheid era trend of ment, ANC votes grew from 60 percent in 2000 nonpayment is still popular despite govern- to 66 percent in 2006, 1.3 million more votes in ment campaigns to counter it (Cashdan 2002), a voter base that had grown by 2 million. Turn- frustrated citizens often vandalize government out did not change and the ANC increased its property and development projects, and violent votes in the areas most affected. So voting and protests are increasing due to insufficient ser- protests go hand in hand. A November 2005 vice delivery. survey focusing on urban areas (56 percent of the South Africa displays a complex paradox: population, with a sample size of 5,143) found comfortable ANC voter returns in local govern- that 52 percent of all respondents agreed that ment elections contrasts with the many town- voting helped in getting better services, and 54 ship and village protests that express anger with percent thought the same of elections (Booysen ANC service delivery. The political party is sel- 2007). These patterns remained unchanged in dom the target of frustration at the local govern- the election of 2009. ment level in South Africa (Bekker and Leildé 2003). To understand this paradox, Booysen Broader political accountability (2007) points out that the “brick” and the “bal- A more fundamental accountability issue may lot” may complement the citizens’ strategy to be at play beyond a mere complaint about ser- improve services. vice delivery. The primary issue, several politi- Booysen describes a pattern of “absentee rep- cal science researchers argue, is that citizens do resentation.” Councilors often do not reside in not perceive that the formal democratic system their communities and are not available for con- serves them in the way they expect. sultation with their constituents. A May 2005 Citizen concern about government respon- AC Nielsen survey records very low interaction siveness is documented in several surveys by the between voters and their municipal councilors, Institute of Democracy in Africa (IDASA), the with 80 percent of voters saying they saw coun- Afrobarometer, and the Human Sciences Re- cilors once a year or less (Booysen 2007). More- search Council. These surveys suggest that citi- over, voters deem the ward committees to be po- zen responses to government are not usually ex- litically biased, and committee members find it pressed as a demand to have services delivered. difficult to even communicate with their coun- Citizen responses, in other words, do not con- cilors since members have no budget to make form to the current service delivery paradigm phone calls or travel (Booysen 2007). in which citizens are like clients in a restaurant­ Protests “were firmly anchored in the fertile —­they want a service and judge the provider by bases of dissatisfaction with social conditions how well it does. People want the government to

28 Accountability in Public Services in South Africa serve them individually, but they judge the gov- national and provincial levels is seen frequently the subservient ernment’s ability to serve using broader, more as an obstacle to political accountability because role of provincial complex factors: people want to be listened to decisionmakers are directly responsible to party and local and they place value on mutual trust. Demands leaders rather than voters. Evidence at the pro- extend to issues such as responsiveness, account- vincial level suggests that legislative account- governments makes ability, and respect. ability and oversight are weak and that the elec- accountability This has implications for our report. If peo- toral system may well be a main cause (Rapoo less likely ple do not feel that the government is broadly 2004). However, voting for local government is accountable than they will not feel that it is ac- intended to make ward councilors directly ac- countable and responsive in providing services. countable to voters, but accountability seems The spate of local protests further indicates this weakest at the local level. feeling. An unpublished IDASA study suggests Second, current interpretations­—­held by that participation in the recent wave of violence many politicians­—­of the subservient role of pro- against African foreigners was related to real vincial and local governments make accountabil- or perceived exclusion from democratic insti- ity less likely. While the Constitution protects tutions. The study found that where people felt the autonomy of local and provincial govern- themselves included they did not take part in ments, it usually regards them as implementing the violence. Similarly, the urban poor have con- agencies for the national government in spite of cerns extending beyond lack of service to how the existence of Integrated Development Plans, they are excluded from certain urban space­—­a which are meant to instill local choices. The sub- remnant of apartheid geographic separation be- servience of local and provincial governments tween blacks and whites. to the national sphere is expressed by national The disconnect between citizens and sector plans and the work of MINMECS (liai- politicians/­policymakers could explain why son forums for national ministers and provincial policy that informs service provision can be se- economic councils), which often lock provinces verely out of touch with the needs and wants of and localities into national plans. grassroots citizens (Friedman 2005; Friedman, Representation in provincial and local gov- Hlela, and Thulare 2005). For instance, the ernment is diluted or eliminated as provinces South African National Housing Forum in the and localities are seen simply as implementers­ early 1990s spent years negotiating how to make —­as service deliverers­—­despite the fact that mortgage financing more accessible to the poor, they are made up of elected representatives. Un- despite evidence that poor people view mort- less provincial and local governments are seen as gages as a trap to be avoided. The first leg of the sources of representation­—­among other things, long route to accountability breaks down if poli- they must be able to make decisions in response ticians/policymakers do not know what services to local constituencies­—­they will not be held citizens want and how citizens will react to ser- accountable. vice initiatives. Subservience of local and provincial govern- So why does this disconnect exist? This re- ments to the national sphere also affects public port cannot fully answer this question, but it representatives’ understanding of their roles. suggests the following points. Local representatives tend to see themselves as First, evidence on how the electoral sys- service delivery agents rather than public repre- tem contributes to the misunderstanding be- sentatives, leading to weak provincial oversight tween politicians and citizens is inconclusive. and lack of provincial legislative activity. At best, the electoral system is a necessary but Subservience of local and provincial gov- insufficient condition for accountability: change ernments also exacerbates tendencies to evalu- in the electoral system may help, but it will ate provincial governments purely on whether not ensure accountability unless other changes they are “delivering” now, rather than on how are made. Electoral systems may be a problem accountability would be best served in the long at some levels of government, but they are not term­—­by having an elected government at the in themselves a cause of the misunderstanding. provincial level willing and able to hold public A single-list, proportional electoral system at servants accountable, for instance.

Accountability in Public Services in South Africa 29 Box 2.2 Improving local development through community-based planning accountability by substituting for, rather than complementing, vigorous representative gov- To improve the participatory content of the Integrated Development Plan, South ernment. While not flawed in design or lacking Africa has applied community-based planning in 12 municipalities, including the sufficient commitment, they may simply be inca- large Ekhuruleni, eThekwini, and Mangaung as well as other smaller, rural munici- pable of enabling the poor to make office holders palities. Several hundred participating wards have already received initial open- ended grants ranging from ZAR50,000 to ZAR200,000, which they have accounted accountable. Rather than simulate democracy for with minimal problems. The approach­—­developed across four countries by the through structured forums, institutions must be Khanya–African Institute for Community-Driven Development, in partnership with made to work as intended by the Constitution to the South African Local Government Association and the Department of Provincial make them accountable. and Local Government­—­has been summarized in user-friendly manuals accredited This is an important area for further research by the State Information and Technology Agency. The program has also been inde- and innovation: how do you induce more active pendently evaluated and is ready for national scale-up. citizen engagement with government? Account- Community-based planning is a week-long process in which communities ana- ability will be weak where citizens cannot en- lyze opportunities and constraints in their areas and their livelihoods (but also in- cluding the unemployed, young women, and single mothers) and discuss desired gage, and information is an important part of outcomes and risks. Communities select the top five priorities, decide which can the solution. How the government informs citi- be implemented by the community and which by the municipality or others, and zens is also as important as whether it does so. develop an action plan. The plan is then presented to the entire community, at Public relations–style communication of govern- which point the community can access the unrestricted grant to begin implement- ment plans will do little to enhance accountabil- ing projects. ity; true engagement involves explaining options Similar approaches have been shown to make plans more relevant to peo- in a way that allows citizens to choose (box 2.2). ple worldwide. These approaches involve citizens in the provision of public ser- Citizens should choose services and how vices, help citizens control their lives, and promote community action to overcome dependency. they are provided, and links with grassroots or- Community-based planning was evaluated by the University of the Free State ganizations could initiate government-citizen in 43 wards, with the following results: interaction to increase accountability. Details • Of 43 wards, 42 had approved plans and 41 had spent on their projects using can also be critical: if a conversation is held only the grants. in English, and only in settings chosen only by • Participation was widespread, including by the poor and disadvantaged. officials, grassroots people are unlikely to feel • The priorities of the poor were represented in the final plans. that (if they know enough English) what they • Community-based planning was useful to most wards. have to say matters. • Facilitators, drawn from a wide variety of jobs in the municipality, said they felt positive about the experience. The fourth point regarding the communi- • The manuals were found to be very helpful. cation gap between politicians and citizens is • Training and support to facilitators, however, could have been improved. that many people choose to remain silent even when they wish to speak, especially in societies Source: Goldman 2009, 2010. where undemocratic hierarchies have long been entrenched. This is as true in participation ex- ercises as in the workings of formal democratic Centralization in political parties is also government. Simple interventions centered on common. Unlike in other new democracies, “educating” people about their rights, or about South African parties have clear identities and the importance of holding officials accountable, instill deep loyalty in politicians and voters. will not work. Scott (1987) and Monga (1996) While enhancing loyalty to the democratic sys- show that grassroots people know accountabil- tem, strong party affinity has prompted national ity is important but feel powerless. Greater ac- parties to subjugate provincial and local repre- countability will ensue only if people feel they sentatives to serve party leadership rather than can truly act and speak without reprisal, rather the citizenry. Again, this increases the sense than telling people what they already know, or that elected officials are implementing agents by “educating” them to be citizens. Citizens or- of national government rather than public ganizations can play a crucial role, but govern- representatives. ments can also reduce the costs of speaking and Third, some formal forums created by acting while increasing the rewards. the government­—­such as the Integrated De- Citizen silence manifests in many ways velopment Plan­—­can inadvertently weaken in South Africa. Obviously, apartheid has

30 Accountability in Public Services in South Africa conditioned some to believe that they will not Conclusion Changing electoral be heard if they speak, and South African soci- In principle, South Africa’s parliamentary systems will not ety is divided between the organized and those democracy, combined with a free press, allows be enough—­ more­ who lack the means to organize; some are given citizens to exert significant power over policy- a route to hold officials accountable while oth- makers and express their voice. But a number of fundamental ers are denied this route. The government should issues weaken citizen voice, or political account- issues of political try to interact with citizens to encourage a sense ability more broadly, including the unclear sta- accountability of political efficacy, particularly with citizens tus and rules for electing ward committees. whose circumstances and histories have condi- Committee establishment and election should are at play tioned them to believe that they do not matter. be mandatory, follow well-defined and pub- Citizen action should enhance service provision licized rules, and have clear roles and a small accountability, not create an obstacle to it. budget. The final point is that widespread informal- Changing electoral systems will not be ity and government reactions to it can erode ac- enough­—­more fundamental issues of politi- countability. For instance, shack-dwellers often cal accountability are at play. The existing ser- do not enjoy an automatic right to be heard be- vice delivery paradigm has undermined citizen cause their residence or other aspects of their choice of what services are needed and how they status are technically illegal, greatly weakening should be delivered. Provincial and local gov- their ability to hold the government account- ernments serve as implementing agencies of able. This is clearly the case in South Africa, the national sphere rather than as representa- where evictions or informal access to services tive institutions accountable to their local citi- are routine. The government should be particu- zens. Communication with all citizens must be larly thoughtful about upholding the citizenship institutionalized­—­particularly with the poor rights of people living and working in informal and those living in informal settlements­—­to en- settings. sure that they feel truly politically empowered.

Accountability in Public Services in South Africa 31

3

The Compact

The relationship between the politician/poli- frontline service providers (teachers and nurses, cymaker and the service provider is the second for example) need to make many context-spe- leg of the long route to accountability, increas- cific actions and decisions, which are difficult ing the internal capacity of the state to deliver to monitor and control centrally. Contrast this services. Public sector management reform is the with, say, macroeconomic management, which primary instrument often advocated for improv- can be accomplished by a small group of techno- ing this relationship. One could argue that it was crats. The challenge for public sector reform is to mainly this leg that was emphasized in much find ways to give frontline providers incentives of the “old” theory of the public service. For to do their jobs well without intensive central instance, Max Weber’s bureaucracy is entirely monitoring. Strengthening the short route to ac- governed by “internal” principles of continu- countability can help because it increases local ity, strict adherence to rules, and clear lines of oversight on service providers. authority, responsibility, and supervision; direct accountability of the bureaucrat to the citizen The compact does not play a role in Weber’s “ideal bureau- The compact is “the broad, long-term relation- cracy” (Weber 1978). ship of accountability connecting policymakers International experience shows that tech- to organizational providers. This is usually not nocratic public sector management reforms as specific or legally enforceable as a contract” in a context of weak voice­—­a political system (World Bank 2003, p. 48). Narrowly interpreted, with weak accountability between citizens and reforming the compact implies that the compact policymakers­—­are largely ineffective (Fiszbein The compact 2005). This implies not that reforms should be abandoned but that they should be linked with other innovations and improvements in the ac- Politicians/policymakers countability system. In particular, linking pub- lic sector management reforms with better The compact outcomes­—­as opposed to better input and out- put management­—­is promising. The improved Citizen voice outcomes will likely create political “wins” that would lead to political commitment to reforms. Citizens/clients Client power Providers One reason that public service reform is so difficult is that in most cases, large numbers of

Accountability in Public Services in South Africa 33 Strategic is made up of internal public service manage- In its comprehensive 2008 midterm review frameworks and ment systems that can be improved. Many of the transformation of the public service, the medium-term developed countries­—­Australia, New Zealand, Public Service Commission (PSC)­—­South Af- and the United Kingdom, for instance­—­have rica’s administrative monitoring and evaluation budgeting are reformed public service management systems body­—­organized its analysis around nine con- well developed with considerable success. stitutional principles. These principles were also in South Africa, However, in developing countries the record used as the analytical basis for the 2009 PSC re- of reform is mixed. Reforms of states’ internal port on the state of public services: but the two often organizational structures and public manage- • A high standard of professional ethics must be do not align ment systems in Latin America, for instance, maintained. The PSC commends the es- have led to some improvement in capacity for tablishment of a National Anti-Corruption core economic management of fiscal adjustment, Hotline but also notes that follow-up on re- macroeconomic planning, financial manage- ported cases has been poor. It warns of an ment, and tax administration, but more fun- increase in reported financial misconduct damental improvements to state systems have cases­—­from 583 in 2003/2004 to 1,042 in proven more elusive (Fiszbein 2005). 2006/2007­—­and that only about two-thirds In South Africa, reforms have met equally of senior managers in the civil service are in mixed results. Many recent initiatives to im- compliance with the Financial Disclosures prove services focus on the compact between Framework. The PSC worries about the lack policymakers and public service providers, but of more decisive follow-up in this matter as successful economic and fiscal management re- well. forms have not met parallel success in the reform • Efficient, economic, and effective use of re- of the more fundamental compact. Assessing re- sources must be promoted. Government sim- form of the compact shows that despite elaborate plification and budget standardization has planning, critical elements of the long route to improved monitoring and benchmarking. accountability are missing. Underspending­—­currently only about 1 to Strategic frameworks and medium-term 2 percent of the total budget­—­is not a sig- budgeting are well developed in South Africa, nificant problem. Coordination between but the two often do not align, making it dif- budgets is lacking, however, as is an un- ficult to link the strategy, outputs, and outcomes derstanding of the important difference be- to the budget. Results-based planning and bud- tween outputs and outcomes: departments geting have been adopted governmentwide, but report major variances in basic data from one the results focus more on inputs and outputs report to the next. The PSC also notes that than on outcomes. The new Presidential depart- it is often not possible to align departmental ment on Performance Monitoring and Evalua- strategic plans with their annual reports and tion will focus more on outcomes. The depart- budgets, which makes performance-based ment’s 12 desirable development outcomes and management difficult. associated output indicators will underpin per- Personnel management reform is es- formance agreements with departments, prov- sential for accountability of service delivery inces, and local governments, which will be sectors, which are usually labor-intensive. supported by improved data gathering and bi- In Latin America, as elsewhere, organized monthly reviews by the President. union labor resistance to performance-based Further, detailed managing-by-results pay and other incentive systems have often frameworks are not supported by strong moni- undermined reforms. Again, strengthening toring systems that measure deviations from the short route provides a promising alter- goals fast enough to take action. Only in excep- native approach. In Brazil, for example, vil- tional circumstances do sanctions incur when lages were given oversight over family health results or outcomes are not achieved. Finally, workers and school director elections (Fisz- South Africa’s compact includes public partici- bein 2005). pation as a constitutional principle, but such In South Africa, government and unions participation is often missing. fought throughout the 1990s and early 2000s

34 Accountability in Public Services in South Africa over how best to evaluate schools and teach- this poor performance, the PSC again notes Izimbizo is an ers. This adversarial relationship is based in the lack of follow through: “The most per- innovative, easy, history: anti-apartheid teachers struggled tinent reason pointed out is that there is an and effective against autocratic, judgmental, and puni- absence of sanctions. . .” (PSC 2008a, p. 50). tive “panel inspections” of black schools and • People’s needs must be responded to and the pub- mechanism to teachers, and inspectors were often removed lic must be encouraged to participate in policy- let citizens voice violently when they attempted to set foot in making. The PSC notes that several innova- their problems black schools. Given this history, perfor- tive participatory and feedback mechanisms mance management to evaluate teachers and have been established to support the Batho and expectations schools has been fiercely resisted by teacher Pele principles. Started in 2000, Izimbizos unions. Agreement was reached recently on are forums where senior government lead- an Integrated Quality Management Sys- ers discuss issues with the public. Two Im- tem (2004), which is currently being imple- bizo periods are held each year and include mented (HSRC 2005). a Presidential Izimbizo and other govern- • Public administration must be development- ment-led Izimbizos. oriented. The PSC notes progress on the The African Peer Review Mechanism Millennium Development Goals and over- (APRM) found Izimbizo to be an innova- all progress in service delivery: piped water tive, easy, and effective mechanism to let cit- reached 88 percent of the population in 2006 izens voice their problems and expectations, and basic sanitation facilities reached 70 but it also noted­—­as does the PSC­—­the percent. About 2 million subsidized houses lack of a feedback loop for communities to were made available by 2007; the public track whether their concerns are being ad- works program created 3.6 million person- dressed. The Izimbizo mechanism cannot work days over 2004–2007; and social grants replace more institutionalized consultation reached 12 million beneficiaries, 25 percent processes, which would ensure people’s own- of the population. ership of development programs (APRM The PSC remains concerned, however, 2007). about the lack of participation in these ser- The PSC notes other participatory pro- vices. It “has over the years raised concerns grams established by the government, in- about the manner in which poverty reduc- cluding Citizens’ Forums, Citizen Satis- tion projects are being implemented. Part of faction Surveys, and the placement of over the implementation weaknesses is the non- 3,000 Community Development Workers involvement of beneficiary communities in in about 2,000 municipal wards to improve the planning and implementation of pro- communication between the government grams.” “. . . departments generally do not and the public. However, the PSC finds that adopt project management approaches that participation overall has not been successful. allow for effective beneficiary participation The constitutional principles need more for- and alignment of projects with local devel- mal policies and procedures to be realized. opment plans” (PSC 2008a, p. 40). Perhaps more worrying: “. . . there has been • Services must be provided impartially, fairly, reluctance by government departments to equitably, and without bias. The PSC assesses follow suit and engage with citizens in their the implementation of the Batho Pele prin- implementation processes” (PSC 2008a, ciples and the Promotion of Administrative p. 56). Justice Act (PAJA) of 2000 as being highly In one of its strongest statements, and re- unsatisfactory: “. . . very few departments ferring to the continuing countrywide town- have service standards, and when citizens at- ship riots, the PSC states: “It is unfortunate tempt to complain, redress mechanisms are that the Public Service has not managed to either not available or ineffective. It may be perform successfully in entrenching pub- for this reason that invoking PAJA to hold lic participation measures, resulting in high government to account by citizens has been levels of citizen dissatisfaction and unrest. very low” (PSC 2008a, p. 49). In explaining There are clearly differing interpretations as

Accountability in Public Services in South Africa 35 The government­ to what constitutes adequate consultation; out between 80 and 90 actions in 2007, and wide monitoring consensus is required around this to pre- less than 10 percent resulted in measurable and evaluation vent the current impasse experienced” (PSC positive outcomes. The focus of all actions 2008a, p. 94). was biased heavily toward inputs (more than system—­ a­ critical • Public administration must be accountable. 50 percent of actions) and outputs (more accountability Since 2006, it has been mandatory for senior than 30 percent of actions).1 mechanism for managers to include the Batho Pele prin- The PSC also notes that most reports ciples in their work plans and performance are hard to read and understand, and prov- the second leg of agreements. However, using the Depart- inces, where the bulk of service deliv- the long route­—­ ments of Education and Social Development ery takes place, lag in reporting. The 1995 has been under as examples, more than 50 percent of senior White Paper on the Transformation of Pub- managers in these departments did not con- lic Service Delivery2 introduces the idea of construction for clude a performance agreement in their first an “‘annual citizens’ report,’ which in simple several years plans. For 2007/2008 only, 16 percent of language, often produced in local languages, national department heads and 13 percent explains what a department does, how it has of provincial department heads filed perfor- met its service standards and how it intends mance agreements with the PSC by the due to improve its services” (PSC 2008a, p. 72). date. “In a study conducted by the PSC, not a sin- The governmentwide monitoring and gle national department could confirm that evaluation system­—­a critical accountability it had produced such a report, and only 18 mechanism for the second leg of the long provincial departments had produced them” route­—­has been under construction for sev- (PSC 2008a, p. 73). Exceptions exist how- eral years. As discussed, the Department of ever: Limpopo and the Free State have ex- Performance Monitoring and Evaluation is cellent annual citizens’ reports, which are working hard to make the system a reality. distributed to the public every year on Batho The picture on accountability in finan- Pele day. cial management is spotty. According to • Good human resource management and ca- the PSC, some departments have received reer development practice to maximize human qualified audits four years in a row, but no potential must be cultivated. The PSC notes action has been taken. National departments high vacancy rates and large discrepancies receiving qualified audits increased from 7 between the vacancy numbers reported in in 2004/2005 to 11 in 2006/2007. For pro- the automated personnel system and those vincial departments, audits increased from reported by departments. In 2005, 10 per- 0 in 2004/2005 to 44 in 2006/2007 (PSC cent of public servants may have carried 2008a), still low compared with the total HIV/AIDS, a quarter of whom may die by number of provinces. The PSC continues to 2012. In 2004, 12.7 percent of educators in ask why audits are not conducted on a larger public schools were carrying the virus (PSC scale and why the President and Parliament 2008a). do not take action. • Public administration must be broadly repre- • Transparency must be fostered by providing the sentative of the South African people. Employ- public with timely, accessible, and accurate in- ment and personnel management practices formation. Progress has been made: depart- should be based on ability, objectivity, fair- ments have shifted from reporting inputs ness, and the need to redress the imbalances and activities to reporting outputs. The PSC of the past to achieve broad representation. urges this shift to move further, from report- This last area may be the most successful in ing outputs to reporting outcomes. The gov- terms of achieving the constitutional prin- ernment’s Programme of Action should be ciples: the racial composition of the public leading the way, but it is very much a work service now largely mirrors the demography in progress. According to the Programme of of the country, and women are more repre- Action­—found­ on the Internet and updated sented in senior management, reaching the every two months­—­the social cluster carried target of 30 percent in most departments.

36 Accountability in Public Services in South Africa In 2007, 35 percent of civil servants were fe- “teaching to the test:” when teacher compensa- Compacts male. But performance is disappointing on tion or promotion is determined by student per- would benefit employment of disabled people: only 0.2 per- formance on a standardized test, teachers will from regular cent of civil servants are disabled; the target tend to downplay those aspects of pedagogy not was 2 percent. covered by the test and concentrate on those that and accurate In conclusion, the compact is still perform- are. measurements of ing less than satisfactorily despite the emphasis The emphasis of South African reforms in inputs, outputs, put on it by policymakers. Planning and budget- public sector management seems to have shifted: ing systems are well developed, but it remains emphasis on pro-poor legal and policy frame- and outcomes difficult to link inputs to outcomes. Depart- works with participatory design and implemen- ments continue to focus too much on inputs and tation has moved toward a “delivery targeting” outputs, which distracts from outcomes and re- approach with strong focus on budgets and in- sults, making it difficult to monitor actual per- put-output targets. Consistent with the expe- formance. Where deviations from the compact riences of many other countries, this focus on are identified, follow-through is often missing. targets and delivery has come at a cost: depart- Finally, South Africa’s compact puts strong ments chase sectoral inputs and outputs through rhetorical emphasis on participation, but par- “silo-ized” and centralized delivery systems, ticipation itself is often lacking, suggesting that often while knowing that critical inputs and out- critical gaps in the long route to accountability puts from other departments are not in place to remain. ensure good outcomes. In their quest to deliver at scale, they often stifle and bypass citizen-user Measuring inputs, outputs, and participation. Although unintended, this places outcomes excessive focus on outputs, sometimes even at Compacts would benefit from regular and accu- the expense of outcomes. Since service providers rate measurements of inputs, outputs, and out- are rewarded for delivering outputs and inputs comes. But for the kind of services discussed rather than for successful projects (outcomes), here (transaction-intensive and discretionary providers become less responsive to beneficiaries’ services), they cannot be complete or have per- power to reward good outcomes or punish bad fectly measured outcomes. In education, test ones. scores may adequately reflect certain education Our hypotheses are: goals, but abstract thinking and social adapt- • First, despite the shift to outcomes empha- ability are not so easily measured. Similarly in sized by the Department of Performance health care, the alleviation of pain is a subjective Monitoring and Evaluation, existing com- judgment by the patient; many outcomes, even pacts still focus too much on inputs and when observable to the patient and the doctor, outputs. are not “contractible” in the sense that compli- • Second, compacts continue to focus primar- ance can be proved to a judge or other mediator. ily on the accountability between policy- Attributing impacts to provider actions is makers and service providers rather than on difficult in almost all social services. It is thus the relationship between service providers impossible to have performance contracts that and clients. While perhaps inadvertent, this make payments to providers dependent on out- undermines participation. comes; all contracts will require at least some payment of wages independent of outputs. Procedures When the actions of the provider are specified, Simplifying procedures is often the key to the results are often less than optimal because improved services. Administrative procedures of inflexible response to local variation. Further, need to be understandable and practicable to since the provider does many things, some or all the average users, not just those with greater of them hard to observe by the policymaker, the technical knowledge.3 In the Philippines, for risk that paying for measured outcomes will dis- instance, the government established a single place hard-to-measure tasks is always present. online public procurement portal to allow sim- In education, this can be seen in the practice of pler, speedier provision of common supplies. The

Accountability in Public Services in South Africa 37 simplifying “e-procurement” portal minimized face-to-face equivalent rates), the widest range of tariffs, and procedures is interaction between civil servants and poten- the second highest level of dispersion (as mea- particularly tial service providers, ensured equal information sured by the coefficient of variation) among de- access by various bidders for public contracts, veloping countries” (Belli, Finger, and Ballivian important and saved money (World Bank 2005a). 1993). The administration could afford high lev- because the Similarly, in the 1990s the Government of els of complexity because the apartheid govern- administration Vietnam launched a Public Administration Re- ment delivered services only to a minority. This form initiative for administrative procedures in legacy of bureaucratic complexity continues to inherited many vital service areas. As part of this effort, a one- constrain the black majority’s access to services. complexities from stop shop was piloted in Ho Chi Minh City in The remnants of the apartheid regulatory the apartheid 1995 to provide administrative services in busi- framework are visible in sector after sector. The ness and cultural activity licensing, land admin- transition did away with the most egregious rac- bureaucracy istration, construction, and notarization. The ist laws, policies, and regulations, while embark- one-stop service counter was intended to help ing on far-reaching reforms of the regulatory citizens and businesses interact with the pub- framework, but the magnitude of the task may lic administration. Ten years later, one-stop have been underestimated. shops had been established in a third of all pro- Moreover, many of the old regulatory frame- vincial departments, most districts, and many works have found new constituencies. For in- communes. stance, apartheid restrictions on the subdivision Public feedback indicates that one-stop of land intended to prevent “die verswarting van shops “have helped improve the efficiency, effec- het platteland” (literally: the blackening of the tiveness, transparency, and equality of admin- countryside) are now touted as tools to ensure en- istrative services in Vietnam” (Asian Develop- vironmental sustainability and high incomes for ment Bank 2005, p. 4). Procedures for dealing new black producers (Van den Brink, Thomas, with applications are listed, rationalized, and and Binswanger 2007). In another example, the linked in a single unit where a team has been Township Planning Act of 1970 forced blacks in established to process applications. The process peri-­urban areas into uniform, dense settlements is transparent, with clear published guidelines with standardized housing patterns, reminiscent indicating the documentation needed, the time of military barracks. Today, the Act is still in allowed to process applications, and the fees. place to prevent urban sprawl and to minimize This helps streamline and clarify the processes, infrastructure delivery costs, but it also prevents significantly reducing red tape and processing black communities from establishing mixed-use time. After the one-stop shop model had been settlements (known as villages in the rest of the implemented for five months for land adminis- world; Berrisford and others 2008). Annex 4 tration in Quang Tri Province, for example, the summarizes two communities’ experiences with time to issue land tenure certificates fell from 91 the complexities of South Africa’s procedures for days to 23 for urban land and from 90 days to 13 accessing land. for rural land. Finally, South Africa is pursuing one-stop In South Africa, simplifying procedures is shops under the name “Multi-purpose Commu- particularly important because the administra- nity Centres,” intended to provide communities tion inherited many complexities from the apart- with integrated, coordinated information and a heid bureaucracy. Some observers have charac- wide range of services. To our knowledge, there terized the complexity as deliberate, intended to has been no rigorous assessment of the perfor- make the administration as opaque and inacces- mance of these centers. sible as possible for the black majority while pro- tecting the interests of the white minority. The Silos number of tariff lines was so excessive, for in- Poor coordination of basic and developmental stance, that it almost looked like each producer local services has plagued planners and imple- had its own tariff line: “By the end of the 1980s, menters of development programs all over the South Africa had the most tariff lines (greater world. Their experiences show that the chal- than 13,000), most tariff rates (200 ad valorem lenge of coordinating and integrating programs

38 Accountability in Public Services in South Africa Figure 3.1 Example of a silo’s functions in a land reform project becomes particularly severe when programs are provided by separate sectors of a higher sphere National Provincial Local of government­—­a province alongside the central government government government Community government, for instance. In that case, financial Agriculture resources for the execution of local projects flow Housing through sector programs, often called silos, to Land National reform the local areas (figure 3.1). treasury project Land affairs These silos typically develop their own pri- orities, procedures, and timetables, creating Other several problems for the local government and communities. First, their priorities are invari- Source: Authors’ analysis. ably different from those of the local govern- ment and communities. Second, local officials and communities confront different, sector- problem”­—­the Department of Agriculture in specific procedures and timetables for the prep- the same ministry. aration of projects, application and approval for To understand the cause of these problems, funding, procurement of goods and services, we must look at the basic implementation and and disbursement of funds. This creates unnec- funding architecture in place­—­the silos. Of- essary complexity and drains local capacity to ficials cannot deliver desired outcomes because comply with the procedures. Inability to apply the ministerial departmental financing mecha- the financial resources provided by one sector nisms (the silos) do not fund the complete proj- to bridge funding gaps in other sectors adds to ect, just individual aspects. This silo “syndrome” the difficulties. In South Africa, the two case cannot be blamed on officials as it stems from studies of land reform highlighted in annex 4 the incentive structures of compacts: depart- show the severity of these coordination prob- ments until recently were held accountable only lems and how they stifle local initiative and for what they could deliver measured by inputs capacity. (budget) and outputs (for example, hectares of Sector silos not only create problems for land), not for joint activities with other depart- local government and communities, they also ments and the outcomes in terms of a benefi- undermine development priorities of an indi- ciary’s quality of life. Notwithstanding incessant vidual ministry. Acquisition is only the begin- calls by policymakers, alignment, collabora- ning of a land reform project; transforming land tion, and coordination among departments are into a successful farm that provides an attractive poor because incentives do not encourage these livelihood almost always requires complemen- practices. A senior manager in the City of Cape tary investment in fixed farm assets­—­irrigation, Town laments: “Even when we had a compre- roads, inputs, technical advice, and housing, hensive picture, we created silos” (Smith and among others­—­which can be more than twice Vawda 2003, p. 37). the land costs. In South Africa, the Ministry Another important side effect: since service for Agriculture and Land Affairs is responsible providers are rarely rewarded for successful proj- for financing, or cofinancing, all these program ects (outcomes) but instead for delivering out- components, except for housing. Nevertheless, puts and inputs, they may be less responsive to the senior ministry officials have considerable the voice and power of beneficiaries’ to reward difficulty convincing staff in the Department of good, or punish bad, performance. Accountabil- Land Affairs that simply delivering land is not ity between service providers and beneficiaries is enough. Even when staff know that these other limited. components are not in place and that the likeli- Very aware of the problems of output-fo- hood of outcome failure is very high, they will cused silos, the government has promoted often push ahead with the project to “deliver” “joined up service delivery systems,” which focus on land and budget targets. In their view, the on outcomes. “However, this is a difficult reform provision of the complementary and crucially agenda, because most municipalities have inher- important services is “another department’s ited, and in some ways deepened, the historical

Accountability in Public Services in South Africa 39 The establishment legacy of hierarchical and sector-focused depart- voice and client power. How is this Community of Community ments, which operate like silos” (Pieterse and Development Workers program fairing? Development others 2008, p. 9). Community Development Workers go Internationally, the crippling effect of silos through one year of training and receive a Workers is one on project and program execution is well doc- monthly stipend. They are supposed to be well government umented. We use the case of Area-based Rural informed on the services offered by all govern- response to local Development Projects as promoted and financed ment spheres, help communities access these by the World Bank and many other donors in services (grants, identification cards, and oth- coordination the 1970s and 1980s (see annex 4). South Af- ers), and identify problems with service delivery. and information rica also seems to experiment with variants of Formal accountability for Community De- problems this older approach: “One of the most important velopment Workers has evolved over time. In institutional innovations to shift this legacy of the first few years of implementation, Com- silo-driven service delivery is the shift within munity Development Workers did not report municipalities to area-based delivery models and to ward councilors or ward committees, instead agency-based management of precincts within reporting upward to provincial administrations. the municipalities” (Pieterse and others 2008, Currently, Community Development Workers p. 9). International experience suggests that trained through provinces as “municipal facili- these innovations may be difficult to implement. tators” are seconded by municipalities who take Again, well aware of the problems created responsibility for placing them. by silos, other government responses include A Handbook on Community Development the creation of “clusters” of departments­—­an Workers states that the workers are accountable attempt to centralize the coordination of sector to the Community Development Manager of ministries; Multi-purpose Community Centres­ the municipality and should be seen as “part- —­one for each district/metro; one-stop govern- ners” of the ward committees. Given their role, ment centers; and Community Development which includes ensuring “an organized voice for Workers, a more recent innovation. the poor between elections,” holding Commu- nity Development Workers more accountable to Community Development Workers ward committees seems appropriate. This could The establishment of Community Development allay ward councilors’ perceptions that Com- Workers is one government response to local munity Development Workers are interfering. coordination and information problems, which More trust needs to be built between council- are particularly pronounced in the South Afri- ors and Community Development Workers, can national and provincial silo system. The fol- starting with a clear delineation of roles and lowing quote describes the problem: responsibilities. “One of the causes of lack of service deliv- A second challenge in implementing the pro- ery is attributed to the shortage of expertise gram is the often poor mentoring and support and skills at local government level. The harsh given to the Community Development Work- reality is that in many cases, officials who are ers by the departments to which they are sec- employed in this sphere of government often onded. This is unfortunate because the potential do not know how to forge strategic links and for Community Development Workers to foster engage communities on the wide spectrum of communication between citizens and the state government-sponsored programs and projects remains untapped. As an example of how Com- which aim to better the lives of the disadvan- munity Development Workers can help, 10 pro- taged” (Foundation for Contemporary Research vincial Community Development Workers were 2007, p. 9). deployed in Cape Town to educate township While one can quibble with the perspective­ communities on water and sanitation services. —­blaming the local level instead of the complex- Because the water department put considerable ity and inaccessibility of the departmental silos effort into the management, training, and dele- driven from the higher spheres of government­ gation of responsibilities to the Community De- —­instituting a link between citizen-users and velopment Workers, residents of these 10 town- government programs can improve both citizen ships blossomed. The workers became effective

40 Accountability in Public Services in South Africa communicators to the water department on the competent professionals who were generally bet- Special “enclaves” service delivery issues raised by users. They be- ter paid than civil servants in the core ministries within the came critical intermediaries between communi- and who operated with a degree of autonomy bureaucracy ties and the city, helping the water department from governmentwide rules governing person- become more responsive to issues that could not nel and financial management. are sometimes be channeled through a call center. Enclaves initially benefit from strong politi- established to cal and donor support and commitment, but in bypass government Enclaves the long run they are vulnerable to the same sys- Special “enclaves” within the bureaucracy are temic weaknesses in accountability that justified systems sometimes established to bypass government their creation. In addition, their arms-length systems and the coordination problems inher- relationship with government, and the result- ent in sector-based programs. An enclave is a ing loss of oversight, can lead to opportunistic special agency that combines the various func- behavior, including reckless financial activities, tions of the line agencies into one organiza- corruption, and favoritism (Fiszbein 2005). tion that receives special status. The enclave is The creation of “public entities” is a grow- isolated from regular bureaucratic systems­—­ ing trend in South Africa. During the 1990s, 60 personnel hiring or financial management, for public entities were created, while only 10 were instance­—­to produce quick results. The enclave created in the previous decade. “Public entities solution is often pursued when the coordination were created to allow for flexibility in condi- problems of the regular line agencies (the silos) tions of service, to retain income from revenue seem insurmountable in the face of an urgent­ raised, to expedite systems (for example, pro- —and­ usually politically high-profile—­ delivery­ curement), to ensure operational autonomy, and goal. to implement commercial principles and prac- Social action funds for community-driven tices. Challenges however remain, around per- development represent successful examples of formance, corporate governance and commit- enclaves. Social action funds have been particu- ment of some of these institutions to the overall larly effective at channeling investment funds for development objectives of government” (PCAS basic, small infrastructure projects to previously 2003, p. 12). underserved areas. They can channel invest- Enclaves have been used for municipal ser- ments into poorer, remote rural areas less expen- vices. More recently, enclave-type legal enti- sively and faster than can line ministries, which ties for various 2010 World Cup projects were often focus on infrastructure in big urban cen- established, and “special purpose vehicles” have ters. Several impact evaluations have shown that been proposed to develop urban land and hous- social action funds have reduced poverty and the ing. In Johannesburg, the implementation of cost of infrastructure provision and maintenance the Igoli 2002 decentralization strategy led to (for example, see Rawlings, Darling Sherburne- the creation of 14 municipally owned entities. Benz, and Van Domelen 2003). While service delivery became more efficient, Considerable evidence suggests that social policymaking responsibility and coherence was action funds have been effective in expand- lost amid the decentralization of so many core ing service delivery in Latin America (Fiszbein functions. 2005). In many Latin American countries, en- International experience shows that enclaves clave agencies were set up to provide specific have had success delivering desired projects, social infrastructure services through these but enclaves also risk perpetuating parallel and funds. Often first launched as “social compensa- uncoordinated bureaucracies instead of funda- tion” programs, the funds softened the impact mentally resolving coordination problems. Spe- of structural adjustment policies through con- cial project enclaves play a role in certain situa- struction of small-scale infrastructure projects tions and can also provide good testing grounds in poor communities. In the 1990s, Peru created for new approaches, but they must be accom- a number of “autonomous” agencies for revenue panied from the outset by a clear strategy to administration, regulatory activities, and social fully integrate these programs into the regular programs. The agencies tended to be staffed with bureaucracy.

Accountability in Public Services in South Africa 41 Contracting Contracts manage the private contractors efficiently. Other the provision of Contracting the provision of public services to problems included a lack of clarity of roles, an public services private sector agents is an alternative to creating inadequate incentive structure, and an unclear enclaves. We define a contract as a “formal per- bidding process in the selection of contractors. to private sector formance agreement captured in a legal docu- The ill-designed concessions also did not pro- agents is an ment with a private sector provider.” vide incentives to the private operators to extend alternative to How does accountability work in a contract? water services to the poorest parts of Jakarta The best way to ensure service provider account- (World Bank 2006). creating enclaves ability is to have payment depend on the actual, Contracting social service delivery to non- successful provision of the service. In South Af- governmental organizations (NGO) is also com- rica, such service contracts exist for toll roads, mon. To improve efficiency and reduce costs, garbage collection, and airport management, a growing number of ministries of health in among other services. Central America began to purchase primary Internationally, solid evidence shows that healthcare services from NGOs­—­usually pri- contracting can be successful when the condi- vate, nonprofit entities­—­in limited geographic tions are right: the services provided must be af- areas. Through a formal, competitive process, fordable for the poor, which often requires some El Salvador, Honduras, Nicaragua, and Panama form of subsidy or regulation; and contracting awarded contracts to NGOs to provide coverage should lead to an increase in choice for the users. to remote (often indigenous) populations with Conversely, contracting is usually not successful little or no access to care, in return for a fixed, when consumer choice is reduced and­—­in the annual capitation payment. absence of a strong regulator working in the in- In education, the best known case of con- terest of citizens­—­a monopoly results. tracting with NGOs is the Fe y Alegría schools, State capacity may be insufficient to man- which teach roughly a million students in 14 age the complexities of long-term contracts or Latin American countries. Evaluation of the those particularly difficult to design and moni- program in nine countries indicates that Fe y tor. When the Johannesburg-owned waste man- Alegría schools outperform traditional public agement entity­—­Pikitup­—­started subcontract- schools: on average they have lower repetition ing refuse collection, for instance, service quality and dropout rates, higher overall school prog- declined and the city’s regulator had no way to ress rates, and higher overall retention rates. hold these subcontractors accountable (Smith These schools are based on an open environment and Morris 2008). Long-term “capture” prob- promoting effective parental involvement in the lems have also been identified in the regulation learning process, community involvement in of contracts with the private sector. One main school management, and student-centered strat- difficulty is ensuring a close relationship be- egies (Fiszbein 2005). tween the regulator and the executive branch. Indonesia has also had success contracting Balancing investor interests with those of con- services to NGOs. Since 1987, an NGO in Ja- sumers is also a challenge. karta has been running a school-based intestinal Private sector participation in Latin Amer- parasite control program. The scheme includes ica is credited with enabling significant im- providing information on the causes and conse- provements in the utilities sector: connections, quences of the infections, teaching pupils about labor productivity, and service continuity have the importance of personal hygiene, carrying out improved. Electricity coverage has increased biannual examinations, and treating infected broadly. Water and sewerage services have pri- children. The prevalence of intestinal worms marily benefited lower income groups, while a among children attending schools serviced by surge in telephone penetration primarily bene- the NGO declined from almost 80 percent in fited the middle classes. 1987 to less than 10 percent in 2003, while the By contrast, in Indonesia the contracting of average incidence in Indonesian schools remains water services to private consortiums led to neg- between 60 and 80 percent (World Bank 2006). ative outcomes. The public operator failed to ne- How is contracting working in South Africa gotiate comprehensive service agreements and to after more than a decade? The White Paper on

42 Accountability in Public Services in South Africa Municipal Service Partnerships (DPLG 2004) contracting. Healthcare contracting also pres- decentralizing and the Municipal Systems Act (as amended ents difficulties: our bodies naturally cure many service in 2003) envisage contracts between a munici- diseases, so it is difficult to ascertain whether responsibilities pal council and a private sector partner (a pub- care was appropriate and cost-effective without lic-private partnership), a public sector partner controlled studies. to lower spheres (a public-public partnership), or a community- of government Decentralization: international based organization or NGO. Smith and Mor- depends on ris (2008) argue that the current requirements of experience the Municipal Finance Management Act make Decentralizing public services is prominent on many factors these NGO contracts unattractive due to vast the reform agenda of many developing coun- complexities and management-intensive pro- tries. Decentralization has the potential to cesses. As a result, their popularity has waned. enable governments to better respond to local Arrangements with public sector part- conditions and demands. A main objective of ners have not fared much better. For instance, decentralization­—­usually linked to political “ESKOM supplies around 40 percent of the City motivation­—­is to strengthen citizen voice by of Cape Town’s electricity, but does not share bringing services and elected politicians closer any information with the city’s planners. This to the client, thereby strengthening the first makes effective coordination of energy planning leg of accountability. Decentralization also dis- in Cape Town almost impossible” (Pieterse and places the second leg of accountability from the others 2008, p. 11). national sphere to the local spheres. By itself, The South African experience with contract- decentralization does little to strengthen the ing should be assessed in further research. Have accountability of service providers to citizen- the shareholder and customer incentive problems users, although it clearly has the potential to do associated with a municipally owned company so. As we have already seen, this potential has been well managed? Have the specifics needed not yet been realized in South Africa. to agree on a contract and outputs tended to Decentralization can also provide a testing come at the expense of flexibility and outcomes? ground for innovations. Central government When contracts require substantive investments plays a critical role in stimulating such innova- that take a long time to recoup, has contract tions by providing the appropriate incentives and management been up to the task? an enabling environment, monitoring and evalu- It is necessary to determine whether some ating performance based on outcomes, harness- existing contracts may be inappropriate due to ing the lessons learned for the benefit of other the difficulty of monitoring and the need for local governments, and intervening when ser- continual decisions about when, where, and how vices fail. to deliver services. Possible improvements to The desirability of decentralizing service re- these contracts to obtain better outcomes need sponsibilities to lower spheres of government de- to be identified. pends on many factors, including the: The hypothesis to be tested is that con- • Political environment. tracting works best for services that are easy • Technology of the service involved (for ex- to monitor, are nondiscretionary (that is, they ample, the extent to which the quality of the do not require additional, context-dependent service can be measured and monitored). decisionmaking by the service provider), and • Skills intensity required. can be contracted for fairly short periods of • Extent of scale economies involved in deliv- time, making terminating the contract easier. ering the service. In other words, contracting works best for ser- • Existence of an effective system of intergov- vices delivered in a way that mimics a simple ernmental fiscal transfers and own-revenue market transaction in a competitive market raising capacities. without large informational problems involv- • Level to which services are being decentralized. ing impact. Natural monopolies­—­which exist • Clear assignment of roles and responsibili- in many services that rely on a large network­ ties among the spheres of government and —­are less likely to be good candidates for the fiscal discipline imposed on each sphere.

Accountability in Public Services in South Africa 43 • Capacity of national government to effec- political and administrative­—­authorities lack tively monitor and evaluate the other spheres the necessary fiscal resources. Moreover, if as in of government. India until the 1990s, decentralization just re- The 2004 World Development Report pro- places the central ministry with a slightly lower vides a useful framework of power devolution tier of government (a province or state) but ev- (figure 3.2). erything else about the environment remains the The success of decentralization depends on same—­ political­ accountability, administrative how it affects accountability relationships in systems, and client power­—­substantial positive terms of political authority, administrative man- change is not likely. dates, and fiscal resources (figure 3.3). It also Another risk of decentralization is that depends on whether the elements of political, closer proximity to the client can easily be off- administrative, and fiscal decentralization are set by the equally closer proximity to local power properly balanced. Fortunately, South Africa structures and capture (Bardhan and Mookher- has done well in this area, moving authority jee 2000). In Latin America, such proximity has and resources to local governments along with been identified as a main reason why the assess- the election of local councilors. In many African ment of decentralization has been mixed. Other and Asian countries, decentralization was only reasons include the creation of gaps and overlaps in roles and responsibilities, lack of political au- Figure 3.2 Decentralization of services thority and fiscal resources, and lack of effective supervision of local governments by national and Central government financing with 1 Easy to YES contracting provincial authorities, as well as by citizen-users

monitor? NO YES 2 Central government provision (Fiszbein 2005). Homogenous? In many Latin American countries, trans- Local government financing with NO 3 Easy to YES contracting ferring fiscal resources without having perfor-

YES monitor? NO 4 Local government provision mance accountability standards in place has Pro-poor? Client power—experiment with 5 undermined decentralization. In decentralizing YES contracts NO Easy to monitor? the education sectors in Bolivia and Colom- NO Client power—experiment YES 6 with self-monitoring providers bia, perverse incentives were created for each Homogenous? Client power—experiment NO 7 level of government to maximize its budget and Easy to YES with community control, vouchers monitor? powers, and local stakeholders found it diffi- NO 8 Client power—imitate market cult to participate effectively in school manage-

Source: World Bank 2003. ment. In Argentina, unclear compacts between the central government and provinces resulted in large differences in coverage and quality of Figure 3.3 Decentralizing accountability relationships education services across provinces; few or no targets and objectives were specified, and large Center transfers of money were not tied to any effi- Politicians Policymakers ciency indicator. Devolution Subnational government Economies of scale also affect decentraliza-

Politicians Policymakers Delegation tion in important ways. For instance, a network service such as water supply is characterized by Compact

Long route Deconcentration economies of scale in its management (for exam- Voice ple, water quality assurance and chlorination). These economies of scale were sometimes lost Citizens/clients Providers when many Latin American countries decen- Client power Nonpoor Poor Frontline Organizations tralized water services to municipalities (Fisz-

Short route bein 2005). Similarly, in South Africa the rapid

Services decentralization of waste water treatment to municipalities has compromised water quality in

Source: World Bank 2003. the poorer rural municipalities because of loss of economies in chemical costs and skills.

44 Accountability in Public Services in South Africa Good examples of arrangements where Most of the resources are no longer earmarked, strong intergovernmental compacts coexist with while raising local revenue remains difficult­—­not large fiscal transfers can also be found. In China, an ideal situation for accountability. A triangu- decentralization involving large transfers has lar system of accountability involving the heads worked well across many sectors. In Chile, the of the regions, the local councilors, and the com- education delivery responsibility was transferred munities is being improved gradually over time. to municipalities after clearly defining the re- Since 1998, the Indonesian government has sponsibilities, rights, and obligations of the new also implemented the community-driven Na- providers. Central resources allocated to munici- tional Program for Community Empowerment palities were linked to increased numbers of stu- (PNPM), which has emerged as one of the best dents attending class (Fiszbein 2005, p. 21). national community-driven development pro- Between 1999 and 2001, in the wake of the grams in the world. Community mobilization Asian Financial Crisis, highly centralized Indo- under the PNPM is coordinated with the trans- nesia built on longstanding analytical work to de- fer of funds for community projects and with ca- centralize in a “big bang;” the country has contin- pacity-building to strengthen communities and ued to improve on the decentralization framework decentralized local governments. Communities ever since. Local and provincial governments are have implemented hundreds of thousands of proj- autonomous territorial entities within the unitary ects in infrastructure, livelihoods, social services, state of Indonesia, with 33 provinces, about 440 and natural resources management. The PNPM districts, 100 urban municipalities, and 80,000 has enhanced top-down accountability across villages. To avoid centrifugal tendencies, de- local governments, introducing innovations in centralization was directed straight to the local accountability such as community audit com- governments, largely bypassing the regional au- mittees, NGOs and the press as monitors, and a thorities. Administrative responsibility for service comprehensive monitoring and evaluation system. provision was transferred to subnational govern- Since about the mid-1990s, many African ments, which employ more than three quarters of countries have pursued decentralization osten- the 2.4 million civil servants. The decentralization sibly as a means to improve accountability and laws require the central government to transfer at service delivery. A description of the “degrees” least 26 percent of net government revenues to the of decentralization in African programs is pre- subnational level, mostly to local governments. sented in table 3.1.

Table 3.1 Forms of decentralization

Degree of decentralization Political features Fiscal features Administrative features Deconcentration • No elected local government • Local government is a service delivery arm of the center • Provider staff working at local level are employees of center (minimal change) • Local leadership is vested in local officials such as a governor and has little or no discretion over how or where services and accountable to center, usually through their ministries; or mayor but appointed by and accountable to the center are provided weak local capacity is compensated by central employees • Voice relationships are remote and possibly weak • Funds come from the center through a central ministry or • Accountability remains weak if provider monitoring is weak; department citizens could have to rely on a weak long route stretching • No independent revenue sources to politicians at the center; a strong compact between policymakers and providers can compensate to some extent

Delegation • Local government may be led by locally elected politicians, • Spending priorities are set centrally, as are program norms • Providers could be employees of central or local government, (intermediate but it is still accountable fully or partially to the center and standards; local government has some management but pay and employment are typically set by center change) • Voice relationships are more local but can be overruled by authority over allocation of resources to meet local • Local government has some authority over hiring and location center circumstances of staff but is less likely to have authority over firing • Funding is provided by the center through transfers, usually a • Both long and short routes to accountability are potentially combination of block and conditional grants stronger; greater local knowledge can allow better • No independent revenue sources matching and monitoring of supply with local preferences, strengthening both the compact and client power

Devolution • Local government is led by locally elected politicians expected • Subject to meeting nationally set minimum standards, local • Providers are employees of local government (substantial change) to be accountable to the local electorate governments can set spending priorities and determine how • Local government has full discretion over salary levels, • Voice relationships can be very strong, also subject to best to meet service obligations staffing numbers, and allocation, and authority to hire and fire capture by elites, social polarization, uninformed voting, and • Funding can come from local revenues and revenue-sharing • Standards and procedures for hiring and managing staff clientelism arrangements from center may still be established within an overarching civil service • A hard budget constraint is imperative, creating incentives for framework covering local governments generally accountable service • Potentially the strongest long and short routes to accountability, but now also more influenced by local social norms and vulnerable to local capacity constraints and politics

Source: World Bank 2003.

Accountability in Public Services in South Africa 45 A sound Several common factors have impeded suc- to specific constituencies. Again, the tendency is decentralization cessful implementation of decentralization in to look upward to the central authority. framework has Africa. Finally, successful decentralization can be First, many countries may have focused described at least partly by the division of capital been established prematurely on devolution, the full transfer of spending among the tiers of government. In Or- in South Africa delivery responsibilities, staff, and financial re- ganisation for Economic Co-operation and De- with an advanced sources to local governments. Devolution is al- velopment countries, for instance, subnational ways a politically difficult process as centralized spending on public services and infrastruc- intergovernmental bureaucracies generally resist the loss of func- ture accounts for more than half of total pub- fiscal framework tions, staff, and financial resources. Premature lic spending. In a study of fiscal decentralization devolution usually slows capacity-building for carried out in seven African countries (Ghana, sustainable delivery of services and infrastruc- Senegal, South Africa, Swaziland, Uganda, ture. A few countries, including Ethiopia, have Zambia, and Zimbabwe) in the late 1990s, less been more effective by first building the capac- than 10 percent of public spending, on average, ity local governments need to successfully imple- flowed through subnational entities. In such a ment core local service delivery mandates. Ethi- situation, where 90 percent of public spending opia has established a national capacity-building is by national agencies, the ability of people­—­ conditional grant mechanism. Three and a particularly the poor­—­to exercise the short route half years after its creation, the mechanism is to accountability is limited (Steffensen and Trol- complemented by local government-matching legard 2000). grants, conditioned on broad indicators of per- The literature abounds with principles of de- formance, rather than on specific infrastructure centralization, and many are instructive. Recent and service investments. experience in Africa indicates that three could Second, many African countries have be critical: retained—­ and­ even expanded—­ local­ govern- • Do not focus prematurely on full devolution. ment systems established during the colonial era. Accountable local governments must first These systems typically include rural local gov- focus on successfully delivering their core re- ernment jurisdictions entirely dependent on the sponsibilities, and devolution can follow later center for resources­—­thus accountable mainly to as politics allow. the center rather than to citizens. • Subnational governments that are mainly In almost all large and complex societies, dependent on transfers from the center will some jurisdictions will not be able to sustain ser- usually focus on upward accountability. vice delivery through locally generated resources • “Allocative efficiency” is not just jargon. The due to remoteness or lack of natural resources, share of public spending carried out by ac- for instance. Nonetheless, limiting the num- countable subnational agencies is positively ber of local governments dependent mainly on related to development achievements. transfers from the center is a sound principle for public administration of decentralization. The system of decentralization in District local governments in many former South Africa British colonies do not conform to this prin- A sound decentralization framework has been ciple. Almost all of the more than 50 district established in South Africa with an advanced local governments in Zambia, for instance, de- intergovernmental fiscal framework. Condi- pend largely on central transfers. As in South tional grant funding from national to provincial Africa, this problem is often compounded by government is allocated using an equitable share unsustainable, intervening layers of subnational calculation (population- and poverty-linked). In entities: provinces, regions, or states. Unlike the addition, creditworthy municipalities can access 50 states in the United States, for example, sub- subnational sovereign borrowing, and real estate national entities in Africa frequently have little property taxes provide a source of local revenue. or no resources of their own, few real delivery as- A brief review of the recent history of de- signments (and those they do have are often per- centralization in South Africa may be useful. As formed poorly), and limited direct accountability negotiations to end apartheid began in the early

46 Accountability in Public Services in South Africa 1990s, a document entitled ANC Regional Policy4 in progress. Disputes over certain municipal most citizens clearly expressed the consensus that three con- boundaries still occur, and the Board contin- are not fully stitutionally defined spheres of cooperative gov- ues assessing and adjusting demarcations. But familiar with all ernance should be established. Chapter 3 of The it is indisputable that all South Africans now Constitution of the Republic of South Africa have, for the first time in the country’s history, a the government (Act 108 of 1996) succinctly lays out the “prin- clearly defined governance framework of munic- entities created to ciples of co-operative government and intergov- ipalities, provinces, and a national government meet their needs ernmental relations,” which all three spheres are operating through legislation and programs that expected to follow. Chapter 7 defines the roles emanate from the 1996 Constitution. and types of local governments (now known So who is accountable for what? In all com- popularly as metros, districts, and locales), and plex societies, most citizens are not fully familiar Schedules 4 and 5 define the functions of the with all the government entities created to meet three spheres, including functions concurrent their needs. Most citizens are generally aware with the national and provincial spheres (Sched- that national defense, central banking, and in- ule 4, Part A) and functions that focus on local ternational diplomacy are national government governments (Schedule 4, Part B). functions. Beyond these, most citizens look to As noted earlier, the clarity of South Af- the nearest and most accessible sphere of gov- rica’s Constitution is admirable: its 14 chapters ernment for assistance with day-to-day needs, and 7 schedules are concise, and it is written in including the delivery of basic, critical services comprehensible language. But as with all negoti- and infrastructure. As argued previously in this ated compacts, the Constitution reflects a num- paper, this short route to accountability works ber of settlements on critical issues, not least of best when the services in question are direct and which are the functional assignments included easily measurable: water, streets, streetlights, in Schedules 4 and 5. electricity, refuse removal, shelter, public edu- The Constitution was adopted in May 1996, cation, health services, identity cards, and the and the Municipal Demarcation Board was for- recording of life events such as birth, marriage, mally established in February 1999. The Board and death. completed its main mission of demarcating 284 Obviously, and for a variety of reasons spe- municipalities covering all of South Africa­—­ cific to each situation, public service provision is including 6 metros, 232 local municipalities, 46 never as clear and direct as citizens prefer. Econ- districts, and 3,754 wards­—­in time for the first omies of scale, measures of effectiveness and ef- truly democratic local government elections on ficiency, and history always produce a public ser- December 5, 2000. In just 15 months, the Board vice delivery and accountability framework more held hundreds of public consultations through- complex than citizens expect. South Africa is no out the country, received thousands of written exception. submissions, and digested massive amounts of Most of the public functions to be carried demographic, financial, economic, and geo- out by the three spheres of government are set graphic information system data to replace the out in Schedules 4 and 5 of the Constitution. 843 racially segregated local governments extant The functions clearly assigned to local govern- at the end of the apartheid era (Municipal De- ments in Part B are predictable: electricity and marcation Board 2004). The thoroughness and gas reticulation, firefighting, potable water sup- pace of these crucial reforms is viewed by most ply, waste water and sewage disposal, and mu- knowledgeable practitioners as nothing short of nicipal planning and public works. Part A of astounding. the same schedule lists public functions that are The full framework for democratic gover- concurrently the responsibility of national and nance and accountability in South Africa has provincial government: education at all levels, been in place for over nine years, and there have health services, the environment, housing, in- been three rounds of local government elections dustrial promotion, public transport, and urban so far (2000, 2005, and 2009). The full frame- and rural development. Schedule 5, Part A goes work for accountable governance in South Af- on to list functional areas of exclusive provin- rica is, therefore, very young and still a work cial competence, while Part B lists additional

Accountability in Public Services in South Africa 47 service delivery functions that may be “assigned” by provinces to functional assignments may not have been ex- accountability municipalities (see Section 156 (1) (b)). Section amined as thoroughly as necessary. in South Africa 99 provides that a national Cabinet member may Finally, the local governments were in tur- assign any power or function to a member of a moil in the mid-1990s; transitional councils is not so clear Provincial Executive Council or a Municipal throughout the country struggled just to main- after all Council subject to legislation and a Presidential tain basic services. Johannesburg, for instance, Proclamation. was technically bankrupt by 1998, and “stabili- The above shows that service delivery ac- zation” was the objective of the day. So assign- countability in South Africa is not so clear after ment of additional functions to local govern- all. This situation has developed for a number of ments with severely strained capacities and still reasons. indistinct jurisdictions was not in order. First, the truly remarkable consensus-build- How, then, does South Africa rate in terms ing exercise that led to the 1996 Constitution in- of the three principles outlined above? volved, of course, a number of negotiated settle- ments. Contention over the need for provinces Do not focus prematurely on devolution. South at all was considerable. The African National Africa has recently focused more on centraliza- Congress (ANC) favored a “weak” regional ar- tion than devolution. First, the more than 800 rangement but the “federalists,” particularly in segregated local governments of the apartheid the Cape and KwaZulu-Natal, insisted on the era were conglomerated into only 284 metro, dis- strong, elected provincial structures that finally trict, and local municipalities. The need to dis- prevailed. Once established, the provinces had mantle apartheid local government was unques- to have functions­—­hence Schedules 4 and 5. tionable, but South African citizens today have During these negotiations, the “localists” argued fewer and more distant local governments to for metropolitan area independence from prov- hold accountable. Main delivery responsibilities inces,5 but this approach was successfully op- have also been centralized, notably housing and posed by both the federalists (who worried that electricity distribution. So while South Africa the ANC would dominate in Durban and Cape has not confused decentralization and devolu- Town, thereby eviscerating the KwaZulu-Natal tion, it also has not rigorously examined the sub- and Cape provinces) and more hierarchically sidiarity principle,6 although the Ministers of oriented elements of the ANC. Finance and Provincial and Local Government Second, as it prepared to govern, the ANC did introduce a motion to do so in Parliament was designing the Reconstruction and Develop- in 2007, and the Department of Planning and ment Programme, aimed particularly at redress- Local Government is reviewing structures and ing service delivery and housing imbalances in functions. favor of the disadvantaged majority. The Re- construction and Development Programme was Upward accountability is encouraged by dependence consciously supply-driven, and given the vast on central transfers. Of the 9 provinces and 284 disparities imposed by the apartheid regime, local governments in South Africa, only the 6 this choice was widely viewed as justified. The metros and a number of the larger towns and cit- supply-driven approach meant that functions ies do not depend significantly on central trans- that could be delivered effectively by local gov- fers. Again, it is unquestionable that redistribu- ernments and communities, particularly hous- tion of revenues has been, and will remain for ing, were reserved for national and provincial some time, a top priority in South Africa. It is implementation. disconcerting, however, when subnational gov- Third, the implications of constitutional as- ernments are not held accountable primarily by signment of certain functions to the national the citizens who contribute the majority of the and provincial spheres may not have been fully financial resources for services. considered­—­notably environmental regula- tion. Given the breakneck pace of change that Allocation of public spending is done locally, and typified South Africa throughout most of the the data is not encouraging. While total reve- 1990s, it is entirely understandable that certain nue directed to local governments is increasing

48 Accountability in Public Services in South Africa sharply, the percentage of revenue going to local attract capable professionals. Municipalities also governments is not. In fact, over 90 percent of suffer from decades of neglected investment and the national budget is still allocated to national development, and the resource base to support and provincial governments (table 3.2). the functions they must perform is inadequate. The Equitable Share “unconditional” grant The assessment also found widespread mu- and the conditional Municipal Infrastructure nicipal corruption and abuse of political office Grant combine to make up most transfers to and job status for personal gain. “A culture of local governments (figure 3.4). Both of these patronage and nepotism is now so widespread transfer mechanisms are well designed. The Eq- in municipalities that the formal municipal uitable Share allows local governments to fully accountability system is ineffective and inac- determine how funds will be spent, while Mu- cessible to most citizens. There is now a lack nicipal Infrastructure Grants must address ser- of citizen confidence and trust in the system” vice and infrastructure backlogs. So the main (CoGTA 2009a). The report noted inappropri- South African intergovernmental transfers do ate appointments to meet employment equity allow local decisionmaking and, in theory, effi- targets, deployment of unqualified ANC cadres cient allocation of funds. in senior professional positions, and decision- In terms of the three principles outlined making and resource allocation based on party above, the high level of provincial and local gov- and factional politics rather than assessments of ernment “grant dependency” is most concern- needs (CoGTA 2009b). ing. A second concern is that nearly half the na- Table 3.2 Vertical division of revenue, 2003/2004–2009/10 (ZAR million) tional budget­—­including high public spending on health, education, and housing­—­continues to 2003/2004 2004/2005 2005/2006 2006/2007 2007/2008 2008/2009 2009/2010 flow through provinces. Level of government Outcome Estimate Medium-term estimates In conclusion, decentralization has occurred National 148,142 167,289 194,723 212,629 246,937 273,929 302,672 in a number of areas, either mandated by the Provincial 122,673 138,511 154,368 178,871 205,224 238,076 268,158 South African Constitution or as part of a min- Local 11,581 13,808 16,682 26,051 37,127 47,651 47,651 isterial delegation (for example, Land Affairs). Total 282,396 319,608 365,773 418,001 489,288 559,656 618,481 The Constitution explicitly lays out the principle National (percent) 52.5 52.3 53.2 50.9 50.5 48.9 48.9 of subsidiarity as a guide for decentralization. In Provincial (percent) 43.4 43.3 42.2 42.8 41.9 42.5 43.4 Local (percent) 4.1 4.3 4.6 6.3 7.6 7.6 7.7 practice, however, only limited decentralization Total 100 100 100 100 100 100 100 to municipalities has happened. The main hypotheses to be explored further Note: Data does not include spending of own resources by local governments. in the next chapters are: Source: National Treasury local government database. • Some local government service assignments are underdeveloped. Figure 3.4 Transfers by type, 2003/2004–2009/2010 • More critical services could be directly de- livered by local government, following the ZAR billions Equitable share Municipal infrastructure grant Capacity building principles of subsidiarity (for example, hous- 60 2010 World Cup Indirect infrastructure Other direct infrastructure ing, education, and health). Other recurrent 50 Performance of local governments in South Africa 40

State of Local Government in South Africa, 30 released in October 2009, found that local gov- ernments nationally suffer from low confidence 20 and that structural weaknesses account for many 10 of the challenges municipalities face. While given an exceptionally wide-ranging and chal- 0 lenging mandate, many of the administrations 2003/04 2004/05 2005/06 2006/07 2007/08 2008/09 2009/10 are less than 10 years old and have had to estab- Source: National Treasury local government database. lish themselves in areas where it is difficult to

Accountability in Public Services in South Africa 49 Facing national local government elections with dedicated national and provincial task in early 2011, a comprehensive Local Govern- teams. It acknowledges the extent of variation ment Turnaround Strategy (LGTAS) has been across municipalities­—­from established met- prioritized, with strong emphasis on strength- ros to still-maturing municipalities serving pre- ening performance monitoring, governance, dominantly rural areas­—­and proposes different and accountability (figure 3.5). The approach, approaches according to the capabilities of the endorsed by the Cabinet in December 2009, re- local government: competent authorities would quired every municipality to draft its own turn- receive greater latitude while struggling authori- around strategy by April 2010 to remedy prob- ties would receive greater support. lems in its area of jurisdiction. These strategies The Department of Cooperative Governance would then form part of the Integrated Develop- and Traditional Affairs (CoGTA) has hinted ment Plan for the coming year. that some municipal powers and functions may The LGTAS has five core objectives: be revised. CoGTA provides for a special pur- • Ensure that municipalities meet the basic pose vehicle to fast-track infrastructure develop- needs of communities, with systems to ac- ment to accelerate service delivery for munici- celerate quality service delivery, considering palities that need support. The strategy reaffirms each municipality’s conditions and needs. the government’s commitment to closer engage- • Build clean, responsive, accountable local ment with its citizens, primarily through a re- governments. fined ward committee system “based on the will • Improve functionality, performance, and of the people.” professionalism in municipalities. Supporting the new emphasis on account- • Improve national and provincial policy, sup- ability and professionalism across President port, and oversight to local government. Zuma’s government, all mayors will sign per- • Strengthen partnerships among local gov- formance and service delivery agreements with ernment, communities, and civil society their respective Provincial Member of Execu- (CoGTA 2009b). tive Committee, who will sign complementary Turnaround strategies will focus on immedi- agreements with their ministers. The ministers ate priorities before the 2011 local government have already signed performance agreements elections and on longer term objectives once new with the President. Municipal managers will municipal councils are elected. also sign performance agreements with their re- The LGTAS proposes more coordinated pro- spective mayors. vincial and national support to local government, The new emphasis on performance objectives and desired outcomes, and monitoring progress Figure 3.5 Key interventions proposed by the Local Government toward targets, introduces a new degree of ac- Turnaround Strategy countability in public office. The new system does raise questions about how performance will National government be monitored, however. The real measure of the (including state- owned enterprises) success of the LGTAS will be how much it shifts will organize itself better in relation to the focus of service delivery from infrastructure Good citizenship local government Provinces will campaign with improve their provision to good service by allowing citizens governance values support and oversight a meaningful voice in holding government ac- will be launched to responsibilities over unite the nation local government countable. Will citizens be invited to contribute Interventions their own assessments of how government is per-

Political parties Municipalities will forming? Will their own perspectives and local will promote and reflect on their enhance the own performance experiences be taken into account? If ward com- institutional integrity and develop their of municipalities own tailor-made mittees are to be the primary communication All three spheres turnaround strategies will improve forum, how will their credibility be transformed? intergovernmental relations in practice Notes

Source: CoGTA 2010. 1. See www.gov.za. Examples of actions that we classified as outcomes include areas of

50 Accountability in Public Services in South Africa progress, such as a decline of malaria in a to accountability, it also has its place in the given zone, an increase in HIV/AIDS pa- short route. Since the early 1990s, simplify- tients receiving antiretroviral therapy, a rise ing and delegating infrastructure and ser- in the number of households with proper vice provision to communities through com- sanitation, among others. Examples of out- munity-driven development has become an puts include measurements, such as the important feature of World Bank–financed number of low-income houses delivered to programs all over the world. Among other citizens, the number of flags installed in innovations, it requires the direct transfer of schools, and others. In some instances, the financial resources to communities for their distinction between outcomes and outputs own management. As shown in chapter 8, is blurred. An increase in households with scaling-up this innovation was only possible sanitation or the creation of jobs through the after radical simplification of disbursement, Growth Programme, while outputs strictly procurement, and the financial accounting speaking, can be considered outcomes as systems of communities, governments, and both achievements clearly improve the well- the World Bank. being of households. By contrast, the open- 4. ANC Constitutional Committee and De- ing of 10 new schools, or registration of partment of Local and Regional Govern- unemployed people to the National Youth ment and Housing, ANC Regional Policy; Service, does not automatically translate into draft discussion document published by the positive outcomes. Finally, inputs included Centre for Development Studies, University all policy documents approved, strategy of the Western Cape, October 1992. notes completed, studies presented, projects 5. Similar to “chartered cities” in the Philip- implemented, and resources distributed. pines and elsewhere. 2. http://www.info.gov.za/whitepapers/1995/ 6. Subsidiarity is an organizing principle that transformation.htm. matters ought to be handled by the small- 3. While simplifying procedures is here dis- est, lowest or least centralized competent cussed as part of leg two of the long route authority.

Accountability in Public Services in South Africa 51

4

Citizen-User Power

“Development is not about the delivery of goods All that remains is to get the finances and tech- to a passive citizenry. It is about active involve- nical capacity in place. In short, developmental- ment and growing empowerment. In taking this ism becomes predominant, with democracy al- approach (that is, ‘A People-Driven Process’) we most a luxury. The problem with this view is that are building on the many forums, peace structures, it is not always obvious what people living in and negotiations that our people are involved in specific communities actually need, and in par- throughout the land” (ANC 1994, pp. 4–5). ticular it is often unclear how best to deliver ser- This quote captures a powerful vision of vices. If we accept that each context is different participation in the African National Con- then we cannot assume that ‘one-size-fits-all’ is gress’s (ANC) Reconstruction and Development going to work” (DPLG 2008, p. vi). Programme­—­its development strategy for the Despite the minister’s vision, Pieterse and new South Africa. Friedman (2006a) reminds others (2008) conclude that the early phase of us that the post-apartheid government did not developmental local government was indeed need to be convinced of the virtues of participa- participatory and transparent but that the pen- tory governance mechanisms because of how the dulum has now swung back toward technical struggle had been conducted. and financial imperatives. While service deliv- The short route to accountability­—­the di- ery policies continue to demand that citizens and rect interaction between citizens and service civil societies participate, this study shows that providers­—­is all about participation. Citizens practice and quality across the board are scant should be able to exercise varying degrees of in- and uneven. fluence over service providers­—­we call this “cli- Client power ent power.” Sydney Mufamadi, former Minister for Pro- vincial and Local Government, describes the Politicians/policymakers tensions in South African policymaking: “There are those within the wider policy de- The compact bate who focus exclusively on the need for ac- celerated delivery of services and question why Citizen voice we need such complex time-consuming partici- patory and democratic processes. This view as- Citizens/clients Client power Providers sumes that everybody has the same needs and therefore it is obvious what must be delivered.

Accountability in Public Services in South Africa 53 public South Africa’s Batho Pele principles for- information to citizen-users and holding con- involvement malize strong short-route accountability in ser- sultations with them is insufficient to create a would require vices. In its Country Report for South Africa, strong accountability link: citizen-users must be the African Peer Review Mechanism (APRM) able to alter service provider behavior. a confident highlights the Batho Pele principles as “best citizenry that practice.” It goes on, however, to say that “[i] The Batho Pele principles of consultation and redress holds officials to nnovative initiatives like Batho Pele are still poorly understood and have not addressed the Two of the most important Batho Pele princi- account without fundamental issues behind poor service delivery. ples supporting citizen voice and user power are fear of reprisals Public participation is usually ad hoc with little consultation and redress. The PSC assessed the information of public awareness” (APRM 2007, application of these principles, concluding: “Ini- p. 104). tial steps to improve service in line with Batho The Public Service Commission (PSC) de- Pele have had uneven success. Weaknesses scribes the origins of the Batho Pele principles include poor senior management involvement and, in a prescient statement, what happens in the monitoring of the implementation of the when they are ignored: principles driving Batho Pele, and lack of public “It should come as no surprise that public involvement in the enforcement of these prin- participation was enshrined in the Constitution ciples. Such public involvement would require a as one of the values and principles of public ad- confident citizenry that holds officials to account ministration. Equally important also is the fact without fear of reprisals such as the withholding that given its history, ours is a citizenry that not of services and benefits” (PSC 2008a, p. 13). only values but also expects public participation. Practically, this means that if state institutions Consultation do not institutionalize and adequately promote The PSC’s report on Batho Pele and consulta- public participation, citizens are likely to find tion (PSC 2008b) found that the definition of other ways to express themselves and attract at- consultation used by most departments reflects a tention, even if this involves using less construc- misunderstanding of the intentions of the White tive mechanisms” (PSC 2008a, p. 53). Paper on Transforming Public Service Deliv- Recently, some citizens have drawn atten- ery. In most cases, departments cite instances of tion to the need for public participation through “consultations” when departments only inform protests and activism. This should signal to the customers and stakeholders about services ren- government that effective communication and dered. But informing is not consulting; consul- public participation must remain a fundamen- tation should engage the public through a plat- tal priority. “A key part of these battles is about form where they can express views on the level what constitutes consultation and how citizens and quality of services they receive. get involved in decisions which affect their lives” The report found that only 5 percent of the (PSC 2008a, p. 53). departments indicated that they consult on their The short route to accountability is particu- service standards. Not surprisingly, the PSC larly underdeveloped in South Africa. A 2007 found in an earlier study that citizens are gener- Citizen Satisfaction Survey focused on the ser- ally not satisfied with the consultation mecha- vices delivered by the Departments of Agri- nisms of several departments. In contrast, only culture, Land Affairs, and Water Affairs and 12 percent of interviewed departments per- Forestry (PSC 2007). The survey reported that ceived that they were performing poorly on while the departments generally met clients’ consultation. needs, little consultation with clients took place. Departments cited mainly budgetary con- Its main policy recommendation was to increase straints to consultation. This constraint seems such consultations (PSC 2007). self-imposed, however, because most depart- The challenge of improving the short route ments do not budget specifically for its imple- to accountability is not unique to South Africa, mentation, monitoring, and evaluation. but a feature of many dysfunctional service sys- Departments should set and implement tems in the developing world. Merely providing practical consultation objectives and standards

54 Accountability in Public Services in South Africa and experiment with those most feasible, effec- half of provincial departments said that they The government tive, and efficient, including Citizen Satisfaction had written guidelines specifying how different should experiment Surveys, Citizens’ Forums, Izimbizo, and ser- types of complaints should be handled. Only 29 with facilitating vice delivery hearings. They should clarify con- percent of national departments and 41 percent sultation roles and responsibilities, hold people of provincial departments had set standards for community groups’ assigned these roles accountable, and improve redress. participation monitoring and evaluation. Follow-ups display a similar pattern: the in logging and PSC study found that only 29 percent of na- Redress and complaints mechanisms tional departments and 47 percent of provincial monitoring Complaining can lead to provider accountability departments reported conducting follow-ups to complaints to clients, but the scope for poor people to com- monitor the fairness of their complaints han- plain individually is narrow. In rich countries, dling system. Only 29 percent of national de- individuals get help from tort law and govern- partments and 18 percent of provincial depart- ment-sponsored ombudsmen or consumer pro- ments indicated that they have a specific system tection offices. This kind of help is rare to -non to monitor and evaluate performance on redress. existent for the poor in developing countries. The main reason for these low numbers is that In South Africa, many services operate cus- departments do not have formal systems in tomer complaints desks, typically helpful only place. Only 27 percent of all departments indi- for users that are literate, are numerate, can call cated having a system that allows external stake- a customer service center, and have a sense of holders to assess their performance on redress. their rights as a client. This list of prerequisites In a self-assessment, 39 percent of national de- excludes most users, who would need quite dif- partments and 21 percent of provincial depart- ferent mechanisms to voice their opinions. ments indicated that redress implementation is Some problems for which voice might be inadequate and requires further attention. expected to work are intractable. Take corrup- The PSC study recommends that depart- tion: the public might resent the taking of un- ments implement the model of good practice for der-the-table payments, but they have no incen- redress based on the steps outlined in the Batho tive to complain because the public service is still Pele Handbook. Departments must also identify cheaper than a private doctor. targets or standards for redress, benchmark re- A deeper constraint exists: even when there dress performance, develop indicators for mea- is an opportunity to redress complaints, moni- suring compliance with redress, and establish toring and follow-through are “public goods”­ guidelines for ensuring accountability for com- —­the benefits accrue to the entire group while plaints handling systems. the costs are borne by a few. Poor people would The government should also experiment probably benefit from government facilitation of with facilitating community groups’ participa- community groups to take on complaining and tion in logging and monitoring complaints. monitoring. The PSC also assessed the implementation Improving information flow of the redress principle of Batho Pele (PSC The Bill of Rights (chapter 2) of the Constitu- 2006). Most government departments have tion has a simple and clear vision about informa- complied partially with redress: the PSC study tion flow: “Everyone has the right of access to found that 90 percent of national departments any information held by the state.” and 84 percent of provincial departments had International experience shows that improv- some form of complaints handling system. But ing information flow can help citizens influence many of these systems are not formalized: they service providers. Public disclosure, Public Ex- do not have written guidelines, record com- penditure Tracking Surveys, Quantitative Ser- plaints, or monitor and evaluate service delivery. vice Delivery Surveys, citizen report cards, re- Standards and complaint procedures are not re- porting of absenteeism, Citizen Satisfaction viewed to determine if the departments’ redress Surveys, citizens-based budget analyses,1 service and client-care objectives are being met. Two- benchmarking, program impact assessments, thirds of national departments and more than and media strengthening all help improve

Accountability in Public Services in South Africa 55 information flow. Such mechanisms not only im- The impact was dramatic: in 2001, 82 per- prove access to services, they also create a virtu- cent of the central grants reached schools (fig- ous policy feedback loop, allowing government ure 4.1), and pupils in districts highly exposed to continually update and improve its policies. to the information campaign scored better A brief overview of four of these mechanisms­—­ on the national exam than those less exposed. Public Expenditure Tracking Surveys, citizen This inexpensive but well-targeted policy action report cards, reporting of absenteeism, and Citi- (the provision of mass information) shows how zen Satisfaction Surveys­—­is presented below. strengthening both the compact (central min- istries increasing their oversight on the sector) Public Expenditure Tracking Surveys and citizen-user power (improving information Uganda is a good example of the positive impact flows and complaints mechanisms) can be highly that improving information flow can have on effective. accountability. A 1996 Public Expenditure Tracking Survey of Ugandan local govern- Citizen report cards ments and primary schools revealed that only Several countries have used a citizen report 13 percent of the annual per-student capitation card, or citizens’ evaluation of service providers, grants from the central government made it to to establish a monitoring and sanctioning rela- schools over 1991–1995. Local officials captured tionship of accountability between citizens and the rest, and most schools received no grants at service providers. A citizen report card collects all. Financial flows were rarely audited or moni- citizens’ opinions about one particular service tored, enrollment data were unreliable, and or a range of services. This information is then schools and parents had little to no information fed back to the citizenry so that they can get a about their entitlements to the grants. sense of how the whole community views service To respond to the problem, the Ministries delivery. In theory, this improved information of Finance, Education, and Local Government flow should facilitate collective action to hold began jointly publishing data on monthly dis- the service provider more accountable and thus trict grant transfers in newspapers and broad- trigger the provider to improve service delivery. casting the data on the radio. Primary schools The basic assumption is that “naming and and district administrations were required to shaming” poor service delivery would improve post notices on all inflows of funds. This pro- it. What would argue against this assumption? moted accountability by giving schools and par- • First, citizens would rather not spend the ents access to information to understand and effort to monitor and sanction their service monitor the grant program. Parents were also providers, hoping that other mechanisms able to use a new complaints mechanism to voice would provide accountability (the “free rider” their concerns. problem). • Second, if citizens hardly use the service Figure 4.1 Share of funds reaching Ugandan schools while many alternative, private providers are Uganda shillings Percent available, they will be even less interested in 6,000 100 Average capitation grant P1–P7 (Uganda shillings) spending the effort. • Third, if citizens’ complaints do not pose a 5,000 80 credible threat, why should a provider care

4,000 about receiving a bad report card? With no 60 response from the service provider, citizens 3,000 will lose interest in the exercise. Average share of funds 40 • Fourth, elite capture may occur: elites may 2,000 reaching the schools (percent) distort the monitoring and sanctioning to Median share of funds 1,000 reaching the schools (percent) 20 their own advantage (Björkman, Reinikka, and Svensson 2006). 0 1991 1992 1993 1994 1995 2001 0 Citizen report cards have spread beyond Source: World Bank 2003. Bangalore—­ where­ they were introduced—­ to­ urban and rural services in 24 Indian states as

56 Accountability in Public Services in South Africa Box well as cities in Kenya, Mozambique, the Philip- 4.1 Citizen report cards: Bangalore and Uganda pines, Ukraine, and Vietnam. But do they work? The citizen report card on public services was created by a group of civil society Many practitioners believe so, but the evidence institutions in Bangalore, India, that were frustrated by poor service delivery. The is building slowly (Banerjee and Duflo 2006). group decided that the best way to stimulate an informed debate on the state of The latest scientific evaluation was positive. In a public services in the city was to enable users to give feedback on their experi- randomized use of citizen report cards on health ences with the services. A report card was designed asking respondents to rate service delivery institutions. The exercise produced a user evaluation of the main service delivery in Uganda, local nongovernmen- service providers, ranking each institution according to its customer satisfaction tal organizations facilitated two rounds of village rating. The survey asked about staff behavior, quality of service provided, use of and health staff meetings where health service “speed money” (bribes), and degree of responsiveness to complaints, among other 2 delivery standards were discussed. Villagers and things. The results were shared with the heads of all rated agencies and covered staff were then asked to develop a joint action extensively in the press. plan. In the short term, this intervention paid off. The report cards had a remarkable impact on public awareness of the need for However, the use of citizen report cards for improved service delivery and were instrumental in mobilizing public pressure for health providers in India warns that the posi- improvement, which in turn triggered reform in several of the agencies receiving unfavorable ratings. A follow-up survey in 1999 showed significant improvement in tive effects of such a demand-side intervention user satisfaction for most of the public service institutions. Note, however, that this may not last, underscoring a point in this report: “before-after” exercise took place during a period when Bangalore experienced introducing one accountability tool is unlikely profound economic improvement. to be effective where a systemic accountabil- Recent research by Björkman and Svensson (2007) in Uganda provides more ity problem exists (box 4.1). In this regard, two convincing evidence. They evaluated the impact of a citizen report card project in- differences between the institutional settings in tended to help rural communities hold primary healthcare providers accountable. Uganda and South Africa must be considered: The impact evaluation used a randomized field experiment approach to discover health services in Uganda are decentralized to that the report cards improved two important elements of accountability: access to information and local organizational capacity to use the information effectively. As the district level whereas health services in South communities began to monitor the provider, both the quality and quantity of health Africa are managed at the provincial level; and service provision improved. One year into the program, service use rose 16 per- in Uganda, there is a mechanism for commu- cent, weight-for-age z-scores for infants improved significantly, and deaths among nity oversight over the local health facility­—­the children under five fell 33 percent. These findings suggest that staff behavior was Health Unit Management Committee­—­while affected, and increased effort by health unit staff to serve the community improved no such mechanism exists in South Africa. the quality and quantity of health service delivery. Overall, the results suggest that community monitoring can play an important Reporting of absenteeism role in improving service delivery. What incentives do service providers have to fulfill their service obligations? Is service pro- Table vider pay affected by whether services are actu- 4.1 Absenteeism rates in developing countries (percent) ally provided? Even more basic, do service pro- Country Primary schools Health centers viders show up for work? Bangladesh 16 35 The absenteeism rate of frontline service Ecuador 14 — providers is one of the most telling indicators of India 25 40 whether the second leg of the long route to ac- Indonesia 19 40 countability is working. Citizen-user reporting Peru 11 25 on such absenteeism, and linking such reporting Uganda 27 37 to rewards and penalties, can strengthen service Unweighted average 19 35 accountability. Source: Chaudhury and others 2006. Chaudhury and others (2006) show that­—­in a sample of developing countries including Ban- gladesh, Ecuador, India, Indonesia, Peru, and show up for work on any given day, with the Uganda­—­teacher absenteeism averaged at about worst student outcomes in Bihar and Jharkhand, 19 percent and that for health workers at about where absenteeism rates are around 40 percent 35 percent (table 4.1). (table 4.2). Kremer and others (2005) provide discon- The focus group conducted in preparation for certing evidence from Indian states: on aver- this study found high teacher and nurse absen- age, almost a fourth of Indian teachers do not teeism rates in schools and clinics (Khanya-aicdd

Accountability in Public Services in South Africa 57 Table 4.2 Teacher absenteeism rates in Table 4.3 Teachers absent during an Indian states (percent) unannounced visit to randomly selected schools

State Teacher absenteeism Maharashtra 14.6 Country Percent absent Bangladesh 16 Gujarat 17.0 Ecuador 14 Madhya Pradesh 17.6 India 25 Kerala 21.2 Indonesia 19 Himachal Pradesh 21.2 Papua New Guinea* 15 Tamil Nadu 21.3 Peru 11 Haryana 21.7 Zambia* 17 Karnataka 21.7 Uganda 27 Orissa 23.4

Rajasthan 23.7 Note: All surveys were fielded in 2002 or 2003. All surveys, West Bengal 24.7 except those marked by *, report the average of two visits Andhra Pradesh 25.3 separated by several months.

Uttar Pradesh 26.3 Source: For Papua New Guinea, World Bank (2004a); for Zambia, Das and others (2005); for all others, Chaudhury Chhattisgarh 30.6 and others (2006). Uttaranchal 32.8

Assam 33.8

Punjab 34.4

Bihar 37.8 The absenteeism study of a number of coun-

Jharkhand 41.9 tries (table 4.3) also showed that good work- Delhi — ing conditions­—­particularly good school All India weighted 24.8 infrastructure­—­were associated with lower ab- senteeism, suggesting that approaches that use Source: Kremer and others 2005. both a “carrot and stick” are likely to be most effective. But evidence from India and Kenya shows that reducing absenteeism over the long 2008). Other researchers also found absentee- term can be difficult (box 4.2). ism to be significant, with the average teacher We suggest that South Africa experiment in rural areas reportedly spending fewer than 16 with mechanisms that report on service pro- hours per week on actual teaching (Chisholm vider absenteeism and link such reports to in- and others 2005). Yet all schools have School centive mechanisms for providers. A comple- Governing Bodies that are typically viewed mentary suggestion is to monitor absenteeism favorably by the parents in the focus groups­ at more aggregate levels, creating an atmo- —­parents do have a place for their voice to be sphere in which everyone understands the prob- heard. Yet these governing bodies are not effec- lem and that can spur people to increase local tive in most poor areas, and it does not appear monitoring. that they have the power to discipline teachers or have them dismissed or posted elsewhere. Citizen Satisfaction Surveys Voice without power is not sufficient to ensure The PSC has been commissioning Citizen Satis- accountability. faction Surveys, a variant of citizen report cards, How can absenteeism be reduced? National since 2002. Three surveys of more than 15,000 and provincial departments­—­or a reformed in- citizens on 34 services and 10 departments have spectorate department­—­could conduct random been conducted. Note that these surveys provide audits of their frontline staff. This increased only national averages, which preclude prov- monitoring has been a standard mechanism in inces, cities, and municipalities from compar- many developed countries and has been demon- ing data. It does allow departments to partici- strated to be effective (Duflo and Hanna 2005). pate, on the premise that the information would Having communities report absenteeism in their provide an incentive for collective action at the locality is also promising, again remembering department level. However, we do not have that positive effects of demand-side interven- information on whether these surveys pressure tions, such as a citizen report card, may not last. departments to improve.

58 Accountability in Public Services in South Africa Box The PSC concludes that “one of the key con- 4.2 Reporting absenteeism in the health and education sectors cerns of citizens was the lack of information on Banerjee, Deaton, and Duflo (2004), using evidence from a survey of 143 govern- the services provided as well as the processes for ment primary health facilities in the Udaipur district in the Indian state of Rajasthan, accessing them. The concern was that some of found that the average absenteeism rate was 36–45 percent. Duflo and Hanna the services involve complex processes of com- (2005), using evidence from another survey in Udaipur, showed that the average pleting forms and assessing citizens’ eligibility” absenteeism rates for teachers in education centers was 36 percent. The conse- quences of absenteeism are dire, often resulting in people avoiding public facilities (PSC 2007). In other words, citizens are gener- altogether. ally not aware of the services available, and they In a 2006 paper, Banerjee and Duflo describe how a member of the commu- have difficulty accessing them even when they nity in Udaipur was paid to make weekly surprise visits to health clinics to verify are aware. In addition, ineffective complaint whether assigned staff could be found at their workplaces. No sanctioning or re- mechanisms were a recurrent concern in all Citi- ward mechanisms to punish negligent workers or incentivize good performers were zen Satisfaction Surveys (PSC 2007). used; villagers were simply given the choice of how to deal with the findings of this Citizen report cards have also been in- monitoring exercise. troduced in South Africa in recent years. The While absenteeism fell in the short term, the community monitoring exercise showed that absenteeism after one year was not much different than initial levels: Human Sciences Research Council (HSRC) it was almost the same in clinics where the experiment was conducted (treatment of South Africa and the World Bank piloted a group) as in clinics where no change was introduced (control group). This suggests citizen report card in the municipality of Tsh- that community monitoring alone is not enough to have a positive impact, no matter wane (formerly Pretoria), the nation’s capital how regular and well structured the scheme might be. As Banerjee and Duflo (2006, and one of six metropolitan municipalities with p. 125) said: “One possible explanation for the lack of impact of community monitor- more than 2 million people. Clients’ evaluations ing is that everybody­—­including the service providers­—­knows that the community of services were especially important for Tsh- lacks any formal authority to reward or punish providers.” wane, which has no system to assess outcomes Evidence from Kenya shows that monitoring and incentive schemes that rely on “insiders” can have some effect. The NGO International Child Support (ICS) Kenya despite a detailed service delivery compact that introduced an incentive program for preprimary school teachers where school obligates the municipality to regularly monitor headmasters were made responsible for monitoring teacher absence. As an in- its activities. centive, a bicycle was given to teachers with good attendance records after a year. The exercise produced a user evaluation of By the end of the study period, all teachers in participating schools had al- the main services provided by the municipal- legedly attended work sufficiently to receive the prize, according to headmasters’ ity, ranked according to customer satisfaction assessments. However, in unannounced visits, ICS Kenya researchers discovered rating. The survey collected data from 1,200 no difference in absenteeism between control and treatment schools (Kremer and respondents, with a 100 percent response rate, Chen 2001). Banerjee and Duflo (2006, p. 122) suggest that headmasters inflated attendance rates either to avoid personal confrontation with their coworkers or out indicating high interest in the initiative. How- of compassion: “When human judgment is involved in a system where rules are ever, customer interest was not matched by the often bent, incentives may easily be perverted, either as in this case in an equitable government: “Several efforts were made to orga- direction or to favor some specific individuals or groups.” nize key informant interviews with the City of Intensively monitoring service providers is one way to try to increase work Tshwane officials but without success” (HSRC attendance, but monitoring alone might not be sufficient when it is not carried 2006, p. 16). Remarkably, the survey found that out independently or when it does not minimize individual discretion in monitoring only 17 percent of all residents knew about the procedures. Further, monitoring alone is not likely to yield long-term benefits when Integrated Development Plan, the government’s not combined with effective sanctions and clear incentives to make work more at- tractive to service providers (Banerjee and Duflo 2006). flagship instrument for citizen participation in the development of Tshwane. Unfortunately, municipal management has been hard to engage on the reform implications of the report card, on three racially and socioeconomically diverse and the results have not been made public. geographic areas and explored such issues as Similarly, the eThekwini (formerly Durban) trust and overall satisfaction with the services, Water and Sanitation department (EWS) col- billing and pricing, infrastructure, environment, laborated with the University of KwaZulu-Natal conflict, and health and hygiene. Results from to conduct a pilot study to assess user satisfaction the surveys were made public and all partici- with water and sanitation services. The main pants were offered the opportunity to participate aims were to improve the dialogue between in public meetings on the issues. To strengthen EWS and the local communities and to build the dialogue between local communities and trust among stakeholders. The study focused EWS, senior managers attended the meetings

Accountability in Public Services in South Africa 59 Consumer choice and answered questions and concerns (Wilson, reform implemented in Bali’s district of Jem- of providers is the Malakoana, and Gounden 2008). This pilot, ex- brana included a scheme that granted free pri- essence of client panded to six areas of the city, has proven to be mary health insurance to all enrolled citizens, an effective mechanism for recourse mainly be- granted free secondary care for poor members, power in market cause of EWS responsiveness in addressing ser- and allowed clients to choose between private transactions vice issues raised at quarterly meetings. and public clinics. The scheme improved the In conclusion, citizen report cards are an- coverage and quality of health services. Clinics, other way to strengthen service provider ac- both public and private, competed for clients and countability to citizens. They can also inform therefore had incentives to improve their client politicians and policymakers, so this tool can orientation (World Bank 2006). also be used on the long route to accountabil- Provider choice in services is often combined ity. Presumably, citizens have an informational with user fees and copayments. International evi- advantage over, say, the National Treasury in as- dence shows that user fees or copayment schemes sessing whether service providers are doing their can improve the quality and coverage of public job. Citizens also have greater incentives to im- services and increase provider accountability. In prove service delivery than do a remote adminis- Zambia, for instance, truck drivers were asked to trator or national politician. pay into a common fund used to build and main- If consumers are empowered further to di- tain roads: drivers started policing bridges to rectly influence provider behavior, through pow- keep overloaded vehicles, which would damage erful parent-teacher committees, for example, the “common good,” from crossing. Similarly, in citizens could use this knowledge to reward or the Indian state of Andhra Pradesh, when farm- sanction provider behavior. In such a context, ers started paying for water for irrigation, local citizen report cards can become a powerful ac- government accountability to the community countability mechanism. The right incentives, increased significantly: more community water combined with an information instrument (re- services are available compared with states where port card) and client power, can exert a powerful water is free (World Bank 2003). In South Af- impact on the coverage and quality of services. rica, the APRM found that segments of the The following hypotheses can be deduced population did not favor free primary education about improving service delivery from the dis- because people feared that eliminating the cost cussion above: would compromise the quality of education in • Information without client power is unlikely public schools (APRM 2007). to be an effective tool for improving account- Together, choice and payment can be pow- ability in services. erful mechanisms to strengthen service provider • On the demand side, information is not accountability. The use of free choice, user fees easily accessible, and communities do not and copayments, insurance, or vouchers does have the means to hold service providers not mean that public services must be provided accountable. by the private sector alone; public sector service • On the supply side, information does not providers can also be forced to compete to make flow. Providers are inadequately accountable them more accountable. to clients and lack responsiveness to users Choice and payment for services are only and communities. powerful when combined with free choice of providers. When choice of provider is restricted­ Provider choice and user fees or —­as when a local monopoly is present in the copayments delivery of critical services such as water and Consumer choice of providers is the essence electricity­—­the fact that users have to pay, but of client power in market transactions. Under cannot walk away, only provides incentives for certain conditions, consumer choice can also users to organize and pressure the service pro- be introduced into the provision of basic ser- vider. Such incentives can be strong or weak, vices. Bali is a good example of how expand- depending on the specific arrangements for user ing citizens’ choice of provider can enhance representation in oversight committees and how service coverage and quality. One health sector much they can influence service providers.

60 Accountability in Public Services in South Africa User fees and copayments often collide with Participatory local government User fees and equality considerations, especially where much budgeting and development planning copayments of the population is too poor to afford payments. The South African Constitution, the Develop- often collide The constitutional right of access to adequate ment Facilitation Act (1993), the White Paper housing, healthcare services, sufficient food on Local Government (1998), and the Munici- with equality and water, social security, and education imply pal Systems Act (2000) all support public par- considerations, that user fees and copayments cannot hinder ac- ticipation in policymaking and program imple- especially where cess. In practice, these constitutional impera- mentation, but they do not specify the concrete tives have led to free services for the poor: water mechanisms to facilitate such participation much of the and power, housing for low-income households, (Smith and Vawda 2003). population is too health care for pregnant women and children up Integrated Development Plans (IDPs)­—­ poor to afford to five years old, and primary education and a the principal planning instruments at the local school nutrition program in areas below a prede- level­—­are main programs for citizen participa- payments termined poverty index. tion and input in policymaking. The Local Gov- The Presidency identified the need for com- ernment Transition Amendment Act (1996) in- munities to pay for services beyond free basic troduced IDPs so that local governments could services in Towards a Ten Year Review. The main reflect the public’s priorities in the municipal challenge for the future is to “promote national budget. IDP consultations with communities identity and pride, ensuring that households de- are intended to identify needs, target service de- velop civic responsibility, including a culture of livery to those needs, allocate the budget, and paying for services (beyond the free basic ser- measure government performance in land use, vices) and taking responsibility for protecting transport, and infrastructure and economic de- the infrastructure in their communities” (PCAS velopment planning. The IDPs emerged in ear- 2003, p. 115). nest only after 2000 when the “wall-to-wall” de- Even when services are to be provided for marcation of municipalities integrated formerly free, or for a small fee, a number of developed white and black areas across the country. countries have devised methods to introduce A tool for policymakers to elicit citizen pref- user choice of providers. In health, France, Swit- erences and feedback and participatory budget- zerland, and the Netherlands, among others, ing and development clearly fit into both the have publicly funded or subsidized health insur- long and short route to accountability. In South ance. In education, some U.S. cities and Chile Africa, moreover, citizens view IDP consulta- have used school vouchers. tions as an opportunity to complain about ser- Copayment can also increase incentives for vices and influence service provider behavior­—­ lower spheres of government to use national gov- the short route to accountability. No attempt to ernmental fiscal transfers more economically. resolve this ambiguity will be made here. The water and sanitation chapter in this study, for The IDPs have several limitations. First, example, attributes wasteful municipal spending they are five-year strategic visions without con- to ample funding provided entirely as grants. In a crete mechanisms to influence annual budgeting copayment system, the fiscal grant from the cen- and priority setting. Second, although popular tral government becomes a matching grant, and participation in the IDP process is mandatory, the provincial or local government has to finance in reality such participation is limited (UNDP its matching contribution from its own revenue. 2002). A citizen report card study we undertook Alternatively, a local government could finance in Tshwane showed that only 17 percent of resi- its matching portion from a general purpose dents in the selected townships had even heard grant such as Equitable Share, making this grant of IDPs. While the consultation mechanisms in an “opportunity cost”­—­if spent it cannot be used the IDP process allow for voice, participation is for another competing purpose. not widespread enough to make planners truly The absence of choice of provider, user fees accountable to citizens­—­that is, the short route for basic services, and effective citizen oversight to accountability is underdeveloped. committees has deprived especially the poorest In practice, officials drive the IDP process­ citizens of power over providers. —­rather than councilors or citizens­—­without

Accountability in Public Services in South Africa 61 Community-driven meaningful community participation. IDP con- Information materials outlining the IDP and development sultations often became simply an item to check setting the stage for public participation are is a useful off on the “. . . technocratic checklist for meeting seldom easy to digest or in local languages, the municipal budget cycle” (Smith and Vawda and IDP documents are written in technical complement to 2003, pp. 32–33). Councilors are “informed” English “inaccessible” and “too high-flow- decentralization about activities, not involved in decisionmaking. ing.” In addition, notices for meetings are “In practice, participation has largely been re- not widely distributed, meetings happen on stricted to input-gathering exercises of IDP short notice and at inconvenient times and processes with little effectiveness, because com- places, facilitation of meetings is often poor, munities are not compelled to explore and re- and the outcomes targeted are not clear. In solve competing demands and trade-offs within many municipalities, reports following the themselves and more especially between com- meetings are often not created. munities with different needs” (Pieterse and • Fourth, citizens participate as individuals: others 2008, p. 15). civil society organizations are strikingly ab- In a review of IDP participation in Gauteng sent from the IDP process. However, “[re- Province, Marais, Everatt, and Dube (2007) search] suggests that where civil society conclude that: organizations participate more fully, service- • First, public knowledge of IDPs is low. This delivery gaps are better identified. In such is not a Gauteng-specific phenomenon: even cases, prioritization of service-delivery needs in the Integrated Sustainable Rural Devel- is also more appropriate to the needs of the opment Programme and Urban Renewal communities” (PCAS 2003, p. 14). Programme­—­the focal points for integrated • Fifth, IDP participation remains only con- and coordinated poverty reduction­—­only sultative: it is not connected to budget re- one in ten residents had heard of the IDP. source devolution and lacks joint budget- In Alexandra (a township in Gauteng), only ing between departments. Williams (2005, 7 percent of residents had heard of the IDP. p. 22) confirms this point, saying that IDP However, nine in ten citizens participated community participation in Cape Town is once they were informed. This pattern has “largely a ceremonial exercise and not a sys- been found throughout the country: partici- tematic engagement of communities to in- pation is very high once people know about fluence the Development and Service Deliv- the IDP. Thus, lack of IDP participation is ery Programs.” due to a lack of public knowledge not to a The Presidency, in Towards a Ten Tear Re- lack of public interest. view, concludes: “Although the IDPs are a form • Second, interpretation of the purpose of the of local consultation that seeks to involve local IDP differs between government officials communities in identifying their needs within and citizen-users. Government officials see a given locality, there remains a need to involve the IDP as strategic rather than operational; communities more in decision-making espe- they often view public participation as merely cially in poor communities” (PCAS 2003, p. 14). a legal requirement and simply seek public Other observers agree that the evidence suggests agreement on the plan. Friedman (2006) that South African citizens do not have suffi- cites instances where citizens were asked to cient influence over politicians and policymakers comment on IDP resource allocations that to express their priorities and shape policies (Van administrators had already agreed on, while Donk and others 2008). The PSC concludes that ward councilors had not even seen the draft. “[t]he acceleration of participatory mechanisms Citizens, by contrast, see the IDP meeting is a pressing requirement” (PSC 2008a, p. 94). as a forum to raise demands and complaints about service delivery. In other words, offi- Community-driven development cials see the IDP as “their” plan but citizens Community-driven development is a useful feel the IDP should reflect all needs. complement to decentralization; it is a low-cost • Third, the IDP process in Gauteng suf- and rapid method to implement small-scale fers from major flaws in information flow. local development projects. Community-driven

62 Accountability in Public Services in South Africa development refers to a set of practices that seek • Empowering local authorities. The short route to give communities control of decisions and • Realigning the center to support local and to accountability resources. The communities plan, implement, community-driven development. is the weakest manage, and maintain their projects or services • Promoting accountability. themselves, often with the help of coproducers • Building capacity. link in the service such as local governments, facilitators, and sec- Case studies of South African community- delivery triangle tor specialists. driven development since 1994 (Everatt and Community-driven development has been Gwagwa 2005) suggest that the ideal legal, fis- introduced successfully in many countries, most cal, and policy environment exists but that it is often in association with the World Bank or considerably underused. Some promising com- other donor programs. It has been most success- munity-driven development initiatives during ful in small-scale infrastructure projects such as the 1990s, such as by Mvula Trust in water pro- the building of schools and classrooms. Success- vision, have even been abandoned. ful community-driven development projects also The short route to accountability is the exist in health care. In most of these programs, weakest link in the service delivery triangle, not only do communities control planning and due in part to the disappearance of community- contracting, but money for the investment (less driven development from the South African de- their copayment in cash or in kind) is transferred velopment landscape. Community-driven de- directly to community accounts. velopment has substantial potential to improve Communities function in a complex insti- services that are not prone to economies of tutional environment. They require skills and scale, that do not require uncommon skills, and leaders that must also be held accountable by that can be monitored easily by local oversight community members and higher authorities. committees. Community-driven development cannot func- tion unless the environment is conducive. It Notes functions best, as in Indonesia or Burkina Faso, 1. Several studies of participatory budgeting in as part of a broader decentralization framework­ South Africa have already been conducted, —­we call this local and community-driven de- including one funded by the World Bank velopment. Local and community-driven de- Institute. velopment is encouraged by focusing on five 2. See Björkman and Svensson (2007). About simultaneous processes of change: 5,000 households from 50 communities from • Empowering communities. nine districts in Uganda were sampled.

Accountability in Public Services in South Africa 63

5 Participatory Assessment of Service Delivery

The purpose of this participatory assessment was of Botshabelo and Thaba Nchu. It also includes to assess the “voice of the people” on service deliv- a large rural area with commercial farming and ery issues and to study differences in perceptions former Bantustans, where rural-to-urban migra- of accountability among citizen-users, service tion is considerable. providers, and politicians (the local councilors). Maluti-A-Phofung is a large rural municipal- ity with three main service centers­—­Harrismith, The municipalities Kestell, and QwaQwa (a former homeland)­—­ This study used participatory methods to elicit and Phuthaditjhaba as the municipal center. perceptions on housing, health, agriculture, QwaQwa has declined economically in recent water and sanitation, education, and social years due to the closing of many industries, re- grants. These municipalities were selected to sulting in more than 50 percent unemployment obtain a large variation in size, rural/urban mix, in 2000, the highest in the sample. and economic base and dynamics (table 5.1). Naledi is the least populated municipality Mangaung has the largest municipal popu- in the sample and includes three small towns, lation. It includes the thriving service city of which serve the surrounding commercial farms Bloemfontein and the smaller, declining towns that employ many farm workers.

Table 5.1 Characteristics of the six municipalities

Issue Mangaung Maluti-A-Phofung Naledi Greater Tzaneen Bela-Bela Sol Plaatje Province Free State Free State Free State Limpopo Limpopo Northern Cape

Size (square kilometers) 6,283 4,421 3,412 3,240 4,000 18,771

Population 645,438 400,000 (2000) 27,471 375,584 52,124 (2001) 201,464 (2001)

Urban/rural Metropolitan, small rural, Large rural, former Small rural, only small Urban and rural, large Small urban serving Urban, mining, medium small former homeland, homeland, mostly townships townships townships commercial farms, small townships large townships township

Main economic activities Services, agriculture Agriculture, tourism potential Commercial agriculture, Services and commercial Commercial agriculture, Mining, services, nascent extensive livestock agriculture game farming, tourism tourism

Economic growth and Growing rapidly Declining with out-migration Declining with out-migration Fastest growing in Limpopo Growing because of tourism Mining in decline, but most conditions vibrant

Former Bantustan? Yes Yes No No Yes No

Percent white 12 1 1 2 17 13

Percent unemployed 40 51 37 41 30 41

Source: Government of South Africa and authors’ calculations.

Accountability in Public Services in South Africa 65 Greater Tzaneen is the third largest mu- The methodology nicipality, and includes Tzaneen, 3 other rural The main quantitative and qualitative research towns, and 125 villages. The Tzaneen economy methodologies included: depends largely on commercial agriculture, • Participatory tools and methods such as with rural communities engaged in subsistence Venn diagrams, accessibility score cards, ser- agriculture. vice provider performance score cards, and Bela-Bela has one of the smallest munici- long-term trend diagrams. pal populations in South Africa, but it covers • Focus groups using semi-structured inter- a large geographic area of 4,000 square kilo- views and checklists during and after par- meters. Commercial agriculture and game ticipatory methodologies. farming (or ecotourism) are the main economic • Questionnaires followed by focus group activities, but it has a dynamic tourism sector discussions. and the lowest unemployment rate in the sam- • Workshops with service providers and ple. It also contains areas derived from former councilors. homelands. The Venn diagrams and accessibility score Sol Plaatje is the fourth most populous mu- cards, prepared by the communities, were trans- nicipality, but with an additional eight smaller lated into quantitative scores of community per- towns, it is geographically the largest in the ception of accessibility of services (table 5.2). sample. Kimberley is an important diamond Qualitative methods also rendered a long list mining center, largely responsible for the eco- of comments from community members, politi- nomic growth of the municipality. cians, and service providers in all six municipali- Are perceptions of service accountability ties. The lists were used to assess perceptions of affected by the different characteristics of the the quality of each of the services: the number of municipalities? We will compare the densely positive comments made about the services (the populated Mangaung, Maluti-A-Phofung, and compliments) and the number of negative com- Tzaneen with the sparsely populated Bela-Bela ments (complaints) are summarized in table 5.3. and Naledi to consider the influence of size on Issues raised repeatedly were counted only once, services and accountability. We will also com- so the scores do not reflect the full intensity of pare Mangaung and Sol Plaatje­—­municipalities feelings on some issues. dominated by a city­—­with the predominantly The same user comments were analyzed a rural Maluti-A-Phofung, Naledi, and Bela-Bela second time to locate the nature of the account- to study how perceptions of service account- ability issue causing complaints. The comments ability differ by geography. Finally, we will were allocated to a leg in the accountability compare Mangaung, Tzaneen, and Bela-Bela­ model: for example, citizen complaints about —municipalities­ experiencing growth—­ with­ never seeing councilors were allocated to the Maluti–A-Phofung, Naledi, and Sol Plaatje­—­ accountability between users and politicians/ municipalities that are economically stagnant or policymakers (leg 1); complaints about poor declining. staff attitudes were counted as inadequacy in the

Table 5.2 Satisfaction with access to services across municipalities (percent)

Sector Mangaung Maluti-A-Phofung Naledi Greater Tzaneen Bela-Bela Sol Plaatje Overall Education 80 100 100 40 60 60 73

Health 20 33 40 20 20 40 29

Water and sanitation 80 66 100 60 100 97 84

Average of three 60 66 80 40 60 66 62

Social grants 80 N/A 60 100 100 60 80

Housing 80 N/A 20 60 30 80 54

Average of five 68 N/A 64 56 62 68 64

N/A is not available. Source: Authors’ calculations.

66 Accountability in Public Services in South Africa Table 5.3 Complaints and compliments about quality of services

Municipality Housing Education Water and sanitation Health Social grants Total Naledi 2 5 3 8 1 19

Bela-Bela 3 4 2 3 3 15

Tzaneen 1 1 2 0 0 4

Mangaung 0 1 0 6 1 8

Maluti-A-Phofung 0 2 1 10 2 15

Sol Plaatje 2 2 1 6 2 13

Complaints 8 15 9 33 9 74

Compliments 1 5 9 3 9 27

Complaints (percent) 75 50 92 50 73 89

Source: Authors’ calculations.

short route to accountability between citizens Satisfaction with access to “water and sani- and service providers (leg 3); and absenteeism tation” and to “social grants” rank the highest complaints were counted as a weakness between among the sectors, both receiving less than a policymakers and providers (leg 2), as well as third of the complaints that health received. between citizens and service providers. (Absen- But complaints equal compliments in water teeism could be remedied by action in either or and sanitation and in social grants, suggest- both legs.) The same methods were used to allo- ing that quality remains a major issue in both. cate the comments of service providers and poli- The study also finds challenges in water qual- ticians. (Note that senior researchers sometimes ity accessibility: farm workers continue to re- had to make subjective judgments in allotting ceive poor quality water, and many rural areas comments.) in municipalities and informal settlements have no access to water at all. Sanitation ser- Results by service vices are generally poor: communities still use Table 5.2 shows community satisfaction with the bucket system in many municipalities­—­ access to five services delivered across the six particularly in rural areas­—­and toilets block municipalities. We coded accumulated responses easily, while bucket collectors are often corrupt by color: a score of 39 percent or less is in red, a or late. score between 40 percent and 69 percent is in Satisfaction with access to “education” ranks yellow, and a score of 70 percent or greater is in third, with a score of 73, but quality is a big green. problem with 15 complaints against only 3 com- Consistent with the Public Service Com- pliments. Complaints include: mission’s national user satisfaction surveys (PSC • Lack of teacher discipline. 2007, 2008a), “health” ranks the lowest across • Teacher absenteeism and drunkenness. all municipalities in satisfaction with access and • Fraternization between teachers and stu- quality, with by far the most complaints and very dents, undermining the boundaries that pro- few compliments. Among the most frequent mote respect for teachers, managers, and the complaints in the health sector were: education institution. • Personnel are uncaring, often late, and disre- • Lack of student materials and poor quality spectful of opening hours. of available materials. (Pupils often share • Clinics are often understaffed and poorly textbooks.) managed. • Poor quality of school infrastructure. • Staff attitudes toward patients imply that • Bad school management and administration. service is a privilege rather than a right. Satisfaction with access to “housing” ranks • Staff is often viewed as unresponsive to pa- fourth, with a score of 54. The extremely low tient needs. scores in Naledi and Bela-Bela weigh down the • Information confidentiality is rarely average, suggesting that rural municipalities are respected. at a disadvantage in access to housing. While

Accountability in Public Services in South Africa 67 living in a Box 5.1 Tzaneen politicians’ housing had the lowest number of complaints, views of their role rural area is not with eight, it also received only one compli- • To involve the youth in development and uniformly an ment, suggesting that problems in housing may be worse than in water and sanitation and social participation. • To ensure integration of services in all impediment to grants, which received about the same number of approaches. service access complaints and compliments. • To ensure that communities are consulted Water and sanitation is the most decentral- and their inputs are reflected in the municipal ized service, with responsibility residing largely documents. with the municipal government. This sector has • To mediate between the municipality and the the most satisfaction with access and received community. few complaints, possibly due to the close prox- • To ensure that communities get services as imity of the decisionmakers to users. planned in the Integrated Development Plans. • To ensure that service delivery responds to In contrast with water and sanitation, social needs. grants are disbursed from a central agency but • To ensure that service delivery is not sabo- with highly deconcentrated service access points­ taged and that people get the services they —­economies of scale in disbursement are very deserve. high. Eligibility, however, is determined by mu- • To maintain or strengthen the livelihoods of the nicipal social workers reporting to the province­ people in the wards. —­no economies of scale are present. This is a • To educate community members on their rights good example of coproduction­—­responsibilities and responsibilities to services. • To ensure that accurate service delivery infor- are allocated to different coproducers accord- mation is disseminated to the communities. ing to differences in scale economies. Satisfac- • To ensure that basic needs are responded to tion with access to social grants was almost as as per the mandate of the national government. high as that for water and sanitation and had about the same number of complaints, but so- cial grants received more compliments. Clearly, that rurality may reduce quality of services rather decentralization alone cannot explain per- than access to services. ceptions of accountability. Coproduction also Naledi, the least populated and most rural seems to add to the complexity of accountability municipality, has the greatest satisfaction with relationships. access (80 percent) but the highest proportion of quality complaints (19 percent)­—­and therefore Results by municipality the highest accountability challenges. Tzaneen Because satisfaction-with-access scores were provides an interesting contrast in that it has not available for social grants and housing, we the least satisfaction with access but the most cannot calculate an average score for all five satisfaction with quality. Politicians in Tzaneen services. We begin instead by comparing the have a particularly impressive commitment to average scores for the three services for which service delivery­—­much more so than elsewhere­ we had complete data­—­education, health, and —­which may explain this curious finding water and sanitation. (box 5.1). The data suggest that living in a rural area is Access to services is also unrelated to mu- not uniformly an impediment to service access: nicipal population: Mangaung does no bet- Tzaneen­—­a mixed rural-urban municipality­ ter than Naledi and Bela-Bela, the least popu- —­is less satisfied with access than are Maluti-A- lated municipalities, on satisfaction with access. Phofung, Naledi, and Bela-Bela­—­all rural mu- Meanwhile, the urban areas of Mangaung and nicipalities. The same conclusion can be drawn Tzaneen have the fewest complaints about qual- when the five services are compared across the ity of services. municipalities not including Maluti-A-Pho- In conclusion, satisfaction with access to ser- fung. But total complaints across all services vices does not systematically vary by population are highest in the three rural municipalities, and rurality, but urban municipalities may have while Tzaneen has the lowest number of com- fewer service quality challenges than do rural plaints, followed by Mangaung. This suggests areas.

68 Accountability in Public Services in South Africa Perceptions of communities Table 5.4 Total adverse comments across municipalities, by accountability leg We have already discussed how the comments of communities were translated into “weights” Number of adverse comments allotted to the three different accountability Municipality Leg 1 Leg 2 Leg 3 Total legs. In figure 5.1, weights are represented by the Naledi 0 12 12 24 heights of the bars, but the average weight across Bela-Bela 1 14 9 24 Tzaneen 2 9 5 16 the three legs is the same for all the services. The Mangaung 2 14 3 19 bars reflect only the number of times the prob- Maluti-A-Phofung 4 5 7 16 lems were allocated to the legs, not differences Sol Plaatje 1 6 4 11 in perceptions about the seriousness of problems. Total 10 60 40 110 Community dissatisfaction with account- Source: Authors’ calculations. ability between service providers and politicians/ policymakers (the compact, or leg 2 of the long route to accountability) is high in all services, Figure varying from 40 to 50 percent. Dissatisfaction 5.1 Scores of accountability, by service and accountability leg with accountability between service providers Weighted scores Housing Education Water Sanitation Health Social grants and users (leg 3) is slightly lower, on average, 60 and varies from 30 to 50 percent. Community dissatisfaction with political ac- 50 countability (citizen voice, or leg 1 of the long 40 route to accountability) is much lower, varying from only 5 percent for water to about 18 percent 30 for health. (Note that we counted community dissatisfaction with both policies and politicians/ 20 policymakers.) Dissatisfaction with policies ap- 10 pears minimal and political representation well developed, inconsistent with other findings on 0 Leg 1 Leg 2 Leg 3 leg 1 of accountability and suggesting that local accountability problems are fewer when com- Source: Authors’ calculations. pared with provincial and national problems. Dissatisfaction in the health sector again stands out: politicians are seen as fairly account- Maluti-A-Phofung. Similarly, rural municipali- able to citizens, but health service providers’ ac- ties do not face greater challenges with the com- countability to citizens is the worst across all pact than do urban municipalities. sectors. In water and sanitation and in social grants, the compact between politicians and ser- Perceptions of service providers vice providers is seen as weaker than is account- Politicians and service providers were inter- ability to citizens. viewed in four municipalities. (Data for Man- Table 5.4 shows that Sol Plaatje­—­a fairly gaung and Naledi are missing.) Politicians and populated municipality with a large mining sec- service providers were asked to comment on tor and a strong opposition party­—­has the few- accountability across all services in their munici- est adverse comments, while the least populated palities rather than to rate services separately. municipalities of Naledi and Bela-Bela have the Figure 5.2 shows service provider perceptions of most. The most populated municipalities, Man- accountability across municipalities; figure 5.3 gaung and Maluti-A-Phofung, have scores in the shows politician perceptions. middle. Municipal population alone, therefore, is Like communities, local service providers see not enough to explain accountability problems. few problems with political accountability. But Looking at total adverse comments, we unlike communities, service providers see few can also conclude that rural municipalities do accountability problems, on average, between not face greater accountability challenges than themselves and users (except in Bela-Bela). Ser- do the urban municipalities of Mangaung and vice providers link almost all accountability

Accountability in Public Services in South Africa 69 Figure 5.2 Accountability challenges perceived by service providers problems to state capacity­—­to accountability be- (percent) tween themselves and politicians/policymakers. Politicians’ views of accountability are more Percent Sol Plaatje Bela-Bela Maluti-A-Phofung Tzaneen 70 similar to those of service providers than those of communities: problems with accountability are 60 seen mostly as stemming from the relationship

50 between service providers and politicians/policy- makers. Politicians do not, on average, believe 40 that there are major problems with accountability 30 between service providers and the communities. Politicians’ views of political accountability to 20 citizens vary greatly across municipalities: politi- 10 cians see no political accountability problems in Sol

0 Plaatje, for instance, while 40 percent see political Leg 1 Leg 2 Leg 3 accountability as a problem in Bela-Bela. (Note that Sol Plaatje had the most opposition-party in- Source: Authors’ calculations. terview participants since the opposition party is strongest in Sol Plaatje, where politicians believe that state capacity problems are most severe.) Figure 5.3 Accountability challenges perceived by politicians (percent) Conclusion

Percent Sol Plaatje Bela-Bela Naledi Maluti-A-Phofung Mangaung Tzaneen Two main conclusions about local perceptions of 90 accountability can be drawn: 80 • First, all stakeholders identify serious issues

70 with the state compact­—­the accountability

60 link between politicians/policymakers and service providers. 50 • Second, opinions on client power­—­the link 40 between service providers and citizen-users­ 30 —­are deeply divided: citizen-users see a 20 major lack of service provider accountability 10 to them, while service providers and politi- 0 cians do not see this as a major problem. Leg 1 Leg 2 Leg 3 This participatory assessment of accountabil-

Source: Authors’ calculations. ity in services reveals a major difference in percep- tions of the failure of the short route to account- ability, consistent with the theme of this report and the findings of several other assessments, in- cluding that of the Public Service Commission.

70 Accountability in Public Services in South Africa 6

Education

The success of any education system depends Change and challenges on productive classroom relationships between South Africa’s education system has been radi- educators and students. But education is dis- cally transformed since apartheid. The apartheid cretionary, transaction-intensive, and difficult system had different curricula for different racial to monitor. Teaching and learning require groups and poor quality education for Africans, many decisions, with educators constantly Coloreds, and Indians,2 with the attendant neg- seeking the most appropriate way to convey a ative impacts on education outcomes for these variety of material and ensure it is absorbed groups.3 Education administration was spread by a variety of learners.1 Educators must across as many as 19 departments serving differ- constantly decide how to adapt the content, ent people during the apartheid era. structure, and pace of their lessons to maxi- Today, education administration has been mize student learning. These decisions hap- condensed into two national departments and pen numerous times at numerous locations. It nine provincial departments. This system over- is impossible­—­and probably undesirable­—­to sees education services for about 6.8 million stu- constantly monitor interactions between edu- dents in grades 1–6 and almost 2.9 million stu- cators and students. dents in grades 7–9, about 97 percent of which It is difficult to establish accountability are in public schools.4 While the providers are in this situation. Tight bureaucratic control, predominantly public, many of these schools which might work for routine, verifiable tasks, charge fees and raise voluntary contributions erodes and becomes ineffective. Strengthen- from students’ families, though a no-fee policy ing the role of education services improves the is now in effect for schools in the poorest two short route to accountability; ensuring infor- quintiles. mation flow and appropriate compacts between policymakers and service providers supports Learning outcomes the long route. How does South Africa’s education system per- Information on policies to establish account- form on learning outcomes? International com- ability is growing, but large gaps remain. This parisons show that substantial challenges to the chapter describes information compiled from development of a high-quality education system existing sources, proposes hypotheses and rec- remain: South Africa scores below its peers. ommendations on how to improve education South Africa participated in the Trends service delivery, and discusses some areas where in International Math and Science Study additional research might be useful. (TIMSS), which administered math and science

Accountability in Public Services in South Africa 71 assessments to grade 8 students, and the Prog- 2006 for PIRLS in figure 6.2.) For example, in ress in International Reading Literacy Study 2003 national income in Chile was US$9,851 (PIRLS), which administered an assessment of per capita (in purchasing-power-parity dol- reading ability to students in grade 5.5 South Af- lars) and in Malaysia US$9,000, compared with rica has a low average score even compared with US$10,364 in South Africa. Average test scores other countries with similar income. (Results for in the three countries were 387, 508, and 263, 2003 for TIMSS are shown in figure 6.1 and for respectively. So South Africa performs more than one standard deviation below both Chile Figure 6.1 Average grade 8 Trends in International Math and Science 6 Study math test score and difference between poorest and and Malaysia. Likewise, while national per richest quintile, with share of difference due to within- and capita income in Morocco was about half that between-school differences of South Africa in 2006 (again, in purchasing

Test score Between schools Within schools Average power parity), its average performance on the 700 PIRLS test was slightly higher, and the distribu- tion of scores across test takers was substantially 600 more equal. In an older regional comparative learning 500 assessment, South Africa performs even worse

400 than some African neighbors, several of which have substantially lower national income. In 300 2000, the Southern African Consortium for Monitoring Educational Quality (SACMEQ ) 200 tested samples of grade 6 students in 13 South- Iran Korea Chile Egypt Syria Estonia Latvia Jordan Hungary Lithuania Malaysia RomaniaBulgaria Lebanon Morocco IndonesiaMoldova Botswana Macedonia Philippines ern African countries. Only Seychelles and United States Saudi Arabia South Africa Slovak Republic Russian Federation Mauritius had higher national per capita income

Note: The top of each bar is the average score among test takers in the richest quintile; the than South Africa. Similar to the TIMSS and bottom of each bar is the average score among test takers in the poorest quintile. Quintiles PIRLS results, students in South Africa per- are derived from an index of household ownership of various consumer durables. Countries are ordered by purchasing power parity gross national income per capita. formed poorly (figure 6.3): math scores averaged Source: Authors’ analysis of TIMMS data. 486 and reading scores 492, compared with 563 and 546 in Kenya, for example, whose national Figure 6.2 Average grade 5 Progress in International Reading Literacy per capita income (purchasing-power adjusted) is Study reading test score and difference between poorest and about one-tenth that of South Africa.7 richest quintile, with share of difference due to within- and between-school differences South Africa also has a very large perfor- mance gap in the socioeconomic status of the Test score Between schools Within schools Average family of the test taker­—­larger than all other 700 TIMMS, PIRLS, and SACMEQ participant

600 countries. Most of the difference in learning is across schools: the share accounted for by 500 within-school differences is smaller than that in many other countries. For example, 13 percent 400 of the difference between the richest and poorest quintile in TIMSS was accounted for by within- 300 school differences in South Africa in 2003, com- 200 pared with 19 percent in Chile and 23 percent in

Poland Latvia the Russian Federation. Hungary Bulgaria Morocco Georgia Moldova Singapore Lithuania Romania Indonesia United States South Africa Slovak Republic Macedonia, FYR Iran, Islamic Rep. More dramatically, 23 percent of the differ- Russian Federation Trinidad and Tobago ence between the richest and poorest quintiles Note: The top of each bar is the average score among test takers in the richest quintile; the bottom of each bar is the average score among test takers in the poorest quintile. Quintiles on SACMEQ math scores was accounted for are derived from an index of household ownership of various consumer durables. Countries by within-school differences in South Africa, are ordered by purchasing power parity gross national income per capita. The grade 4 PIRLS data correspond with grade 5 in the South African education system. while 62 percent of the difference was within- Source: Authors’ analysis of PIRLS data. school in Botswana. This suggests that dif- ferences in student performance across South

72 Accountability in Public Services in South Africa Figure 6.3 Difference between poorest and richest quintile in Southern Africa are attributable to school- or learner- Africa Consortium for Monitoring Educational Quality math specific characteristics that differ systematically and reading test score, with share of difference due to across schools. within- and between-school differences Public spending Test score Between schools Within schools Average 700 Math The TIMSS and SACMEQ results are several years old, and some of the country compari- 600 sons might be outdated. New data will emerge 500 from South Africa’s participation in the 2007 400 8 and 2012 rounds of SACMEQ data collection. 300 Nevertheless, the consistency across these inter- 200 national learning assessments suggests systemic 700 Reading problems in South African education. How can education outcomes be improved? 600 The main instruments for improvement are eco- 500 nomic growth and public spending on educa- 400 tion. Growth stimulates private and public re- 300 sources for education, the demand for skills, the 200 economic return of education, and ultimately, Kenya Malawi Mauritus Namibia Lesotho Uganda Zambia Tanzania Seychelles Botswana Swaziland the derived demand for schooling and learning. South Africa Mozambique As discussed in the early chapters of this re- Note: The top of each bar is the average score among test takers in the richest quintile; the port, public spending on basic education is not bottom of each bar is the average score among test takers in the poorest quintile. Quintiles the major concern in South Africa: real expen- are derived from an index of household ownership of various consumer durables. Countries are ordered by purchasing power parity gross national income per capita. ditures on education rose 35 percent between Source: Authors’ analysis of SACMEQ data. 1994 and 2006. Education spending grew from about 5.4 percent of GDP in 1985 to 6.4 percent Figure 6.4 Expenditure incidence of primary and secondary schools, in 1994, before falling to 5.5 percent in 2006. public spending over time in South Africa and selected Despite falling as a share of per capita GDP developing countries between 1990 and 2004, the latter part of this Percent of period saw a more than 50 percent increase in expenditure incidence Poorest quintile Quintile 2 Quintile 3 Quintile 4 Richest quintile real spending per learner, from ZAR4,561 in 50 South Africa 9 1999/2000 to ZAR6,895 in 2006/2007. 40 A major shift in the allocation of education 30 spending in favor of the poor has occurred over 20 the past 10 years. Public spending on education across income quintiles­—­once targeted nar- 10 rowly to wealthier white households­—­has be- 0 1993 1995 2000 2005 come much more equitable, and compares fa- 50 Comparator countries vorably with spending in other countries (figure 40 6.4).10 Much of this shift has been across racial 30 groups: the gap in spending per white student and spending per black student fell from 240 20 percent in 1995 to 122 percent in 2005 (van der 10 Berg and Louw 2007). 0 Brazil 1997 Indonesia 1998 Mexico 1997 Thailand 2002 Resources have shifted considerably toward Source: For South Africa, van der Berg and Louw (2007); for Brazil, Indonesia, Mexico, and poor students and poor schools. The spending Thailand, Filmer (2003) and updates. differential per child between the richest school quintile and poorest quintile is small: students in the richest school quintile receive ZAR5,284 This small difference has two main sources: per year, and students in the poorest school teacher pay in the richer quintiles is still higher quintile receive ZAR4,757 per year (table 6.1).11 than in the poorer quintiles, largely because

Accountability in Public Services in South Africa 73 Table 6.1 Spending per student by type of spending and school quintile, 2005 (ZAR per student) the groups remain. South Africa has some of the largest testing gaps in the world between School Teacher Nonteacher Norms and Total public the richest and poorest quintiles (see figures 6.1 quintile pay pay standards Top-slice current and 6.2). 1 3,955 90 411 301 4,757

2 4,076 115 325 283 4,800 The following exercise illustrates how incre-

3 4,088 188 253 287 4,817 mental public spending is associated with in-

4 4,170 265 197 272 4,904 cremental outcomes. After normalizing the test 5 4,528 428 112 215 5,284 scores from the TIMSS, PIRLS, and SACMEQ Total 4,103 202 273 276 4,854 testing exercises so that the mean for South Af- rica on each test is 500 and the standard devia- Source: Gustafsson and Patel 2006. tion is 100, we estimate the average score for each quintile. This average score is then divided by av- teachers are more qualified and more experi- erage spending per student for that quintile. The enced in richer schools, and almost five times ratio derived approximates the average test score as much (ZAR428) was spent per child on non- per Rand spent on a child in each quintile. After teacher employees in the richest school quintile indexing the ratio to 1 for the poorest quintile, than in the poorest school quintile (ZAR90), figure 6.5 shows the ratios for all quintiles. largely because of more nonteacher employees in Figure 6.5 indicates that each Rand spent in richer schools (administrative and support staff, the richest quintile is associated with about 20 such as gardeners). Despite pro-poor spending percent more learning than each Rand spent in targets using the Norms and Standards formula the poorest quintile. The ratio of spending to test and the so-called “top-slice” (allocations for scores is highly nonlinear: virtually flat across nonpersonnel recurrent spending), the difference the first 3 quintiles (60 percent of the popula- between teacher and nonteacher pay represents tion), increasing slightly in the fourth quintile, most of the remaining gap in spending per child and increasing sharply in the richest. The ratio between rich and poor quintiles. of spending to test scores grows at an increasing rate in the richest quintile.12 Public spending and learning The ratio of spending to test scores in each outcomes: an association quintile cannot be interpreted as a pure reflec- The previous discussion focuses exclusively tion of the productivity of that expenditure: as on expenditure incidence­—­how much public discussed above, spending has been increasingly spending is allocated across quintiles or groups. allocated to precisely where it is needed­—­poorer Despite increased spending on the poor, enor- students­—­which would induce an overall nega- mous differences in learning achievement across tive impact on spending and outcomes. In ad- dition, quality improvement associated with ex- Figure 6.5 Efficiency index of public spending on education in each penditure takes time to materialize. quintile Nevertheless, without more analysis and Index (quintile 1 = 1.00) better estimates of the causal impact of spending 1.25 in each quintile, this index represents an assess- ment of how the relationship between spending 1.20 and education outcomes might differ across the

SACMEQ grade 6 socioeconomic distribution. The hypothesis that 1.15 maths, 2000 emerges is that serious problems and inequalities SACMEQ grade 6 reading, 2000 exist in the effectiveness of public spending for 1.10 the poor majority of the population.

TIMSS grade 8 1.05 science, 2003 Funding the schools PIRLS grade 6 reading, 2006 TIMSS grade 8 maths, 2003 1.00 Funds reaching the schools 1 2 3 4 5 Source: World Bank 2003. The effectiveness of reallocating public fund- ing for education depends on the funds actually

74 Accountability in Public Services in South Africa Table 6.2 Disbursed public spending on school grants that actually reach reaching the frontline­—­the schools. Public schools: results from Public Expenditure Tracking Surveys spending flows through many administrative points­—­what is intended to be spent and what Country/year(s) Percentage of funds that reached schools actually reaches the schools might be quite dif- Ghana 1997/1998 51 ferent. Public Expenditure Tracking Surveys Kenya 2004 (secondary school bursary funds) 78 have proven useful in many countries to high- Madagascar 2002 88 Peru 2001 (utilities/consumption goods) 70/97 light how much of allocated and disbursed Papua New Guinea 2001/2002 72/93 public funds actually reach the clients of pub- Tanzania 2002/2003 62 lic services. The differences among allocations, Uganda 1991–1995/2001 <20/80 disbursement, and receipts by schools have been Zambia 2001 (discretion/rule) 24/90 large in some countries: more than 80 percent Source: Ye and Canagarajah (2002) for Ghana; Republic of Kenya (2005) for Kenya; Francken of funds did not reach students in a survey of (2003) for Madagascar; Instituto Apoyo and World Bank (2002) for Peru; World Bank (2004) for Uganda in the early 1990s, while 32 percent Papua New Guinea; Ministry of Finance, Government of Tanzania (2004) for Tanzania; Reinikka “leakage” in Kenya’s secondary school bursary and Svensson (2005) for Uganda; Das and others (2002) for Zambia. program was discovered (table 6.2). The stud- ies find that the differences are larger for schools service provision, or if these disbursements are serving poor communities. delayed or hard to access, the weak relationship Interpreting the reason for the difference is between spending and outcomes should perhaps sometimes less straightforward than measuring it: not be a surprise. Weak implementation of na- leakage could be caused by misallocation or theft tional policy is caused by substantial provincial but also by reallocation of funds at various admin- discretion over education spending. istrative points. Reallocation of budgets could be Data enabling confident statements about legitimate when funds are rechanneled for higher the extent of leakage or delays in the funding priority needs. Nevertheless, reallocation can lead of schools in South Africa are lacking. Some to less than expected education delivery. schools that have demonstrated good financial A poverty-weighted allocation to schools management capacity (Article 21 schools) re- channeled through provincial budgets­—­ ceive a cash budget to procure goods and ser- commonly called the Norms and Standards vices directly. Since allocation of funds to these allocation­—­is an important part of South Afri- schools is fairly straightforward, concern about a ca’s system for funding nonpersonnel school em- large amount of leakage in these disbursements ployees. This grant was designed to redress past is probably misplaced. inequities by allocating public money to schools A majority of schools receive some of their in poorer communities. funds in the form of a “paper budget.” School Since 2007, a no-fee policy has been insti- managers receive a budget envelope and make tuted. The poorest schools receive a government requests to procure goods and services through allocation to offset lost parent fees. The formu- a provincial procurement department­—­creating las for these allocations are fairly straightforward, the potential for leakage. Anecdotal evidence but debate over these allocations focuses particu- suggests that large delays take place among larly on provincial targeting of the funds.13 One the times allocations are announced, requests recent UNICEF-funded study found that some for goods and services are made to provinces, provinces were “equalizing” expenditures. The al- and goods and services are finally received by location formula calls for a sharp decline in sub- schools. Reporting and monitoring when pro- sidies from schools in the richest quintile in favor cured items actually reach schools is lacking. of those in the poorest, but the actual disburse- Delays and uncertainties lead to unpredictability ments are flatter: schools in the poorest quintile and a lack of materials in classrooms, undermin- receive less than determined by the policy, and ing service delivery. schools in the richest quintile receive more.14 The per-student subsidy allocated for poorer students Financial accountability at the frontlines: legal decreases to only three times rather than six times versus in practice that of richer students. If allocated resources are, School principals perform a dual role: they man- in addition, not actually reaching the frontline of age students and teachers within each school as

Accountability in Public Services in South Africa 75 The largest an educational unit and serve as executive offi- Body members, district officers, and provincial share of public cers of the school as an operational unit. Parents education departments, including procurement expenditures on hold principals accountable for financial and departments, Members of Executive Commit- other decisions through the School Governing tees, and others­—­are held accountable for their basic education Body’s review of the annual financial budget. actions. Generating evidence about these issues is allocated to School Governing Bodies are also required to and analyzing them is a priority. educator pay, submit audited financial statements to the rele- vant provincial education department at the end The enabling—­ and­ disabling—­ ­ particularly of the fiscal year. environment for teachers classroom District offices, whose organization var- The largest share of public expenditures on basic teachers ies from province to province, link the provin- education is allocated to educator pay, particu- cial education department to schools. These of- larly classroom teachers. As described at the out- fices are supposed to communicate provincial set of this chapter, the classroom is the linchpin decisions­—­including financial decisions­—­to to providing educational services­—­and teachers schools, and vice versa.15 are important to that process. A recent qualita- Provincial education departments are re- tive study of “schools that work” involving site sponsible for allocating and distributing most visits to 18 fairly well-performing schools serv- types of funds and for monitoring and evaluat- ing fairly poor communities concluded that the ing these expenditures. The provincial govern- “commitment of principals and teachers” was ment holds the Provincial Member of Executive central to producing their good results (Christie, Committee for Education accountable for the Butler, and Potterton 2007, p 58). To improve performance of the provincial education depart- education, the classroom environment should ments. Individuals can direct complaints about encourage good educators to join the profession provincial education departments directly to the and help teachers perform at their best. national department of education, or Student But teaching is difficult. In a study involv- Governing Bodies can take legal action against ing field interviews with parents, teachers, and the departments. School Governing Bodies in two communi- This legal framework for education fund- ties, all respondents agreed that “teaching is ing provides accountability and transparency a difficult job” (Langa and du Toit 2007, p 9). and accommodates schools’ resource needs, but As described in the “schools that work” report, these ideals are not always implemented in prac- principals “consistently mention concerns about tice. As discussed above, reliance on a provincial attracting good people into teaching, and many procurement agency for non-Article 21 schools referred to the fact that teachers are finding it leads to uncertainty about whether the major- increasingly difficult to stay in the profession” ity of schools receive their resources. Substantial (Christie, Butler, and Potterton 2007, p 62). discrepancies are reported between budgets that schools create before the start of the school year Time on task and the resources used by the end of the year Attracting highly motivated staff and estab- (as reflected through their financial statements). lishing the incentives for them to perform well Importantly, schools receive little information is difficult. A recent multicountry study using on what resources they are entitled to and how unannounced visits to measure teacher absentee- much of these resources they have left at any ism16 shows how pervasive it is in many coun- stage of the fiscal year. Many schools report that tries (see table 4.1, chapter 4). large sums go unspent, and these funds cannot Common patterns emerged across the coun- be rolled over to the next year. tries, but the study data also uncovered reasons The discrepancy between the legal and ac- for absenteeism, and schools’ approaches to deal- tual experience suggests substantial scope for ing with this problem are unique in each coun- improving information on how schools make fi- try. Absenteeism is widespread among teach- nancial decisions, whether resources are reach- ers across countries: higher ranking and more ing schools in a timely manner, and whether the powerful educators, such as headmasters, are various players­—­principals, School Governing typically absent more often than those ranking

76 Accountability in Public Services in South Africa Table 6.3 Time devoted to teaching and learning in basic education lower, and better physical quality of school facil- (percent) ities is usually associated with lower absenteeism. Country-specific results include that parent Country/state Potential time Time present Time on task and community involvement in school affairs Tunisia 100 90 78 correlates significantly with lower absentee- Egypt 100 83 74 ism in Papua New Guinea­—­but not in several Morocco 100 86 71 Yemen 100 77 66 other countries. In Zambia, a country ravaged Pernambuco (Brazil) 100 87 63 by HIV/AIDS, most absenteeism was related to Lebanon 100 63 55 health issues or dealing with family emergencies­ Ghana 100 55 39 —­a similar situation might be present in South Africa. Source: Millot and Lane 2002; Abadzi 2009. Comparable teacher absenteeism data in South Africa are not available; the public ex- penditure and education service delivery work teaching time would increase test scores signifi- plan should seek to fill that gap. But school vis- cantly. Taylor (forthcoming) reviews equity and its repeatedly reveal that teacher absenteeism is efficiency in South African education, conclud- a problem. Respondents in the pilot field visits ing that “increasing the time available for teach- referred to “teachers leaving school early or not ing and learning, which at present is being used even attending when they should be presenting very inefficiently in the majority of the country’s classes,” and “some teachers do not even come schools,” would be one of the most effective ways to school on Mondays and Fridays” (Langa and of increasing educational opportunities, espe- du Toit 2007, pp. 6 and 11). In response to the cially for poor children. “Schools that work” school background questionnaire for the 2000 shows how “time on task and careful use of time round of the SACMEQ learning assessment, was an essential­—­and very visible­—­feature of principals reported that health problems, late these schools in producing good results” (Chris- arrivals, absenteeism, and skipping classes were tie, Butler, and Potterton 2007). the four most prevalent educator “behavioral” Chisholm and others (2005) conducted problems. Analysis of the SACMEQ data re- an extensive study of teacher time allocation veals a strong negative association between “late in South Africa, making it the most compa- coming” and test scores, although the author of rable source of information to the international that analysis cautions that the results depend on study above. Their main data from 900 sam- a poorly measured indicator (Gustafsson 2007). pled schools are derived from reported time use A more comprehensive assessment of teacher ab- rather than observed time use, but the findings senteeism, and its causes and impacts in South are nevertheless revealing. Within formally de- Africa, would help understand service delivery termined hours of duty, educators (principals issues. and teachers) reported spending an average of 16 Absenteeism is just the most extreme ex- hours per week teaching and 1.6 hours prepar- ample of how low “time-on-task” leads to loss ing and planning; the remainder of their time is of productivity. Across countries, instructional devoted to administrative (management and su- time strongly influences learning (Coates 2003). pervision, records and recordkeeping) and other An international protocol to collect time-on-task tasks (professional development, guidance and information in classrooms has recently revealed counseling, extracurricular, pastoral care, and that in Ghana, less than half of the time avail- breaks). able is actually devoted to teaching and learning What constitutes “official” teaching time (table 6.3). is uncertain (Chisholm and others 2005): ac- Again no directly comparable data for South cording to the Personnel Administration Mea- Africa is available, but recent studies have iden- sures, teaching time is 29 hours and 45 minutes tified time devoted to teaching and learning as a week for primary and secondary schools, while a major factor in determining outcomes. Gus- the National Education Policy Act (NEPA) tafsson (2007), using SACMEQ 2000 data, estimates between 22 hours and 30 minutes estimates that increasing formally scheduled and 27 hours and 30 minutes of teaching time

Accountability in Public Services in South Africa 77 a week for different grades. Nevertheless, the is not surprising as the teaching force is large 16 estimated hours of actual teaching time re- and diverse. It is unreasonable to label all teach- ported in the Chisholm study falls between 54 ers as underperforming, but just as unreasonable and 71 percent short of “official” estimates of to assume that all teachers are self-motivated time spent teaching. Moreover, though NEPA and hardworking. The public policy challenge calls for more teaching hours at higher grades, is to create systems that encourage highly moti- the study found that teachers reported teaching vated staff to enter the profession while creating fewer hours as grades got higher: they reported incentives for all staff to work hard. 18 hours a week in grades 1 to 3, 16.3 hours in The next section of this chapter will address grades 4 to 6, and 14.8 hours in grades 7 to 9. accountability in schools and how it might in- These estimates do not include absence, which duce higher teacher performance. Before turn- suggests that time lost is even greater.17 ing to that, this section reviews teacher pay­—­a Clearly, time is not used efficiently in South good starting point because it affects both the African schools, suggesting that improved edu- choice of teaching as a profession and how hard cation outcomes can be achieved through in- teachers work once in the position. creasing time spent on teaching and learning.18 What is the best way to assess whether teachers are overpaid or underpaid? Reviewing Teacher pay published scales is potentially useful, but here we The previous section described two diagnostic focus on results from individual and household indicators for low teacher performance: high surveys. One problem is that it is hard to estab- absenteeism and low time on task. This should lish what teachers might earn in other occupa- not be read as an indictment of all educators, tions: what is the relevant counterfactual? The many of whom are clearly dedicated to, and individual and household surveys compare the motivated by, educating the next generation. But incomes of teachers with the incomes of non- field visits confirm the perception that teacher teachers with similar background characteris- effort varies: while parents acknowledged that tics, including years of education. teaching is a difficult job, they had mixed reac- This method estimates the relationship be- tions as to whether “teachers are hardworking tween the logarithm of earnings as a function and committed” (Langa and du Toit 2007). This of productive factors­—­experience and years of education—­ and­ other attributes—­ sex­ and Table 6.4 Estimates of teacher earnings “premium,” based on analysis of household survey data place of residence. It then includes a variable for whether the employee is a teacher. The estimated Country/year Premium/deficit coefficient on this variable can be interpreted Argentina 2000 0.16 as the teacher “premium” or “deficit:” the per- Brazil 2001 –0.19 cent more pay or less pay associated with being Chile 2000 0.43 a teacher. Costa Rica 2000 –0.03 A recent review of the teacher premium Dominican Republic 1998 –0.05 in Latin America showed substantial varia- Mexico 2000 0.23

Nicaragua 2001 –0.24 tion across countries (table 6.4). In Indonesia,

Peru 2000 0.26 a recent study showed that a moderate overall Venezuela 1998 0.11 teacher premium of about 7 percent was concen- Indonesia 2003 0.07 trated in primary school, where the teacher pre- Indonesia 2003 (primary school only) 0.14 mium was about 14 percent. South Africa 1995–2006 (average) 0.32 From 1995 to 2006, teachers in South Af- South Africa 2006 0.28 rica were paid an hourly earnings premium of 32

Note: This table reports coefficient on a “teacher” dummy variable in log hourly earnings percent (see table 6.4), based on evaluation con- regression. In all cases, the models include years of education, and the sample is restricted to sistent with that in the other countries: teachers only employees with secondary education or more. earned an average of 32 percent more an hour Source: Latin American countries based on household surveys between 1998 and 2001 from Hernani-Limarino (2005); Indonesia based on labor force survey data from 2003 from Filmer compared with other employees in the labor (2005); South Africa based on data from van der Berg and Burger (2007). force. This estimate is based on an analysis of October Household Surveys for 1995 to 1999

78 Accountability in Public Services in South Africa and of Labor Force Surveys for 2000 to 2006.19 earnings have been similar to those of other pri- In 2006, the teacher hourly earnings premium vate employees, but monthly earnings for private was 28 percent. teachers have been lower. Analyzing teacher pay is complicated by the This analysis shows that teachers in South fact that teachers typically work fewer hours Africa are not “underpaid:” the majority of than similarly qualified nonteachers: about 10 teachers earn a pay premium in line with other hours a week fewer in the South African labor public sector employees. But average salary is force data analyzed here. While teachers may not the only pay variable encouraging a highly earn more hourly, their take home pay may be able workforce. Prospective employees might similar—­ or­ even less—­ than­ other employees. weigh relative salaries at different points in the Repeating the analysis above but setting the log- earnings distribution: workers with high ability arithm of monthly earnings as the explanatory or characteristics will be more interested in how variable shows that teachers still received a pre- the higher range of teacher earnings compare mium of 16 percent (table 6.5). with other occupations. Moreover, once having The vast majority of teachers are public em- chosen teaching, the distribution of earnings ployees, which explains most of the teacher pay premium: 95 percent of the teachers in the Labor Table 6.5 Estimates of teacher earnings “premium,” based on analysis of individual and household survey data: pooled data from Force Survey worked in the public sector, while 1995 to 2006 almost no teachers worked in the private sector before 2000. We can adjust the model to distin- Hourly Monthly guish public sector teachers from private sector Simple model teachers and include a dummy variable to com- Teachers 0.32 (0.01) 0.16 (0.01) pare the results with nonteacher private sector Augmented model employees. The results suggest that public sec- Public sector teachers 0.49 (0.01) 0.32 (0.01) tor teachers enjoyed an hourly wage premium of Private sector teachers 0.03 (0.03) –0.19 (0.01) 49 percent more than private sector teachers and Other public sector employees 0.36 (0.01) 0.32 (0.01) a monthly wage premium of 32 percent more. Note: Standard errors are in parentheses. This table reports coefficient on a “teacher” dummy Public sector teachers had about the same variable in log hourly and monthly earnings regression. Models include only employees with 10 or more years of schooling. Models include years of experience and its square, years of monthly pay premium as other public sector em- education and its square, a dummy variable for female employees, dummy variables for race, ployees, while the hourly teacher premium was dummy variables for province, and dummy variables for year of survey. larger because teachers work fewer hours. By Source: Analysis of data described in van der Berg and Burger (2007). contrast, private sector teachers earn the same hourly wage as other private employees but al- Figure 6.6 Premiums in monthly and hourly earnings of public sector most 19 percent less monthly. Private sector teachers, private sector teachers, and other public sector employees, relative to private sector employees (1995 to 2006) teachers are paid only for the time they work; public sector teachers are paid a flat rate. Premium/deficit There has been some change in these pat- 0.8 terns over time (figure 6.6). Prior to 2000, 0.6 Public teachers, hourly monthly earnings of public sector teachers were Other public, hourly higher than those of other public sector employ- 0.4 Public teachers, monthly ees. This reversed after 2000, and in 2006 earn- ings for other public sector employees was 4 per- 0.2 cent higher than for public teachers. Other public, monthly Private teachers, hourly Teachers enjoyed a substantially larger pre- 0.0 mium in hourly earnings prior to 2000­—­as high –0.2 as 39 percentage points in 1998­—­which has Private teachers, monthly since narrowed to around 8 percentage points. –0.4 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 The surveys identified few private sector teachers Note: Points are coefficients on a dummy variable for each type of employee in a model relating (explaining the volatile results), but aside from log monthly or hourly earnings on experience, education, gender, and province. an unusual increase in 2004, the survey suggests Source: Analysis of data described in van der Berg and Burger (2007). a fairly stable pattern: private teacher hourly

Accountability in Public Services in South Africa 79 affects teacher performance: incentives to work turning to how school accountability, autonomy, hard are dampened if performance is rewarded and decisionmaking might affect it. Worldwide, less than in other professions. teacher pay is commonly most related to quali- Van der Berg and Burger (2007) used a simi- fications and experience or tenure, with little lar approach to analyze the full distribution of room for adjustment based on performance. monthly earnings by type of employee, compare Such an approach reduces the scope for incen- the actual distribution of earnings among teach- tives to work hard (Hanushek 1996). ers and among other employees, and then simu- An influential discussion on “merit pay” con- late the distribution of earnings among teachers cluded that pay for performance is inappropri- as if they were paid as nonteachers. The results ate for teachers because­—­among other things­ are revealing: teachers enjoy a large wage pre- —­effort and outcomes are hard to observe and mium at the bottom of the earnings distribution hard to attribute directly to the effort of specific compared with what they could earn in the pri- teachers (Murnane and Cohen 1986). But an ap- vate sector or the rest of the public sector (table propriately designed pay scheme can align the 6.6), but this premium shrinks at higher points incentives of educators with those of parents and in the earnings distribution, becoming a penalty policymakers, while rewarding hard work and at the 75th percentile. productivity to attract and retain high perform- Pay incentives initially attract those with ers to teaching. poorer abilities or endowments into teach- Various attempts to introduce performance- ing, where they will earn more than in the pri- related pay schemes have demonstrated that out- vate sector, but more skilled teachers are later comes can be improved. Group-based schemes encouraged to move into the private sector or can overcome the problem of attribution, for to other parts of the public sector, where they instance. The programs with the largest dem- can earn more. In sum, while the average sal- onstrated impacts have been carefully crafted, ary premium might appear to attract high qual- small-scale programs, which have provided fairly ity prospective staff, the structure of earnings large benefits to teachers and/or schools. These both dissuades the highest ability prospective include: employees and dampens incentives once in the • A scheme in Kenya that paid teachers bo- profession. nuses based on student performance and at- tendance (Glewwe, Ilias, and Kremer 2003). Incentives for performance? • A scheme to monitor teacher attendance in Creating a teacher pay structure that provides nongovernmental schools in India (Duflo, appropriate performance incentives is difficult. Hanna, and Ryan 2007). In the next section, we consider attempts to • A group teacher rewards scheme, which directly link teacher performance to pay20 before measured student achievement in Israel (Lavy 2002). Table 6.6 Actual and simulated teacher earnings at different points in the earnings distribution • Group and individual teacher rewards schemes related to student achievement in Compared with Compared with other employees India (Muralidharan and Sundararaman all other employees in the public sector 2007). Teacher earnings Teacher earnings Actual teacher using nonteacher Teaching premium using nonteacher Teaching premium Some larger programs have shown positive Percentile earnings earnings structure (percent) earnings structure (percent) impacts. In Mexico’s Carrera Magisterial pro- 5th 1,087 302 259 677 61

10th 1,523 462 230 1,132 35 gram, teachers volunteered to be evaluated in

25th 2,225 1,042 114 1,906 17 exchange for the chance at a bonus of up to 25

50th 3,191 2,228 43 2,965 8 percent of their base salary: secondary school 75th 4,184 4,202 0 4,376 –4 teacher effort improved along with student test 90th 5,159 7,428 –31 6,375 –19 scores (McEwan and Santibañez 2005; Santiba- 95th 5,770 10,339 –44 8,417 –31 ñez and others 2007). In 1990, Chile’s National

Source: van der Berg and Burger (2007) based on October Household Survey and Labor Force System of School Performance Assessment Survey data from 1995 to 2006. started offering bonuses to schools that showed high student achievement: student achievement

80 Accountability in Public Services in South Africa Box improved cumulatively (Mizala and Romaguera 6.1 School-based performance awards in Chile 2003; box 6.1). Chile has offered an award for “top performing” schools in each region every two Other programs have shown less impact. In years since 1996. Ninety percent of the award goes directly to teachers, in propor- Mexico’s Primary Education Project, teacher tion to their hours of employment, and 10 percent is allocated to the schools. attendance was monitored and bonuses were Schools are divided into comparison groups within each region based on lo- paid to teachers who attended regularly, but at- cation (rural, urban), education level (primary only, secondary with primary), and socioeconomic status of parents (according to information collected as part of the tendance was not substantially affected (Ezpeleta examination and an official “index of vulnerability”). In 2000–2001, this classification and Weiss 1996). The Carrera Magisterial produced 104 comparison groups, avoiding directly comparing poor rural schools program­—­so successful in secondary schools­—­ with rich urban schools to decrease the disproportionate distribution of rewards did not have an impact in primary schools. to richer schools. Teacher incentive schemes can create un- An index of school performance is calculated based on standardized tests of intended consequences. Some studies in devel- Spanish and mathematics in grades 4, 8, and 10. The index uses a formula weighted oped countries highlight how linking pay to test for average test level (37 percent) and improvement in test scores (28 percent), but scores led to teacher cheating (Jacob and Lev- it also includes other criteria such as “equality of opportunity” (22 percent) and “initiative” (6 percent), based on regular development of group pedagogical activi- itt 2003) or exclusion of low-achieving students ties, among other things. (Figlio and Getzler 2002). Group incentives can have adverse consequences by exacerbating Source: Mizala and Romaguera (2003) in World Bank 2003. teacher “free-riding:” when it is hard to observe the effort of all team members, people have an incentive to shirk their duties while benefiting School decisionmaking with parent involve- from the rest of the team’s work. The challenge ment is clearly undeveloped as a school improve- for any scheme linking pay to performance is to ment strategy in South Africa­—­but it does generate clear, transparent incentives that are ac- exist. As one principal in the recent “schools ceptable to all parties and that avoid unintended that work” study states regarding the hiring of consequences. educators: A recent UNICEF-funded study suggests “The critical thing is selection. I was for- that such a scheme would be acceptable in South tunate that when I came in, the parents had Africa: around 80 percent of parents, teachers, confidence in me. When we conduct inter- and principals reported that they would support views, it is very important to know who you are a system that paid schools a bonus for improving recommending­—­although the appointment is results. As mentioned, creating a scheme that going to be made by the Department of Educa- provides incentives without unintended negative tion. I was fortunate in that I had an opportunity consequences is difficult, but a system that does to go and check on all of the HoDs. It’s people not reward good teaching will both fail to at- who are passionate about education. It makes a tract high quality staff to the profession and fail big difference when parents in the SGB (School to motivate teacher performance. Governing Body) have confidence in you. That’s where it starts” (Christie, Butler, and Potterton Accountability at the school level 2007, p. 60). This chapter has focused on the long route to This quote illustrates how a supportive accountability: national-level learning assess- School Governing Body enabled the principal ment results, changes in public expenditures to make a strong hiring recommendation­—­how post-apartheid, and teacher compensation. parent participation and school autonomy can These issues affect citizen voice and the compact play out. between policymakers and frontline providers. But when outcomes are hard to centrally moni- Parent participation tor, and when populations are heterogeneous, Anecdotal evidence suggests that client power is growing evidence suggests that the short route weak in many South African schools­—­despite to accountability­—­between the direct benefi- the legal existence of School Governing Bod- ciaries of services and the frontline providers ies. While parent participation is consistently of those services­—­can be a powerful tool to reported to be high in more affluent areas, it is improve outcomes. lower in schools serving most of the population.

Accountability in Public Services in South Africa 81 The PIRLS 2006 reading assessment of grade Parents should be able to hold schools ac- 5 students in South Africa included a school countable, particularly for things they can ob- questionnaire completed by the principals: few serve: Are educators absent or tardy? Are text- reported that more than half the parents par- books arriving too late to be used effectively? ticipated in school activities­—­volunteering in Involvement—­ and­ impact—­ increases­ with in- the classroom, attending teacher-parent confer- formation, which can take many forms. In field ences, or attending social and sporting events­—­ interviews, parents said that teachers should although participation rates were substantially “mark the tests of the learners and inform higher for children in the richest quintiles (fig- both the children and parents on the outcomes ure 6.7).21 Whether this difference in partici- thereof” (Langa and du Toit 2007, p. 17). The pation matters for decisionmaking depends on recent UNICEF-funded survey found little re- whether participation in poorer schools is suffi- lationship between parents’ assessment of school cient for enabling substantive input. The qual- quality and the results of the systemic evalu- ity of participation is likely to be the determin- ation, which suggests that parents are not well ing factor, which in turn depends on a variety of informed about their children’s performance in other factors. school. Low parent participation, especially in More generally, publication and dissemina- schools serving poor communities, might be tion of inputs and outcomes can inform parents symptoms of several things, not the least of on how the school is funded and how outcomes which is that poor parents have limited time. compare with other schools­—­potentially spur- But focus group interviews suggest parent en- ring parents to action. Similar approaches have gagement can be increased: parents tend to show yielded positive impacts in other contexts. Pub- up when school fees are being discussed, for in- lishing how much each school was being allo- stance (Christie, Butler, and Potterton 2007). cated by the central budget in Uganda reduced At the same time, parents seem to be gen- leakage of public funds and improved outcomes: erally satisfied with the functioning of School parents could monitor funding for each school Governing Bodies. The recent UNICEF-funded and how schools were using the funds (Reinikka survey of 525 schools found that most parents and Svensson 2005; Björkman 2007). The Bra- not on School Governing Bodies thought that zilian state of Parana experimented with school their Body was doing a good job. report cards between 1999 and 2002 (Winkler 2005). Test scores and information from school Figure 6.7 Students attending schools where the principal reports that more than 50 percent of parents participate in three types censuses and parent surveys were compared with of school-based activities, by socioeconomic quintile of the municipal and state averages to create a simple student three-page summary of both school perfor-

Percent Poorest quintile Quintile 2 Quintile 3 Quintile 4 Richest quintile mance and perceptions of performance, which 50 was then disseminated to parents, teachers, and school administrators. But only anecdotal evi-

40 dence is available to suggest that the program had positive effects.

30 Other experimental studies have been de- signed to identify causality: a study of an infor-

20 mation campaign in three states of India (Pan- dey, Goyal, and Sundararaman 2009) and a

10 test-based student and school report cards study in Pakistan (Andrabi, Das, and Khwaja 2009) 0 both show that information can impact behavior Volunteer regularly to Attend teacher-parent Attend cultural, sporting, or help in the classroom conferences social events at the school and measured learning outcomes. or another part of the school Information is not necessarily a panacea. A re- Note: Quintiles derived from an index of household ownership of various consumer durables. cent carefully designed evaluation in India showed Source: Authors’ analysis of PIRLS 2006 data. that dissemination of information on rights and responsibilities failed to have any measurable

82 Accountability in Public Services in South Africa Box 6.2 Meta-analysis of school-based management initiatives in the impact on outcomes or on community involve- United States ment in schools (Banerjee and others 2008). In a meta-analysis of the effectiveness of school-based management models in School autonomy the United States­—­or Comprehensive School Reform (CSR)­—­Borman and others (2003) reviewed 232 studies with 1,111 independent observations evaluating 29 Parents might not participate in school man- CSR programs. From these observations, they computed the size of the effect agement because they feel they can do little to that these models had on student achievement. The authors regressed weighted effect change. Moreover, parents may judge that effect size on the moderator variables to obtain the residuals from the regression, schools have limited scope for making changes: added the mean weighted effect size on the moderator variables to obtain the it is hard to imagine that parent participation­ residuals from the regression, and added the mean weighted effect size to each —­and pressure­—­could have much impact observation, thus calculating effect sizes that were statistically adjusted for all of on schools unless schools receive a degree of the methodological variables. They found that the number of years of implemen- decisionmaking autonomy. tation of the CSR was a statistically significant predictor of student achievement effect size. Some Article 21 schools that demonstrate financial capacity have authority over expendi- Adjusted effect size tures. The number of these schools is growing, 0.5 but they tend to be in wealthy communities with the most active School Governing Bodies. In 0.4 these schools, certain categories of funding can be “ring-fenced” (as documented in the recent 0.3 UNICEF-funded study), resulting in limited control over expenditures. 0.2 Non-Article 21 schools are not authorized to procure goods and services directly and need 0.1 each request authorized and procured through 0.0 the province. These schools have limited scope 1 year 2 years 3 years 4 years 5 years 6 years 7 years 8–14 years for determining which teachers are hired and Years of implementation fired, though schools in wealthier communities sometimes raise funds to pay for complementary Source: Borman and others (2003) in World Bank (2008). teachers over which they have authority. School decentralization in some very poor settings, such as rural communities in post-civil evaluations are that “SBM policies changed the war El Salvador or secondary schools in Nicara- dynamics of the school, either because parents gua, has shown compelling results. Growing evi- got more involved or because teachers’ actions dence from various countries­—­primarily in Latin changed.” Several studies found that “SBM had America­—­shows that school-based management a positive impact on retention rates, failure rates, (SBM) can improve outcomes, even in very poor and, to a lesser degree, dropout rates” (pp. 12 and communities.22 School-based management in- 13). The studies found mixed impacts on learning cludes reforms that devolve decisionmaking to outcomes, in part due to the timing of impacts. schools: some programs transfer power to prin- Studies from the United States have shown cipals or teachers; others strengthen parental that the devolution of power to schools is a cu- and community decisionmaking. The scope of mulative process that takes place over several decisionmaking devolved differs across programs years (box 6.2). but may include hiring and firing of teachers and Evaluations of school-based manage- other staff, curriculum development, procure- ment initiatives suggest substantial scope to ment of textbooks and other instructional mate- give more decisionmaking authority to poorer rials, infrastructure improvement, and monitor- schools. One of the more intriguing­—­and per- ing and evaluation of teacher performance and haps controversial­—­findings is that school im- student learning outcomes. provement is largely attributable to flexibility in The types of programs are discussed in schools’ ability to hire and fire educators. World Bank (2008) and its references, but the In an in-depth analysis of how school-based main conclusions from the review of rigorous management reforms operate in El Salvador,

Accountability in Public Services in South Africa 83 A broader agenda Honduras, and Nicaragua, Umansky and Vegas Establishing a pay structure that encourages encompasses (2007) conclude that “[teacher] behaviors appear skilled people to teach, and then provides incen- curriculum to respond to the incentives created in locally tives for them to perform to their highest poten- managed schools. As a result of school-based tial, is important for the compact. The current pay reform, teacher management teachers are working more hours, structure attracts those with lesser abilities to the training, and assigning more homework, and meeting more teaching profession where they will earn more infrastructure often with parents” (p. 211). But they also cau- than in the private sector, while pushing more tion that reforms did not change teaching meth- skilled candidates into the private sector or other development ods and that increased job insecurity led better parts of the public sector. The new South Afri- qualified teachers to shy away from locally man- can salary dispensation for teachers, announced aged schools. after long negotiations with teacher unions, goes Nonetheless, school decentralization that is a long way toward creating incentives for teachers carefully crafted­—­especially in decisions about to enter the field. Linking teacher pay to perfor- educators­—­has the potential to complement mance was another positive development, though pay reforms in creating incentives for higher it is still not certain how it will be put into prac- performance. tice. Experiments to find what policies might work best in South Africa could help. Policy recommendations and research The short route to accountability in most agenda South African schools is limited, especially in This chapter has focused on a narrow set of edu- schools serving the poorest communities. Par- cation issues. A broader agenda encompasses ents are not typically active in these schools, curriculum reform, teacher training, and infra- despite the legal existence of School Govern- structure development. Moreover, evidence on ing Bodies. Parents must be made aware of their which to base policy recommendations remains rights and be given information that empowers thin. Nevertheless, this review does suggest them to hold schools accountable. some areas for policy reform. Options to increase school decision- Disseminating information about outcomes making, which would afford clients more power, is a priority for tracking progress, informing should be explored. Carefully crafted school parents on schools performance, and strength- decentralization­—­increased autonomy in finan- ening citizen voice. South Africa’s participation cial decisionmaking, in procurement of goods in international assessments goes a long way to- and services, or in hiring and firing of educators­ ward achieving this objective, as do national as- —­can empower schools and make them ac- sessments (“systemic evaluations”). Timely dis- countable to their clients. semination of this information, in a format easy Information about school autonomy, parent to understand and usefully disaggregated, will participation, and school governance and man- strengthen citizens’ ability to exercise their voice. agement is growing. But important questions re- The operation of the paper budget in non-Ar- main, and work on the following will be helpful ticle 21 schools could be reviewed to strengthen in shaping the future agenda: the compact between government and frontline • How are financial decisions made at schools, providers. Delays in procurement and uncertain- and how does this vary? Are there a substan- ties about remaining budgets create an environ- tial number of schools capable of managing ment in which schools cannot plan effectively. their finances but currently constrained by More generally, a greater understanding of not being able to achieve Article 21 status? school financial operation and decisionmaking­ • Are resources reaching schools on time? The —­whether resources are reaching schools on hypothesis is that this varies across provinces time and whether the various players (princi- and that increased decisionmaking author- pals, School Governing Body members, district ity, combined with adequate accountability, officers, provincial procurement departments, could streamline the process and improve Members of Executive Committees) are held ac- the flow of funds and resources to schools. countable for their decisions­—­would shed light • Over what areas do schools, school admin- on how to start reforming.23 istrations, and School Governing Bodies

84 Accountability in Public Services in South Africa exercise autonomy (selecting students, set- Distribution of School Funds in South Af- ting school fees)? And how does this differ rica.” June 2007. across schools? The hypothesis is that schools • Langa, R., and A. du Toit. 2007. “Quality serving poor communities are less autono- Education in Rural Schools in South Africa: mous and that less autonomy hampers their Explorative Pilot Study.” June 2007. ability to respond to their localized needs. • Van der Berg, S., and M. Louw. 2007. • How problematic is teacher absenteeism in “Mapping Trends in the Incidence of School South Africa, and what causes it? Under- Expenditure in South Africa beyond 2000.” standing whether absenteeism is caused pri- June 2007. marily by complementary inputs (for example, • Van der Berg, S., and R. Burger. 2007. infrastructure), governance (accountability to “Teacher Pay in South Africa.” June 2007. principals and communities), salary struc- 1. South African literature usually uses the tures, or individual factors (such as health terms learners and educators rather than stu- status) will help formulate policy recommen- dents and teachers­—­this chapter uses the dations to address teacher absenteeism. terms interchangeably. • What specific roles do School Governing 2. Consistent with most analytical work on Bodies play, and are they effective? The -hy South Africa, we use the apartheid-era pop- pothesis is that School Governing Bodies are ulation groups. “hands off” entities for most schools, espe- 3. Case and Deaton (1999), for example, em- cially those serving poor communities. Ex- pirically show how the apartheid education ploring what makes some School Governing systematically resulted in lower quality in- Bodies active and effective can shed light on puts for the nonwhite population, which in ways to increase their effectiveness in other turn led to lower education achievement and areas. test scores. • What mechanisms hold schools and com- 4. These statistics are from the Department of munities accountable? Of particular interest Education (2008). The South African pri- are how school management interacts with mary and secondary education system is made parents and communities and how much, up of two “bands:” The General Education and by what means, do parents hold school and Training band includes a Foundation management accountable. Phase (grades “Reception” and 1 to 3), an In- • To the extent that these mechanisms­—­both termediate Phase (grades 4 to 6), and a Senior formal and informal­—­are lacking, what is the Phase (grades 7 to 9); the Further Education appropriate policy response (enhancing capac- and Training band includes grades 10 to 12. ity, providing more local information, con- 5. South African results for TIMSS are de- ducting top-down “inspection,” for example)? scribed in Reddy (2006). PIRLS refers to • Is school-based management, particularly this as the grade 4 reading assessment, but control over the hiring and firing of teachers, this corresponds to grade 5 in the South Af- consistent with the broader policy framework rican education system. and even feasible with high unionization? 6. Scores are normalized so that the average The hypothesis is that it would be difficult to score across all test takers in the 1995 par- reform pay structures given entrenched inter- ticipant countries was 500, with a standard ests. Nevertheless, given the urgency of en- deviation of 100. suring high quality and motivated teachers, 7. Part of the difference can be attributed to the an acceptable arrangement must be found. fact that the group of students is somewhat more selected in Kenya, because a slightly Notes higher percentage of children reach grade 6 In addition to the references provided in the in South Africa. This differential, however, reference list, this chapter draws from various is unlikely to explain the gap in national per background studies. These include: capita income. • Fourie, J., and M. van Niekerk. 2007. 8. South Africa chose to not participate in the “An Investigation into the Allocation and 2007 round of the TIMSS.

Accountability in Public Services in South Africa 85 9. Data is from Republic of South Africa shocks to absenteeism and to allow an analy- (1999) and Republic of South Africa (2007), sis of whether it is limited to certain teachers which report a per learner expenditure of or spread across teachers. ZAR3,232 and ZAR6,895 in nominal 17. The timekeeping instrument used in the terms, respectively. An inflation factor of study did not include an explicit category for 1.41 was applied to the earlier number to time absent. convert it to real 2006/2007 terms. 18. For international comparison, Gustafson and 10. Issues of comparability over time exist, as Patel (2008) use data from the 2003 TIMSS discussed more fully in van der Berg and to derive the instructional hours delivered by Louw (2007). grade 8 math teachers in a number of coun- 11. This analysis derives quintiles on the basis tries. They find a range of 24 hours in Chile to of the student population of those schools, 11 hours in Ghana, with most countries falling which are not strictly comparable with the between 15 to 20 hours. In that study, South quintiles used for the expenditure numbers African grade 8 math teachers are calculated as in figure 6.2. providing 17.6 hours per week of instruction. 12. Note that this analysis is sensitive to the way 19. Details on the dataset used for these calcula- scores are normalized across assessment sys- tions are in van der Berg and Burger (2007). tems. The consistency of the shape of the 20. Santibañez (2008) usefully summarizes the relationship, however, does not depend on literature on teacher incentive programs, and normalization. the following draws from that review. 13. See Fiske and Ladd (2004) and references 21. The exception is for “volunteering” in the therein. poorest quintile (perhaps because there is a 14. This study, currently in draft form, is entitled greater need)­—­although the overall level is “School Funding and Management in South extremely low in all groups. Africa: Findings from the School Survey.” 22. This discussion draws heavily from a recent 15. These de jure procedures are laid out in the review of evaluations of school-based man- South African School Act Articles 34, 38, agement interventions (World Bank 2008). 42, and 43. 23. The recent UNICEF-funded study will help 16. The international protocol involves two fill knowledge gaps in school funding and unannounced visits separated by sev- management and be a building block for eral months to account for potential local guiding policy reforms in this area.

86 Accountability in Public Services in South Africa 7

Water and Sanitation

In 1994, one-third of South Africa’s 40 mil- and rich areas. Service-related protests have in- lion people could not access safe drinking water creased sharply since 2009: angry citizens want within a reasonable carrying distance, and half service as much as they want delivery­—­they did not have a toilet (DWAF 1994). The gov- want municipal councilors to treat them with ernment has set ambitious targets for improv- respect. ing access to water and sanitation. South Afri- Water services are managed by local govern- ca’s government pledged repeatedly to eradicate ments, meaning that remedying these problems water backlogs by 2008 and sanitation backlogs will require intensive customized interventions by 2010. Significant pressure was put on munici- in virtually every municipality in the country. palities to achieve these bold targets, with sub- This chapter argues that public accountabil- stantial funding from the treasury. ity mechanisms for water services need urgent In May 2008, the Minister of Water Affairs strengthening if universal services access is to be told Parliament that South Africa had met the achieved and if the benefits of immense public Millennium Development Goals for both water investment are to be realized. and sanitation, having met the water target some Deteriorating municipal water services has time earlier (Hendricks 2008). Substantial prog- serious implications for national water security. ress has been made toward the infrastructure Even with additional infrastructure, South Af- required for universal service access, and Pres- rica is likely to face a national water deficit of ident Zuma’s administration has committed to between 17 percent and 30 percent as demand providing services in all areas by the end of his steadily outpaces supply over the next 20 years term in 2014. (2030 Water Resources Group 2010). Difficult But service access gains are under severe tradeoff choices will need to be made among threat­—­a crisis in water services is taking place. agriculture, critical industrial activities such as Hundreds of settlements, towns, and cities ex- mining and power generation, and large and perience localized water outages, treatment fail- growing urban centers. Substantial investment ures, disease outbreaks, leaks, bursts, and sewer in water infrastructure will raise the cost of spills arising from poorly maintained infrastruc- water. ture. The institutional requirements for sustain- These realities underline the critical impor- able water services have been severely underes- tance of sound, accountable management of the timated, and weak municipal service delivery country’s water sector. How decisively will the management systems are eroding the value of government be willing to act to safeguard sound vast investment in infrastructure in both poor management of water services in the interests of

Accountability in Public Services in South Africa 87 Strong national water security and the constitutional But South Africa underestimated the re- commitment to right of access to sufficient water? What mean- quirements needed to sustain these water ser- infrastructure ingful avenues will open for more effective pub- vices. Expanding access is straightforward com- lic engagement and public accountability? pared with ongoing service challenges such as development Municipal water supply is the fastest grow- operation and maintenance, integrated devel- and spending ing sector of national water demand, reflecting opment planning, revenue administration, and on servicing population growth and rising living standards relationship building between citizens and mu- (van Rooyen 2008). The effectiveness of munici- nicipal structures. Concurrent local government subsidies achieved pal water services management has critical im- restructuring has meant that infrastructure ex- impressive results plications for national water security, now and pansion has generally outpaced the institutional increasingly in the future. infrastructure needed to run the new services. Managing water services has various dimen- Most municipalities have opted for a techno- sions: managing water demand through effective cratic, top-down approach to delivery, with de- metering and tariff systems, promoting efficient cisions taken and implemented with little public use and minimizing losses, and safeguarding the involvement. But many municipalities lack the quality of return flows of treated wastewater so skills to succeed with a technocratic approach. that surface and groundwater systems remain fit In too many instances this has led to services not for use. Sound asset management is essential to being delivered or poor quality infrastructure maintain extensive water reticulation networks without the resources to keep it working. to limit leaks and physical losses, and sound Government emphasis on meeting service wastewater management is necessary to mini- delivery targets for distributive justice has been mize pollution of water sources. achieved at a high cost. As Seddon (2008) points Institutional issues, rather than technical is- out, target-driven approaches to service delivery sues, are to blame for these problems, reflecting inherently divert resources from areas not tar- a collapse of governance and accountability sys- geted, compromising overall performance. In tems and a lack of sanctions for bad decisions. South Africa, widespread evidence suggests that Many municipalities are ineffective or dysfunc- emphasis on new service connections has led to tional and characterized by political in-fighting, neglected maintenance of existing infrastructure inappropriate appointments, corruption, and and lack of capacity in bulk infrastructure, caus- neglect­—­points made repeatedly by the Presi- ing service failures and bottlenecks. dent and Cabinet ministers in statements and Perhaps more important, pressure to meet government reports in 2009 and 2010. delivery deadlines has limited local engagement on service options, preferences, and approaches­ The changing approaches to service —­and marginalized the intended beneficia- provision ries. Delivery has been prioritized over service, Strong commitment to infrastructure devel- largely removing people from any role in holding opment and spending on servicing subsidies local government accountable for service delivery achieved impressive results in South Africa. (Friedman 2006b). Further, the reciprocal roles Government figures indicate that by 2009 water and responsibilities of citizens in supporting sus- services infrastructure had been extended to 96 tainable service provision are not always agreed percent of the population, up from 59 percent in on or understood. Many citizens complain that 1994 (DWAF 2009a). To put the achievement services are not affordable, causing them to ac- in perspective, far more people were reached cumulate debt. by new water services infrastructure since 1994 Public accountability relationships among in South Africa than the total population of different spheres of government are emphasized Angola, Chile, or the Netherlands; and more over relationships between municipalities and people are benefiting from improved sanitation service providers: few agreements stipulating facilities in South Africa than the total popu- performance parameters are in place between lation of Zambia. Even more remarkably, the municipal managers and their own technical majority of new toilets in South Africa were flush services departments or external service provid- toilets provided in new housing developments. ers, and performance monitoring is inadequate.

88 Accountability in Public Services in South Africa Even less attention is given to the relation- people have full water-borne sanitation, with a ships between citizens and service providers­—­ further 2 percent using a flush toilet that drains client power in the short route to accountability­ to a septic tank on-site. Eight percent have a —­and between citizens and their political ventilated improved pit (VIP) toilet, the gov- representatives­—­citizen voice in the long route ernment-funded default for basic sanitation out- to accountability. The implications for sustain- side of urban areas. Two percent are serviced by able provision of effective water services in South chemical toilets­—­a nonflushing, on-site toilet Africa are profound. treated with chemicals to reduce odor­—­which require frequent emptying. One percent has a Infrastructure provision for water bucket toilet­—­a crude form of on-site toilet rely- services ing on weekly municipal emptying service­—­and Viable technologies for drinking water and sani- 27 percent make their own arrangements, usu- tation improvement are shaped by quality and ally a rudimentary pit toilet or open defecation proximity of the water source, settlement den- (Stats SA 2007; DWAF 2008a). sities, scale economies, water consumption, institutional capacity, funding, and aspirations, Improving rural access to water and sanitation among other factors. Today, as in 1994, the most severe service defi- In low-density settlements, potable water ciencies are in rural areas.1 The new govern- from communal taps stands within a reason- ment’s approach to rural servicing focused on able carrying distance and robust on-site dry providing a basic level of service­—­“some for all, toilets meet basic requirements for public health. not all for some”­—­emphasizing community-led Where settlement densities align with human demand responsiveness. The government recog- and financial resources required for sustain- nized it did not yet have the capacity to provide able management, networked water services and and maintain services directly; the new Consti- connections to each plot are feasible. But the tution assigned responsibility for service pro- increased per capita water consumption from vision to local government, but there were no networked yard or house connections introduces municipalities in most former “homeland” rural the need for off-site drainage and wastewater areas before 2000. disposal. Cognizant of its limited capacity, the gov- The same combination of scale econo- ernment instead promoted community-managed mies and resources supports reticulated drain- projects to improve access to water and sanita- age, generally linked with flush toilets. Sew- tion. Government-supported agencies helped ered flush toilets offer status and convenience residents form community structures, provided but few health benefits when compared with training in project and financial management, well-built and well-maintained dry toilets. As a and assisted with program implementation­—­ consequence of South Africa’s history of racial subsequently community managed and operated. discrimination, however, anything other than Water projects focused on piping water locally a flush toilet is regarded widely as second-best, discriminatory, and interim until a “proper toi- Figure 7.1 Access to water and sanitation in South Africa, 2008 let” can be installed. Access to water Access to sanitation Figure 7.1 summarizes the distribution of Bucket toilet 1% the main water and sanitation technologies Unserved 4% Chemical toilet 2% Unclear 4% Other 5% Ventilated improved used in South Africa. Recent census statistics pit toilet 8% (Stats SA 2007) show that nearly half of South

Africans have in-house water supply, while a Standpipe 20% further quarter have a water connection reticu- House tap Own arrangement 48% below Reconstruction Flush toilet lated to their yard. Twenty percent source their and Development 58% Yard tap Programme 27% water from a communal standpipe within 200 23% meters­—­the supply standard for a basic service­

—­and an estimated 4 percent do not yet have ac- Source: Derived from Stats SA (2007); DWAF 2008a. cess to an improved supply. Fifty-six percent of

Accountability in Public Services in South Africa 89 The decay of local from protected springs, drilling boreholes, and capacity has led to tradeoffs. In rural areas, the accountability linking settlements to new rural water schemes. introduction of Free Basic Water meant that vil- mechanisms­ Sanitation improvement emphasized creating lage water committees no longer had funds to demand, raising awareness of how to stop the pay for unexpected repairs and maintenance. —notably­ the transmission of diseases associated with poor en- They also lost authority over their scheme op- politicization of vironmental health and hygiene, and construct- erator, who was now paid by, and answerable to, ward committees ing enhanced pit toilets designed to minimize remote municipal government. Local oversight odors and insects. mechanisms eroded as the function of village along factional The capital costs were cofinanced by house- water committees was progressively taken over party lines—­ ­ holds, with work opportunities on the projects by ward committees and municipal service pro- has contributed helping to raise household funds. Operation viders, with a commensurate decline in public and maintenance costs were covered by small accountability mechanisms. to failures in monthly contributions from each household, Municipalities themselves had only been es- rural schemes and water supply was managed by a village water tablished in the former homeland areas in 2000. committee, which paid and employed the local Newly vested with responsibility for service pro- operator. By the late 1990s, this approach had vision, they were keen to take charge, whether brought improvements to hundreds of small or not they had the reach, systems, or expertise settlements across the Eastern Cape, KwaZulu- to manage services over a large number of small, Natal, and Limpopo. widely dispersed settlements. Centralized man- But the government was concerned that agement of service provision in numerous dis- the rate of progress on community-driven proj- persed settlements is logistically difficult; few ects was too slow. Supply-side strategies were municipalities have been able to avoid intermit- steadily overtaking community-managed ap- tent service breakdowns and failures in rural proaches, with projects increasingly driven by schemes: pumps stand idle for weeks or months tight engineering management and report- because of a lack of fuel or parts, and water is ing schedules. Community involvement in not treated because of a lack of chlorine (DWAF planning, decisionmaking, and management 2007; DWAF/CSIR 2007; Momba, Obi, and withered, and the cost and complexity of sup- Thompson 2008). The decay of local account- ply-driven schemes grew, making local self- ability mechanisms­—­notably the politicization sufficiency difficult. The high cost of accessing of ward committees along factional party lines­ water supplied by these new schemes was a fac- —­has contributed to these failures. tor behind the new Free Basic Water policy in- Rural settlements­—­where service deficien- troduced in 2001 (Eales 2010a). cies are largest and challenges related to remote Free Basic Water provides important support servicing are most acute­—­are badly in need of to poor and vulnerable households, in line with coproduction for water services. Service partner- the government’s pro-poor objectives. Free Basic ships that combine local skills and accountability Water covers both the capital costs of accessing with the resources of the larger municipality and services and the provision of a defined volume service provider are needed. of water (generally 6 kiloliters, but higher in A potent illustration of the breakdown in some urban areas) free each month to all quali- public accountability relationships came in June fying households, with funds transferred from 2010: City Press reported that cronyism in the national to local government to cover the cost. allocation of municipal tenders in Limpopo’s Department of Water Affairs statistics suggest Mopani District municipality had caused severe that 85.6 percent of households enjoyed access water shortages in over 100 villages to continue to Free Basic Water in 2009 (DWAF 2010); de- despite municipal expenditure of ZAR25 mil- mand-side figures collected by Statistics South lion on borehole maintenance and development Africa suggest far lower coverage, particularly in the past year. Villagers complained that the among poor households (Stats SA 2010a). contractors did shoddy work, installed pumps Free Basic Water clearly provides relief to with inadequate capacity, and drilled boreholes poor households, but “municipalization” of ser- in the wrong places. Allegedly, more than half vice provision in new municipalities with limited the tenders were awarded to local party cadres

90 Accountability in Public Services in South Africa shortly before the African National Congress’s and hazardous. Elsewhere in the world, VIP toi- The tradeoffs (ANC) regional elective conference in Septem- lets work well because they are built and man- involved with a ber 2009, and taxpayer money was used to buy aged as a household asset; pits are large enough supply-side, target-­ votes and secure the election of the mayor as the to meet the family’s needs, and because the toilet party’s regional chairperson. With many taps is their responsibility, users have an incentive to driven approach still dry nearly a year after the tenders were is- limit the rate at which the pit fills, say, by not are particularly sued, residents have reverted to drinking river throwing solid waste into it. evident in basic water or buying water sourced from private bore- The eThekwini Metro municipality is a rare holes at ZAR1 per 25 liter container. The news- example of a more pragmatic approach. It pro- sanitation paper quoted one resident as saying: “The prob- vides a free service to empty pits in households lem is that even if you complain, nothing will with VIP toilets but aims to phase out this li- happen” (Rampedi 2010). ability by providing a dry ecological toilet in areas without sewers, which users can safely Basic sanitation improvement: from assistance manage and empty themselves. It puts as much to service provision effort into public education on user roles and re- The tradeoffs involved with a supply-side, target-­ sponsibilities in integrated sanitation as it does driven approach are particularly evident in basic to building toilets. Households are responsible sanitation. Infrastructure backlogs are far more for maintaining their own toilets, but small or severe than for water, but the priority citizens’ medium enterprises assist households in empty- give to improvements is lower. To meet national ing dry toilets, and the municipality pays local backlog eradication targets, municipal delivery facilitators to guide households in maintenance programs provide on-site, dry VIP toilets where (Gounden 2008). reticulated sanitation is not feasible. Munici- Notwithstanding eThekwini’s commitment palities have been allocating significant grants to expanding the reach of waterborne systems, to contractors to deliver VIP toilets. The build- sewered sanitation is usually not feasible or af- ing contractor is accountable primarily to the fordable in sparsely settled areas. Septic tanks municipality, rather than the household, and are an alternative, but they require a continuous households are given little if any choice about supply of water to the house­—­again not feasible toilet type, location, or appearance. More atten- or affordable in all areas. Yet these constraints tion is given to oversight of funding flows than are not widely understood, and families are of- to on-site performance, and household “own- fended when they feel they have received inferior ership” is limited. Basic sanitation improve- sanitation. ment now focuses on toilet building as part of a Sanitation is very important where HIV in- broader package of government services, rather fection rates are high, and the government faces than as a household responsibility supported by some tough choices. Its supply-driven approach government infrastructure. to providing VIP toilets is not delivering sus- Sanitation improvement has shifted its focus tainable sanitation, and a steady backlog of hun- from health to dignity, and the physical appear- dreds of thousands of full, unusable VIP toilets ance of the structure sitting on top of toilets is is growing. Simple alternatives­—­such as VIP often given more attention than the function- toilets with movable top structures­—­are seldom ing of the technology (Buckley and others 2008). explored. This has led to the construction of brick or con- The focus of sanitation improvement needs crete block structures that cannot be moved when to be brought back to people­—­not toilets. One the pit is full. A growing portion of about 1 mil- option is to halt Municipal Infrastructure Grant lion VIP toilets built over the last decade are now funding for immovable VIP toilets, creating in- full and unusable. In addition, households do not centives for municipalities to explore more sus- regard emptying of these municipal pit toilets as tainable approaches. Another is to put funding their responsibility, and few municipalities have power directly in the hands of citizens­—­through the funds or equipment to de-sludge mass pits. a voucher system enabling them to choose their VIP toilets are inappropriate for municipal toilet type, for example. Doing so would make sanitation because pit de-sludging is unpleasant the supplier accountable to the user, not merely

Accountability in Public Services in South Africa 91 the municipality. It may promote better under- In South Africa, the choice of urban sanita- standing of the respective roles, responsibilities, tion technology is heavily determined by politics: and accountabilities of municipalities and citi- flush toilets are associated with white privilege zens and serve the interests of improved sanita- and dry toilets with discrimination. The dormi- tion more effectively. tory black townships of former “white” South Africa generally had lower grade reticulation Improving urban access to water and sanitation networks than those servicing “white” areas and In urban areas internationally, scale econo- bucket and pit toilets servicing those not on the mies make it feasible to connect each stand to network. Extensive expansion of sewered net- a networked water supply and sewer system. works has brought service to more than 2.5 mil- Building and maintaining networked water sys- lion houses in new developments, replacing more tems is capital- and skills-intensive and tends than 400,000 bucket toilets with flush toilets. to be managed by monopolies. This gives the Attempts to implement less expensive wet monopoly provider enormous power, making systems­—­such as nonsewered flush toilets with effective regulation and accountability to safe- an on-site soakpit­—failed­ in new settlements guard the needs and rights of citizens imperative. such as Ivory Park (outside Johannesburg) in the 1980s, primarily because of poor execution. Government subsidies for free basic services Figure 7.2 User satisfaction with municipal water services, 2005–2009 have muted consideration of cost-effective alter- natives to conventional flush toilets. Attempts Percent in the 1990s to implement condominal shallow 80 sewer systems­—­where users take responsibility for local network maintenance to reduce service 60 Good costs­—­have not been replicated in the 2000s. Commitment to providing networked water and sanitation services has overshadowed con- 40 Average sideration of the intensive skills required for building reticulated systems. The higher the 20 level of service, the greater the skills required to Poor run more elaborate pumping and pressure man-

0 agement systems and the billing and collection 2005 2006 2007 2008 2009 systems to recover the costs. Investment in mu-

Source: Stats SA 2010b. nicipal water services infrastructure is growing by about 26 percent a year, but municipal reve- nue is growing slower than the cost of providing Figure 7.3 Average payments for water services, 2004–2009 services (Savage, Eales, and Smith 2008). In many areas, the gap between expenditure Percent 70 and income is being funded through less spend- ing on maintenance, rehabilitation, and renewal, 60 leading to “asset stripping.” Deterioration in the

50 quality of service delivery is inevitable and likely to decrease payments and compound the down- 40 ward spiral. Recent household survey data col- 30 lected by Stats South Africa shows that this is already underway (figures 7.2 and 7.3). 20 Extending coverage has taken priority over 10 maintenance of both new and existing systems,

0 raising the likelihood of service failures. Waste- 2004 2005 2006 2007 2008 2009 water treatment failures are polluting South Af-

Source: Stats SA 2010b. rica’s water systems and threatening water secu- rity nationally.

92 Accountability in Public Services in South Africa The enormous expansion in networked ser- South Africa is investing heavily in capital- vices since 1994 has occurred concurrently with and skills-intensive municipal infrastructure, a massive loss of municipal engineering capacity, which requires operating and maintenance skills compounded by an acute national shortage of ar- not available in South African municipalities. tisans. In 1989, when local government served Engineers are being seconded into local gov- 14 million people (with the remainder served ernment services through the Siyenza Manje by provincial or “homeland” structures, or not program­—­a joint initiative between the gov- served at all), municipalities employed 2,500 en- ernment and the Development Bank of South gineers. Twenty years later, the population has Africa­—­but short-term deployments cannot grown 20 percent and municipalities are now remedy the substantial gap in skills. The short- “wall-to-wall” nationally. In 2005, only 1,300 age of experienced engineers with time to men- civil engineers were employed in municipalities, tor junior engineers, transfer skills, and build ex- and a further 10 percent have since left (Law- perience is even more distressing. less 2008, 2009; figure 7.4). This represents a fall from 23 municipal engineers per 100,000 peo- Figure 7.4 Key trends in networked water and sanitation ple to less than 3, while coverage by municipal Millions Municipal civil engineers per 100,000 people networked services has virtually doubled. 14 25 Figure 7.5 illustrates the findings of a na- Households 12 tional survey of technical capacity in municipali- 20 ties in 2005 by the South African Institute of 10 Reticulated household Civil Engineers (SAICE). The darker the shad- water connections 8 15 ing, the worse the skills gap, with municipalities 6 shaded black employing no civil engineers. Sewered household 10 flush toilets The SAICE survey found that 83 munici- 4 palities employed no civil engineering staff, that 5 2 Trend in municipal 45 had only one civil engineering professional, civil engineers per 100,000 people and that 43 employed only technicians or tech- 0 Municipal civil engineers 0 1985 1990 1995 2000 2005 2010 nologists under the age of 35, with the majority under the age of 27. Age is relevant because en- Source: Derived from Lawless (2007); Stats SA 2005, 2007; DWAF 2008a. gineering professionals maintain that it takes at least 20 years to gain the experience necessary to Figure 7.5 Civil engineering professionals in local government, 2005 head a municipal technical services department. The situation has deteriorated since this survey, with close to 100 municipalities now employing Civil engineering professionals 0 no engineers (Lawless 2008, 2010). 1 2 The engineering shortage was caused mainly 3 by the policy to correct the historical racial im- 4 ≥ 5 balance of municipal staff, but the pool of avail- DMAs able technical and managerial staff in South Af- rica does not yet align with employment equity targets. Nontechnical personnel are being ap- pointed to engineering positions in many munic- ipalities, while technical positions are left vacant in others (Lawless 2008). Public health engineer- ing is a major factor in municipal water provision, and the costs of inappropriate appointments are high and growing: essential maintenance and planning for infrastructure renewal and expan- sion is being neglected, and municipal councilors are not motivated to allocate appropriate budgets Source: Lawless 2007. for sustainable service provision.

Accountability in Public Services in South Africa 93 Deteriorating water services factors­—­ranging from infrastructure capacity to Consider the dire state of municipal wastewa- skills employed­—­and emphasized compliance ter management (box 7.1). Currently, less than with wastewater treatment requirements. In 4 percent (32 of 852) of wastewater treatment October 2009, 9 percent of municipalities were works nationally meet the criteria for a Green already in crisis, with an acute risk of disease Drop Certificate, awarded by the Department outbreaks; 38 percent were at high risk, with the for Water Affairs to works that scored an aver- potential to tip into crisis by mid-2010; 42 per- age of 90 percent or better against a number cent were at risk, including two of the country’s of defined performance criteria. Of the 449 largest metros; and just 11 percent nationally assessed, 203 scored above 50 percent, while 47 were rated as “functional” (DWAF 2009b). A percent were not assessed because they could not further assessment in March 2010 showed that provide the required data. the rate of decline was even greater than antici- The assessed risk-rating of municipal waste- pated (Eales 2010b). water treatment capacity has deteriorated fur- In April 2010, in response to widespread ther since the first round of Green Drop as- evidence of inadequate wastewater treatment, sessments in mid-2009. The assessment looked the Minister of Water Affairs announced that primarily at technical and operational risk she had requested ZAR23 billion over the next six years from the National Treasury to refur- Box 7.1 Instances of disease attributable to poor management of water services bish and upgrade municipal treatment works nationally­—­roughly half of the funding required Besides compromising the ecology of rivers and raising the costs of treating drink- (Ndaba 2010). This magnitude of expenditure ing water, discharge of inadequately treated sewage also exposes downstream users to severe health risks: is unquestionably needed, but what steps will • In April 2008, a formal inquiry by the Department of Health (2008) confirmed be taken to improve municipal management of that over 140 babies and infants died in uKhahlamba (now Joe Gqabi) District wastewater treatment and water services more municipality in the north Eastern Cape from a combination of poor nutrition, generally? How will municipalities ensure that fecal contamination of water, and inadequate hospital treatment. Further in- funds are allocated to asset management on time vestigation of municipal water services in the district revealed a range of pos- and that preventative maintenance is done before sible causes of diarrhea and suspected shigella dysentery, including erratic assets decay beyond repair? chlorination of drinking water, inadequate wastewater treatment, dilapidated The government faces huge municipal de- infrastructure, and poor process controls. • In February 2009, a senior official in the Department of Health told the Parlia- mand for capital and operational funding for mentary Portfolio Committee on Water Affairs that the cholera outbreak that water services because of mismanagement, poor spread from southern Zimbabwe into Limpopo and Mpumalanga in the sum- revenue collection, and subsidies for households. mer of 2008/2009 worsened significantly due to municipal water services fail- The cost of the current approach is unsustainable, ures: inadequate chlorination, poor maintenance of treatment works and pump calling urgently for far greater professionaliza- stations, interrupted water supplies, and provision of raw water through taps. tion of municipal water services provision and Where water supplies failed, local residents reverted to untreated water from improved performance efficiency. boreholes and rivers, some of which were contaminated by poorly treated ef- Municipal expenditure on water services fluent and sewage spills (Benson 2009). At least 51 people died from cholera in South Africa in this outbreak. has nearly doubled in the past six years from • In early June 2009, the Minister of Water Affairs, Ms. Buyelwa Sonjica, was ZAR6.8 billion in 2003/2004 to an estimated among those who attended the funeral of eight people from the Eastern Cape ZAR13.1 billion in 2009/2010 (National Trea- village of Mpheko, near Mthatha, who died after drinking water contaminated sury 2009). Concurrently, municipal depen- by human feces when the municipal water reservoir ran dry. A number of ser- dence on grants to fund their operations is grow- vice outages, linked to management of the central reservoir, interrupted the ing. The share of service charges in the total supply of water while the water table had fallen because of severe drought, and operating revenue of local government declined many local springs and boreholes had dried up. Local newspapers reported from 49 percent in 2003/2004 to 42.9 percent in two occurrences of widespread diarrhea in late May and then June 2009, con- current with water supply outages, which led to the deaths. Department of 2009/2010, mainly because of the sharp increase Water Affairs tests subsequently confirmed the presence of contaminants in in national transfers (Carrim 2009) but also be- the river water, but because they were conducted many days after villagers had cause of poor billing and collection by munici- died, the precise cause of their deaths could not be established (Sonjica 2009; palities (National Treasury 2008). Groenewald 2009; Daily Dispatch 2009). Growing grant dependence is masking sig- nificant inefficiencies in many municipalities.

94 Accountability in Public Services in South Africa Yet most municipalities do not “ring-fence” in- and obligations in water services legislation have come and expenditure for water services and stalled. thus do not really know what it costs to provide these services. Equally, what the National Trea- Authority-provider relationships sury calls “poor fiscal effort” on the part of -mu Water services accountability relationships are nicipalities (National Treasury 2008)­—­that is, spelled out in the 1997 Water Services Act. The poor billing and collection­—­means that more Act distinguishes between a municipal Water free water is provided than the national sub- Services Authority, responsible for ensuring sidy policy envisaged. Provision of free water, access to services through setting policy and or water provided far below its real cost, con- monitoring performance, and a Water Services tributes to user perception that water has a low Provider, responsible for executing service deliv- economic value; a far-reaching public awareness ery in line with the policies of the Water Ser- campaign is needed to build understanding of vices Authority (figure 7.6). the gravity of the water challenges facing the The government has designated just under country. half of all municipalities nationally as Water In the face of widespread evidence of dete- Services Authorities. Service provision can be riorating municipal water services, it is tempt- undertaken by that same municipality or it can ing to propose wholesale restructuring to con- be assigned to another entity: another munici- solidate existing capacity into fewer entities, pality; an “organ of state” such as a water board, which could make more effective use of avail- which deals primarily with bulk water supply; a able expertise. But institutional restructuring is private sector company; or a community-based enormously risky and disruptive, and it cannot organization or nongovernmental organiza- be certain that changes introduced will deliver tion (NGO). In whatever form, the latter entity positive outcomes. is known in policy and legislation as the Water It makes more sense to start by strengthen- Services Provider. ing existing local government structures by en- The Water Services Authority sets objectives suring that they are properly staffed, managed and goals, defines policy for achieving them, and professionally, and held more accountable to cit- regulates the performance of the Water Services izens. Arguably, stronger public accountability Provider in meeting them. Most Water Ser- mechanisms are the highest priority as account- vices Authorities keep water services provision ability can drive performance improvements. in-house through a municipal technical services The role of ordinary people­—­as citizens, not just department. consumers­—­in shaping and participating in ser- In the Eastern Cape, KwaZulu-Natal, Lim- vice management and decisionmaking urgently popo, and North West provinces­—­where the needs strengthening. Citizens need more power footprint of the former “homelands” was larg- to hold service providers accountable through est, backlogs are greatest, and local government short-route and long-route mechanisms. structures are generally the least mature­—­the government has assigned the Water Services Accountability relationships in the water sector Figure 7.6 The formal division of responsibility for water services provision Formal accountability frameworks in the water sector focus on relationships among spheres of Authority Provider government and between municipalities and • Ensuring universal service • Duty to provide water Formal accountability their service providers. Accountability mecha- access services mechanisms in the water sector • Planning • Effectiveness and efficiency rest on a distinction between nisms between citizens and service providers­ • Regulation • Consumer charter an Authority and a Provider • Provision • Consumer relations function. The Authority has a —short-route­ accountability—­ and­ between • Services to industry • Consumer-friendly billing local regulatory and • Regional schemes • Business plan supervisory responsibility, while citizens and their political representatives­—­ • Duty to provide information • Duty to provide information the Provider is responsible for implementation and operations. long-route accountability­—­are poorly devel- A Water Services Regulator is being established to regulate oped and perhaps atrophying. Amendments the Authority function. proposed by the Department of Water Affairs to strengthen the definitions of consumer rights

Accountability in Public Services in South Africa 95 Authority function primarily to district munici- council, so responsibility for regulation has palities and made them responsible for ensur- passed primarily to local politicians. Figure 7.7 ing service delivery. The government’s intention maps the World Development Report account- was to consolidate available capacity in districts. ability triangle as it applies to the water services Each district municipality encompasses four to sector. eight local municipalities, so municipalities and Citizens elect their political representatives, settlements served are far higher than a simple and national and provincial governments are the tally of Water Services Authorities would sug- primary framers of the compact between citizens gest. This has significant implications for -ac and the Water Services Authority. In the figure, countability as the administrative base of the the left-hand diagram shows the most common Water Services Provider can be more than a arrangement for delivery in line with this com- hundred kilometers away from where a service pact: the Water Services Authority municipal- improvement is needed. Elsewhere, the local ity is also the Water Services Provider, though it municipality itself is both the Water Services might “in-source” support from a water board or Authority and the Water Services Provider. private company. In theory, the relationship between the The right-hand diagram illustrates a separa- Water Services Authority and the Water Ser- tion between the Water Services Authority and vices Provider is regulated by a service delivery the Water Services Provider. For example, the agreement (Van der Merwe 2004). In the few City of Johannesburg contracts its own wholly cases where these agreements exist­—for­ exam- owned utility, Joburg Water, and holds it ac- ple, between the Water Services Authority and countable for delivery against specified perfor- another municipality, municipal entity, water mance objectives. Joburg Water is assisted by board, or private company­—­the outputs are not subcontractors for a range of services, from necessarily specified, and engagement with citi- building VIP toilets to reading water meters. zens about desirable outputs or feedback on per- More commonly, the external Water Services formance is limited. Provider is another municipality. In most municipalities, the distinction be- tween the Water Services Authority and the Long-route accountability relationships Water Services Provider is blurred, and no ser- between citizens and politicians vice delivery agreement defining their relation- Citizens with service problems rely primarily ship or specifying outputs or outcomes exists. on political accountability mechanisms: rais- The policy role intended for Water Services ing issues through their ward representative or Authorities is taken largely by the municipal councilor­—­where the ward committee func- tions or where the councilor is responsive­—­ Figure 7.7 Applying the World Development Report triangle to water servicing in South Africa, 2008 rather than with a service provider directly. The councilor refers the complaint through the Politicians and Politicians and municipal council to the relevant officials, who policymakers policymakers in turn instruct the service provider to remedy the problem. Municipality as Municipality as Water Services Water Services Water Services But the supply chain and accountability re- Provider Authority Authority lationships are often convoluted, with the Water

Associated service Services Provider often reliant on support from a providers—water boards, small Water Services water board, contractor, or external service pro- and medium-size Provider enterprises vider. Accountability relationships within the Associated service providers—water boards, small Water Services Provider are often diffuse, and and medium-size enterprises management oversight of local operators by a Ward committee Ward committee remote Water Services Provider­—­municipal or otherwise—­ is­ frequently weak. Citizens Citizens Citizens increasingly complain that their

Source: Authors’ analysis based on World Bank 2003. political representatives are unresponsive. This highlights a core governance challenge in South

96 Accountability in Public Services in South Africa Africa: electoral arithmetic ensures the ANC a technical services departments­—­with consider- voting patterns in majority, and political office holders owe their able fragmentation and high overheads. Debate South Africa have positions to the party rather than to voters. centers on how best to consolidate capacity to more to do with Habib (2009) points out that where tenure in maximize scale economies and promote good political office is uncertain, political elites are governance without introducing new risks and identity politics obliged to respond to the needs and wishes of discontinuities. than voicing citizens­—­political uncertainty is the essence of Lasting improvement in performance out- dissatisfaction with democracy, in other words. But the ANC enjoys comes requires institutional reforms at a range a strong majority in both national and local gov- of levels: recruiting necessary skills and compe- the performance ernment, and electoral politics in South Africa tencies, emphasizing performance contracts and of government are dominated by the party list system; the party service delivery outcomes, ring-fencing water fi- determines who serves and for how long, and nances and administration, respecting the roles political representatives are accountable primar- of councilors and officials, arranging partner- ily to the party rather than to citizens. ships with public, private, or other entities, and Several commentators have noted that vot- establishing separate utilities, among others. ing patterns in South Africa have more to do An obvious requirement is to insist that mu- with identity politics than voicing dissatisfac- nicipalities appoint managers, supervisors, and tion with the performance of government and works operators with the appropriate profes- that opposition politics play out primarily inside sional and technical expertise. The Department the ANC between factions contesting for party of Water Affairs is now considering regulating dominance. the competencies of critical water services per- Dissatisfaction with service delivery has sonnel: one option is to introduce mandatory not been reflected in local government election minimum qualifications for senior water ser- voting patterns (Booysen 2007; Schoeman and vices personnel, possibly linked to accreditation Puttergill 2007). Despite widespread service through a relevant body (Muller 2009). protests prior to the 2006 local government elec- Most municipalities take on service provi- tions, voter turnout did not fall and the ANC ac- sion without necessarily considering what ar- tually increased its majority. The same was true rangements might best serve the needs of their in national government elections in 2009, when constituents. The criteria and process for decid- the ANC again increased its overall majority de- ing on municipal services mechanisms are clearly spite extensive dissatisfaction with service deliv- detailed in Section 78 of the 2000 Municipal ery. Thus even the long route to accountability­—­ Systems Act. Framed amid strong organized voting­—­has limited ability to influence service labor opposition to the privatization of munici- improvements until the elected representatives pal services, Section 78 is intended to discourage prioritize a comprehensive turnaround. outsourcing by requiring careful consideration of Growing revolt can be interpreted as a prom- how to enable in-house service provision. ising sign that this turnaround is under way. As Section 78 does not always serve the public discussed in chapter 1, in response to poor ser- interest. Many municipalities need to acknowl- vices, unresponsive councilors, corruption, on- edge that they lack the capacity to deliver effec- going job losses, and other grievances, protest tive and affordable universal services. In some action soared in 2009. As many protests were re- municipalities, a nonmunicipal service provider­ corded in 2009 as in the previous five years, with —­a local small enterprise, a consortium of ser- disturbingly high levels recorded in early 2010 vice partners, or a professional utility­—­might (Municipal IQ 2010). offer a better service. The Municipal Systems Act favors the ap- Strengthening institutional delivery systems pointment of internal water services providers Local government is responsible for service through requiring the municipality to assess, delivery, but debate is growing over service pro- for example, whether it could reorganize its ad- vider arrangements. Sector skills are dispersed ministration to develop the necessary human across several hundred municipal Water Ser- resource capacity to provide the service. The vices Providers­—­the majority being municipal municipality is required to consult only with

Accountability in Public Services in South Africa 97 many municipal organized labor on this assessment. If the mu- Whose interests should prevail when ex- workers lack nicipality concludes that it has, or can develop, ploring service provider options, if the objec- the skills for the the capacity to provide the service, Section 78.2 tive is to provide all citizens with a safe and permits it to exit the decision process, fully com- reliable water supply? The Water Services Au- water services pliant with the law. Most Water Services Au- thority is required to consult citizens on pro- functions they thorities take this route and appoint or establish vider options and arrangements only if, having perform internal technical services departments as the considered internal options, it still wishes to Water Services Provider. It seems that most mu- explore external options. Even then, the mu- nicipalities prefer to avoid out-sourcing to keep nicipality is only required to alert citizens that jobs, funding, and control in-house. A broader it is assessing possible service provider options. menu of options is listed in box 7.2. Only in this external review phase of the as- Recent research shows that many munici- sessment process is the municipality obliged pal workers lack the skills for the water ser- to consider the cost-effectiveness of different vices functions they perform; moreover, the job options, performance benchmarks, or whether requirements exceed the training potential of options favor the poor. If the intent of this leg- many water services operators (Momba, Osode, islation (Government Regulation 8534 of 1985 and Sibewu 2006; Momba, Obi, and Thompson or Section 62 of the Water Services Act of 2008; Schoeman 2006). As Muller and others 1997) was to inhibit outsourcing or appoint- have noted, a medical doctor requires at least ment of private sector agencies, it has been seven years of rigorous training before being al- very effective: no public-private partnerships lowed to practice on the general public (Muller and no appointment of private sector Water 2007). Yet compliance with regulations2 that Services Providers has occurred since the Act safeguard public health by specifying the quali- was promulgated. fications and performance requirements of water Appointment of an external water provider care operators are not monitored (Boyd 2004; must follow competitive tendering, unless an- DWAF 2008c), even though the actions of op- other municipality or public entity is selected. erators affect the health of potentially tens of This means that even community-based organi- thousands of people or more. zations and NGOs must tender competitively to be considered as potential Water Services Pro- Box 7.2 Examples of current institutional viders. Most municipalities appear to regard models for water services provision using community-based organizations as a form of outsourcing and prefer to absorb village op- • The Water Services Authority (WSA) under- erators into their administrations, even if the takes the Water Services Provider (WSP) role costs remain similar to previous payments to lo- itself­—­the dominant approach. cally funded village water committees (Balfour • WSA appoints one or more local municipalities as a WSP. 2008b). Residents of a small town or village have • WSA establishes and appoints its own wholly little power to affect the selection of their service owned retail water utility. provider, even if their service needs might not • WSA contracts a multijurisdiction municipal- require them to be networked to a regional dis- owned entity established to serve more than tribution and treatment scheme. one WSA. A local small or medium-size enterprise or • WSA contracts a water board to assist with community-based organization can only be ap- WSP responsibilities. pointed as a Water Services Provider if the • WSA contracts a concessionaire. • WSA contracts a private sector entity, which Water Services Authority opts to look beyond leases the assets. internal service provision and if the enterprise • WSP signs a management contract with a pri- or organization is able to outbid other contend- vate sector support agency. ers in a competitive tender. The combination of • WSA contracts community-based organiza- legislation and municipal preference rules out tions, with a services agent to support com- virtually all local management options. Most munity-based organization providers. rural water schemes are now run by operators who report and account directly to the municipal

98 Accountability in Public Services in South Africa water services provider, with very little local (potable water) and Green Drop (treated efflu- South Africa has accountability. ent) certification system in 2008 to incentivize decentralized Evidence suggests that service failures are municipalities to comply with high management delivery but widespread due to a variety of reasons: manda- standards. A “name-and-shame approach” for tory supply chain management policies make it poor performers is proving effective in pressur- has opted for difficult to stock spare parts, funding for fuel or ing municipalities to raise treatment manage- a highly state- energy is sometimes transferred late, chlorine ment standards. centric approach supplies run out or dosing is erratic, and vandal- The lines between assistance, enforcement, ism and theft of vital components is common. and intervention by the Department of Water through municipal But resistance in most municipalities to formal Affairs are sometimes blurred: it supports mu- structures outsourcing is significant, apparently because of nicipalities in building capacity to meet qual- a strong desire to retain jobs in the public sec- ity standards, sets standards, monitors compli- tor. A recent Water Research Commission in- ance, and intervenes in cases of noncompliance. vestigation into the feasibility of franchising op- Assigning these functions to different divi- erational and maintenance support services to sions within the department is not sufficient to municipalities, with franchisees supported by a overcome this tension. The Department prefers regional technical support team, noted that the to work with a municipality to rectify a non- most significant obstacle was the reluctance of compliance problem rather than to litigate, as most municipalities to outsource any functions it believes the former to be more constructive. (Bhagwan, Wall, and Ive 2008). Moreover, the constitutional principle of coop- Municipal use of publicly owned water erative governance between the three spheres of boards to support water services provision is government (national, provincial, local) has left growing, however. As organs of state, water the national sphere reluctant to prosecute failing boards are exempt from competitive tendering municipalities. procedures. This is a double-edged sword: it re- A welcome recent shift in policy has oc- duces time and cost, but the efficiency and -ef curred, however, with the Department of Water fectiveness of the water board is not subject to Affairs adopting a far tougher stance against the scrutiny of a tender assessment, and a water municipal negligence in the discharge of water board will not necessarily serve citizens better services responsibilities. Directives have been is- than a community-based or private entity. sued against a growing number of municipali- ties, requiring them to remedy noncompliant Regulation of water services behavior. And, in an unprecedented step in early The absence of effective regulation to drive 2010, the Department pressed charges against national accountability has led to few sanctions two municipalities, Madibeng in North West against municipalities that do not comply with Province and Matjhabeng in the Free State, sector legislation, few incentives for municipali- for persistent wastewater treatment failures. It ties to closely monitor the performance of their remains to be seen whether individuals will be service providers, and limited attention to the prosecuted: a major shake-up in municipal ser- outcomes from infrastructure investment. The vice performance, and perhaps even the selection Minister of Water Affairs has been required to of service providers, could occur if a municipal establish an independent economic regulator manager is held personally liable for noncompli- for the water sector by 2014, but work on the ant wastewater discharge. scope of this regulatory function has only just begun. Regulation of noneconomic performance Strengthening citizen-user voice and power parameters is likely to remain with the Depart- Galvin and Habib (2003) note that South Africa ment of Water Affairs for some time. has decentralized delivery but has opted for a The focus of the Department of Water Af- highly state-centric approach through munici- fairs’ current regulatory activity is strengthening pal structures rather than a more integrated and management of drinking water and wastewater inclusive model, which would harness people’s resources. The Department introduced an inno- support and creativity in planning and sustain- vative and internationally acclaimed Blue Drop ing services. To a large degree, ordinary people

Accountability in Public Services in South Africa 99 have been marginalized in the process of munic- the service framework and the compact linking ipal service planning and delivery and­—­as the citizens, the municipality, and the service pro- Reconstruction and Development Programme vider. Establishing new mechanisms for public warned­—­have become the objects, not subjects, recourse without first building public knowledge of development (ANC 1994). to monitor service performance is pointless. The most urgent priority in bringing peo- Reciprocally, improved understanding of ple back to the center of decisionmaking is to citizen-user perceptions and needs helps mu- strengthen public education and information. nicipalities and service providers deliver what Some recipients of new services have not had people want. The eThekwini Metro is working taps or toilets and are not aware of the tech- to improve water and sanitation services through nologies’ vulnerabilities and their roles in basic a range of initiatives to build trust and com- maintenance. Many have limited awareness of munication with citizens. It conducted 20 focus how these services fit within the broader frame- groups and over 200 citizen interviews to drive work of natural, financial, and human resource some immediate service improvements (box 7.3). management, and most did not grow up paying In response to concerns about water affordabil- for water or services. Few citizens know how to ity, for example, the Metro increased Free Basic interact with municipalities or service providers. Water from six to nine kiloliters for qualifying The eThekwini Metro, the municipal area households starting July 2008. Metro officials serving the large port city of Durban, is a rare are also taking citizen report cards to new lev- example of a municipality that for many years els, with an equal emphasis on qualitative and has been working to improve public understand- quantitative information across a wide range of ing of water services. The municipality has used themes: trust and satisfaction, billing and pric- street theater, house visits, school programs, and ing, health and hygiene, infrastructure, envi- pamphlets to educate users about the vulnerabil- ronment, and conflict (Wilson, Malakoana, and ities of sewered sanitation and about user rights Gounden 2008). and responsibilities. South Africa is a young democracy and citi- While resource-intensive, Metro officials zens’ awareness of their own rights and power can prove that the benefits of the public educa- is still maturing; building awareness of rights tion program far outweigh the costs and that and responsibilities will take time. A sobering the program reduces spending (Gounden 2008). reminder of the task ahead lies in the findings of The first step to create short-route accountability a recent consumer survey (Sigodi Marah Mar- mechanisms are to build public understanding of tin 2007): roughly half of the citizens surveyed

Box 7.3 Focus groups in eThekwini Metro had never heard of the Constitution, the ma- jority were not aware of Free Basic Water, and In eThekwini Metro, “focus groups have served to significant confusion about the meaning of the identify where local service centers keep irregular rights prevailed among those who were aware hours, have helped to inform common complaints (box 7.4). These findings raise profound ques- with contextualized local knowledge, thereby al- lowing staff to better understand root causes of tions about the efficacy of current public aware- the most frequent triggers of low service satisfac- ness campaigns and how much many citizens tion levels. have disengaged from public life. Discussions have also shed considerable light on local attitudes towards and impact of illegal User platforms connections and ways to identify households in A monthly local user platform involving the crisis where stigma prevents people from coming ward councilor, citizens, and representatives of forward for assistance, among other issues. Call- the service provider can be valuable for short- centre staff also report that people calling in from areas where the study has taken place are better route accountability by helping to build the informed and better able to direct their complaints relationship between service providers and towards effective resolution.” citizens beyond the monthly bill. Insight into local service challenges and opportunities can Source: Wilson, Malakoana, and Gounden 2008, p. 145. also bolster other public forums such as ward committees.

100 Accountability in Public Services in South Africa Box 7.4 Citizen awareness of rights and • The eThekwini Water and Sanitation (EWS) Enormous responsibilities has a strong track record in innovative ser- water services A 2007 survey of a representative national sample vice provision. Learning from Cape Town’s achievements in of 2,456 urban adults and 900 rural adults showed experience, in 2008 EWS designed a user that: platform that engaged important political the first decade • Just 41 percent of respondents nationally had decisionmakers from the start. Weekend of democracy in heard of Free Basic Water; a quarter of those training workshops for councilors and civil who had heard of Free Basic Water­—­15 per- South Africa may society organization representatives convene cent of the sample overall­—­believed it meant be compromised in “Water is for free­—­as much as you like.” in all 17 city zones, and a growing number • 46 percent of respondents (57 percent in rural now support monthly user platform meet- the second decade areas) had not heard of the Constitution. ings. The city funds the initiative from its • Among those who had heard of the Constitu- core budget. Councilors participate actively tion, 93 percent believed that, constitutionally, in each zone as they find that it supports everyone had the right to a flush toilet, and 95 their work in wards. percent believed that everyone had the right to a tap inside their home. • In Msunduzi’s local municipality, initial progress in training and establishing user Source: Sigodi Marah Martin 2007. platforms stalled when the municipality was declared bankrupt and put under admin- istration. The new mayor saw the Citizens’ User platforms have enormous potential to Voice training as a valuable communication rebuild trust among citizens, municipal officials, tool to rebuild trust with local residents and and councilors in support of service turnaround provide an alternative to service delivery pro- strategies and objectives. Experience in setting tests plaguing the city; the Citizens’ Voice up “Citizens’ Voice” initiatives in four cities initiative has been incorporated into the highlights the opportunities and vulnerabilities city’s planning and communication work. of this approach. • In Ekurhuleni, enthusiastic support from • In Cape Town, the Department of Water hundreds of residents who participated in a Affairs ran a successful pilot Citizens’ Voice series of community training sessions was project for four years, establishing user thwarted when officials refused to allow platforms run by volunteer committees in users to engage with councilors, fearing that 23 townships. Participants were trained to the establishment of user platforms would understand themes in water services man- threaten the councilors. agement and to articulate concerns to mu- nicipal water officials. The user platforms Conclusions and policy proved highly effective, with participants recommendations raising issues later escalated to senior city Enormous water services achievements in the decisionmakers. However, the initiative first decade of democracy in South Africa may stalled when Department of Water Affairs be compromised in the second decade if swift funding ended: high turnover in municipal measures are not taken to mitigate the conse- water service management and a decision quences of the current supply-driven approach not to engage councilors left no champion to service provision. South Africa’s state-centric for the project. Despite progress in build- approach to service delivery is top-down, target- ing understanding of water demand man- driven, supply-led, and generously grant-funded. agement and conservation, the city opted to This approach skews infrastructure investment abandon the Citizens’ Voice project in favor decisions, overestimates the skills available to of a hardware approach to water manage- operate and maintain services, risks high service ment, introducing restriction devices per- failures, and limits citizen power. ceived to be punitive and discriminatory by Public accountability in the water ser- the community. A more positive outcome vices sector is focused primarily on the re- could have been achieved if the user plat- lationships among spheres of government­—­ form had continued. recommendations were made for improvement

Accountability in Public Services in South Africa 101 Citizens need in these relationships­—­while the relationship Section 78.2 to remove the “and/or” option direct, short-route between citizens and municipal service provid- and mandate the assessment of external op- accountability ers is given less attention. The primary respon- tions. Mandate the review of Section 78 sibility for holding service providers accountable decisions when the performance of the mu- mechanisms should not lay with a remote regulator because nicipality consistently fails to meet defined such an approach places even greater emphasis minimum standards. on the relationship between government and • Mandate performance contracts between all providers, further marginalizing citizens. Water Services Authorities and their Water Citizens need direct, short-route account- Service Providers, irrespective of whether ability mechanisms. Accountability must be the Water Services Provider is the munici- rooted locally, on the ground, and at the inter- pality’s internal technical department or an faces between citizens, service providers, and external service provider. elected representatives. Citizens should have recourse through the regulator, but primary Strengthen local accountability relationships engagement and response must be local, where for operators serving settlements remote from problems are felt and where immediate interven- the municipality’s administrative hub. tion is usually required. • Promote the use of locally accountable oper- ators and operational and maintenance crews Strengthen the compact between the to augment service provision capacity. The government and service providers. current reliance on underskilled local opera- • Regulate the competencies required by per- tors and remote technical support saps power sonnel filling critical water services func- from both citizens and operators. tions. Require water services institutions to • In rural areas, where local water operators use qualified personnel to manage water sup- are employed by a remote Water Services ply and sanitation services. Establish certi- Provider (municipal or other), explore local fied water services managers with the mini- mechanisms to hold service providers ac- mum qualifications established under the countable. Consider aligning settlement- National Qualifications Framework. Pro- level structures with the municipal service mulgate regulations in the Water Services provision framework. Operators would be Act to ensure applicability to all water ser- employed by the Water Services Provider or vices institutions. its agent, but the local structure would share • Strengthen public accountability in the se- management and oversight responsibility. lection and appointment of service provid- ers. Municipalities are increasingly enlisting Strengthen financial management to support water boards­—­whose capacity and cost-ef- sustainable water services provision. fectiveness vary enormously­—­to support ser- • Mandate financial ring-fencing of municipal vice provision but without procedures requir- water services in every municipality within ing water boards to demonstrate that they five years. Economic regulation of water ser- are the most appropriate provider. Introduce vices is not feasible without reliable informa- measures to improve transparency in mu- tion on water services income and expendi- nicipal appointment of water boards to pro- ture for each municipality. vide water services. All appointments should • Revise the municipal Equitable Share grant be regulated by contract, with the scope of to make transfers supporting provision of work, performance outcomes, costs, and du- basic water and sanitation to poor house- ration of appointment specified clearly, and holds conditional on delivery of a municipal subject to public comment. report. • Review Section 78 of the Municipal Systems • Review the conditions for accessing Mu- Act to promote a more transparent, inclu- nicipal Infrastructure Grants to incentivize sive, and integrated approach to assessment more sustainable infrastructure investments. of service provision options and mecha- • Use Municipal Infrastructure Grants to in- nisms. Particularly, review the phrasing of centivize better asset management. Explore

102 Accountability in Public Services in South Africa instruments that will reward municipalities systems are improving across the water sec- for good asset management, for example, by tor but from a very low base. Citizens must providing Municipal Infrastructure Grants play a vital role in monitoring and making above a basic threshold only to municipali- information available to the regulator to but- ties that demonstrate good asset manage- tress the regulator’s effectiveness. ment practices. • Establish local user platforms to bring citi- zen, municipal, and service provider repre- Strengthen citizen-user voice and power. sentatives together. • Dedicate significant resources to building public understanding of how services work Notes and to help citizens understand their rights, 1. References to “urban” and “rural” areas sug- roles, and shared responsibilities in service gest a clear dichotomy of settlement types; provision. Dialogue will be strengthened it is more useful to refer to an urban/rural significantly if the municipality and its pro- continuum­—from­ urban to peri-urban, vider invest in understanding citizens’ needs semi-urban, or rural­—­as this acknowledges and perspectives. different contours of settlement density and • Strengthen citizens’ recourse through close economic activity. Space does not permit a partnership with the sector regulator. Sector full discussion of technology options across regulation is only as effective as the informa- settlement typologies or the layers of com- tion available, and both Water Services Au- plexity added by transient populations, inse- thorities and Water Services Providers have cure tenure, and informal settlement. a strong incentive to disguise poor perfor- 2. Government Regulation 8534 of 1985 or mance. Current monitoring and evaluation Section 62 of the Water Services Act of 1997.

Accountability in Public Services in South Africa 103

8 Toward Local and Community‑Driven Development

What if the choice, design, implementation, and overlapping social links and functions. Power- monitoring of services could be decentralized to ful community leaders or members can subvert communities? The principal would become the these accountability links to shift benefits in their agent and incentives would align with informa- favor, a process called “elite capture.” In addition, tion. For services where production technology communities remain accountable for the proper is not characterized by economies of scale, where use of resources to the government or private or- skills are available locally, and where outcomes ganizations that cofinance their investments. are easily monitored by communities, decentral- This chapter will discuss simple but effec- ization to the community­—­and in some cases to tive mechanisms for internal accountability households­—­is a genuine option for organiza- and upward accountability to funders. The ori- tion of service delivery. gins of community-driven development will Community-driven development has put be explained, along with its evolution from ap- these ideas into operation. As experience with proaches based on community consultation and this approach accumulated internationally, it participation toward integrated local and com- became clear that good service outcomes usu- munity-driven development­—­an evolution that ally require other stakeholders­—­notably local has often been affected by conflicting centraliza- governments­—­to become involved as co­ tion and decentralization goals. This approach producers of services. The model is now called could become an important model for selected local and community-driven development. services in South Africa. Community-driven development transfers most accountability mechanisms to the commu- Approaches toward community and nity and has the potential to create the shortest local government empowerment route to accountability. Even here three legs of Community empowerment approaches come in accountability need to be considered, between: different forms (Binswanger-Mkhize, de Regt, • Community members and community and Spector 2010). The first approach, commu- leaders. nity-based development, has two variations: • Community leaders and the subcommittee • Community consultation model, where govern- executing the project. ment agencies or nongovernmental organiza- • Community members and the subcommittee tions (NGO) consult communities but operate executing the project. as direct service providers using their own staff. The three legs of accountability are blurred by • Community participation model, where govern- the proximity of community members and their ment agencies or NGOs invite communities

Accountability in Public Services in South Africa 105 Successful to participate in choosing development prior- • The power to plan, implement, and maintain local and ities and project designs, cofinancing projects projects to serve their needs. community-driven with in-cash or in-kind contributions and op- • The right to hold politicians and officials erating the projects. accountable. development Community-based development can gradu- • The power to command local bureaucrats in- is seen as a ally build community capacity, but it stops short stead of being their supplicants. coproduction of of devolving funds and implementation responsi- • The power to hire, fire, and discipline all bility to communities. Most community empow- who provide frontline services, such as edu- communities, erment programs in South Africa are community cation, health, municipal, and agricultural local government, consultation or community participation models. services. and the sectors The second approach is community-driven • The right to a share of central revenue. development.1 Normally, this approach gives • The power to levy user charges and taxes. funds to communities and devolves to them all or part of project implementation. Government Centralization versus decentralization agencies and NGOs operate primarily as facili- Centralization was in vogue after World War tators and trainers. Successful examples include II. The Soviet, Keynesian, and welfare state the Aga Khan Foundation in Pakistan and India. models all posited a strong central authority as In Burkina Faso, the community funds originate the engine of progress. Developing countries, as intergovernmental fiscal transfers: the govern- believing in dependency theory or fearing neo- ment provides unconditional grants to communi- colonialism, felt that a strong central govern- ties as part of central revenue sharing following ment was essential for economic and political a specific allocation formula. Communities aug- independence. In contrast, most developing ment these resources through cofinancing in cash country populations were overwhelmingly rural and in kind and through collection of user fees. In and poor, making rural development­—­requiring South Africa, however, few projects or programs decentralization­—­another fundamental goal. use this model of community empowerment. India epitomized this duality between cen- In early experiments with community-driven tralization and decentralization. Mahatma Gan- development internationally, it was found that it dhi advocated highly decentralized development could not function in an institutional vacuum: through what he called “village republics,” but local government coordination and support was the Indian Constitution created a fairly central- required from specialized sector agencies. This ized polity. The post-colonial era thus began led to the third approach: local and community- with two opposite perspectives on managing fu- driven development, which links community ture development. The development discipline empowerment, local government development, has only recently begun to find a balance be- and adapted sector programs into a coherent, tween centralization and decentralization. coordinated framework. Successful local and Most developing countries became far more community-driven development is seen as a co- centralized than developed ones. Both India production of communities, local government, and Bangladesh in the 1950s and 1960s pio- and the sectors, whose links create synergies and neered models of community development, and reduce duplication and fragmentation into silos India was the first country to implement com- and spheres of government. munity development over the entire country. Local and community-driven development Since the 1950s, dozens of countries have em- emerged as a response to the inefficient cen- barked on community development and other tralized approaches. The approach attempts to rural development programs. Most of these maximize accountability relationships along the programs started with similar objectives of de- long and short routes to accountability by em- centralized and participatory decisionmaking, phasizing the following rights and powers of local planning and coordination, and the de- communities: velopment of sustainable local and community • The right to be treated as people with capa- institutions. bilities, not as objects of pity or passive re- Yet, for both technical and political rea- cipients of services delivered to them. sons, the process in most countries fell short of

106 Accountability in Public Services in South Africa Box community empowerment. In most large-scale 8.1 The Comilla model of rural development programs, power and implementation shifted The Comilla model of rural development was created by the Bangladesh Academy back to central agencies and their technical staff, for Rural Development led by Akhter Hameed Khan. The model sought to approach and programs became highly bureaucratic (boxes rural development from the viewpoint of villagers­—­because they have the best un- 8.1 and 8.2). The preceding chapters of this re- derstanding of their problems­—­using the community participation model. Its main port suggest that the South African experience element was the creation of an institutional base in rural society around which development programs could be integrated. The institutional base included a two- since 1994 has followed a similar path. tiered system of agricultural cooperatives for savings, credit, and extension­—­and a By the 1990s, public choice theory had strong local government. It sought to create a cadre of institutional leaders in every gained wide intellectual acceptance and showed village to coordinate the activities of government departments and people’s orga- that a strong centralized state could be preda- nizations. The Thana Irrigation Program, for instance, provided irrigation through tory rather than benevolent. Economic failure participatory planning and implementation by communities. and rural neglect were in many countries at- After a highly successful pilot in Comilla district, the government created a tributed to over centralization and top-down national Integrated Rural Development Program in 1972 to extend the model na- approaches. The programs, generally unable tionwide. This was supervised by the Bangladesh Rural Development Board, which eventually dominated rural development. In the process of scaling up, planning and to grasp the priorities of communities or to re- coordination of programs shifted from communities and local bodies to bureau- spond to their needs, foundered amid agency crats. The program drifted back to the community consultation approach, with coordination problems. The collapse of the So- implementation through government agencies and later through NGOs. Although viet Union caused further disillusion with com- district assemblies were created and local elections took place, the government mand systems: democratic systems replaced mil- never transferred financial resources and implementation responsibility to the dis- itary autocracies in much of Latin America, for tricts or to the communities. This negated one of the main drivers of success of the instance, and participation started to show its Comilla model, and the program foundered. power in sectoral development programs. Source: Binswanger-Mkhize, de Regt, and Spector 2010; Banglapedia 2005. Participation in sectoral programs Box 8.2 The contest between centralization and decentralization in Sectoral programs created opportunities for var- India, 1948 to present ious forms of community participation. In a sur- When India became independent, S.K. Dey piloted community development proj- vey of 121 completed water supply projects with ects in 1948 using community participation approaches within a deconcentrated different participatory mechanisms in 49 coun- service delivery model to cover all aspects of rural development: agriculture, roads, tries, substantial benefits accrued to the more animal husbandry, health, education, housing, transportation, and communica- participatory projects, even after controlling for tions. The pilots were quickly scaled up to a nationwide community development 18 other variables (Narayan 1995). Of the 121 program, and a special ministry for community development was created with Dey projects, only 21 percent received a high rating as the minister. Community Development Workers were posted in villages to assist for participation. The degree of community par- communities with their development priorities. However, the approach was quickly ticipation varied widely from user committees bureaucratized and in the end failed to deliver on its promises. In 1957, another champion, Balwantrai Mehta, headed a committee that found to direct community construction and supervi- lack of decentralization to be a root of the failure of community development. It rec- sion of contractors, while effective participation ommended a three-tier system of panchayati raj (local governments at the district, generally did not occur wherever sector agencies development block, and village levels). But implementation was left to state govern- retained control over implementation. ments, and politicians in state capitals wanted to retain power and resources. They The Bamako Initiative is an outstanding ex- held few local government elections and took over management of many municipali- ample of applying participation and many of the ties. Contrary to announced intentions, community empowerment and democratic de- community-driven development principles in centralization were minimal. By the end of the 1960s, most programs had shifted back the health sector (box 8.3). to state governments, but centralized service delivery yielded unimpressive results. In 1989, Prime Minister Rajiv Gandhi championed village empowerment and set The Kampung Improvement Program, in motion a constitutional amendment to provide for regular panchayat elections started in the 1970s in Indonesia, was an exam- and for State Finance Commissions to suggest the transfer of sector programs and ple of a community-based approach to upgrading funds from state governments to local governments. However, weak implemen- slums. This program consulted beneficiary com- tation and outright sabotage by state governments was again widespread. Two munities in Jakarta, which also contributed to states (West Bengal and Kerala) ruled by the Communist Party, pursued earnest improvement costs. The program upgraded some decentralization as a route to electoral success. Most other parties and states did 7,700 hectares of kampungs occupied by 3 mil- not. The struggle between decentralization and centralization continues to this day. lion people. Roads, footpaths, drainage canals, Source: Binswanger-Mkhize, de Regt, and Spector 2010. water supply, sanitation, solid waste disposal,

Accountability in Public Services in South Africa 107 Box 8.3 The Bamako Initiative in Benin, Mali, and Guinea through 9,000 community nutrition centers set The Bamako initiative increased the access, availability, affordability, and use of up in the state. A little over 40 percent of proj- health services. The approach focused on establishing community-managed health ect funds went to upgrading personnel at local centers serving populations of 5,000 to 15,000 people. Priorities included: health nutrition facilities. Appointees were se- • Implementing community-owned revolving funds for drugs with local retention lected in consultation with communities; they and management of all financial proceeds. had to be local, educated residents. Care pref- • Revitalizing existing health centers, expanding the network, and providing erence was given to poor women with healthy monthly outreach services to villages within 15 kilometers of facilities. • Increasing social mobilization and community-based communication. children. The project reduced severe malnutri- • Pricing the most effective interventions below market, through government tion by one-third to one-half in children ages 6 and donor subsidies and through internal cross-subsidies within the system. to 24 months and by one-half in children ages 6 • Analyzing, with community participation, biannual progress and problems in to 60 months. health service coverage and in service planning and budgeting. • Tracing and tracking defaulters. Social action funds • Using community representatives to increase demand. Social funds, or social action funds, often began as emergency programs aimed at getting projects The results Over the more than 10 years of implementation in these three countries, commu- quickly to communities in need or at mitigating nity-owned services restored access to primary and secondary health services negative social impacts of structural adjustment for more than 20 million people. Health outcomes improved, health service use programs. The multisector funds gave communi- increased significantly, and the poor-rich gap narrowed. ties the opportunity to specify their subproject Immunizations increased in all three countries, reaching close to 80 percent priorities. Social funds used intensive commu- in Benin­—­one of the highest rates in Sub-Saharan Africa. (Immunization rates re- nity consultation and community participation mained lower in Guinea and Mali, largely because of access problems.) The use of models and were soon devised for countries health services by children under five in Benin increased from less than 0.1 visits across all continents. As of 2005, these funds per year to more than 1.0. In Mali, exclusive breastfeeding and the use of profes- sional services for antenatal care, deliveries, and treatment of diarrhea and acute have spent around $10 billion worldwide. respiratory infections increased for all groups, including the poorest. Greater ac- An evaluation of social funds in 2005 by the cess reduced travel costs, and the availability of drugs reduced the need to visit Independent Evaluation Group of the World distant sources of care. Prices have been kept below those of alternative sources. Bank found that 96 percent of closed social funds had satisfactory outcomes, against 71 per- Source: World Bank 2003. cent for all World Bank projects. Social funds worked especially well in post-conflict situations schools, and health clinics were all built using a (such as those in Cambodia and Nicaragua). The community-based approach. evaluation rated institutional development as In donor-assisted forestry programs in substantial in 65 percent of projects, better than Nepal, trees on government-owned plantations the average of 36 percent for all World Bank had a mortality rate of over 80 percent. Starting projects. in 1989, the Hill Community Forestry Project in However, the effectiveness of social funds the degraded mid-hills gave complete usufruct was attributed in part to their autonomy from rights over forest produce to local user groups line ministries, and sustainability was rated as consisting of local villagers as well as all others “likely” for only 43 percent of projects, against who used the forest, including nomads. Evalua- 51 percent for all World Bank projects. Mainte- tions estimated the economic rate of return at 12 nance responsibilities and obligations were often percent for an earlier project and 18 percent for a not clearly specified, nor were ownership issues later one with community control. Deforestation after project completion. Also, while communi- slowed to 0.2 percent a year, compared with the ties were usually asked to participate widely in 1.3 percent found elsewhere in the area (Aiyar discussions on priorities, the priorities were not 1996). always met. For example, Tendler (1999) showed The Tamil Nadu Integrated Nutrition Pro- that some Latin American social funds were gram in India over 1980–1990 targeted preg- very supply-driven. Over time, social funds such nant, lactating women, children under age as the Malawi Social Action Fund moved from three, and the children’s mothers. Under- community consultation to community partici- weight children were given additional nutrition pation models.

108 Accountability in Public Services in South Africa Community-driven development and Figure 8.1 Index of sector decentralization in 19 countries decentralization Index of sector decentralization In the early 1990s, Decentralization, Fiscal Sys- 8 tems, and Rural Development2 looked at several sectors and the powers of local institutions over 6 service delivery in 14 countries and 5 provinces.3 The study showed that the best performers were more decentralized. Jiangxi province, in strongly 4 decentralized China, topped the list, followed by other decentralized countries: Colombia, the 2 Philippines, and Poland. The lowest score came from countries with little or no decentralization: 0 Imo State in Nigeria, Côte D’Ivoire, Senegal, Egypt Chile Senegal Tunisia Zambia Poland and Burkina Faso (figure 8.1). Tanzania Karnataka Colombia Côte d’IvoireBurkina Faso Bangladesh Bahia (Brazil) Jiangxi (China) Hidalgo (Mexico) The Philippines From a study of Colombia’s decentralization, Imo State (Nigeria) Punjab (Pakistan) Fiszbein (1997) concluded that competition for political office resulted in accountable and inno- Nusa Tenggara Timur (Indonesia) Source: McLean and others 1998; Piriou-Sall 1997; Aiyar 1995; Aiyar, McLean, and Piriou- vative local governments that improved service Sall 1996. delivery and reduced corruption. Faguet (1997) showed how decentralization improved account- ability and reduced poverty in Bolivia. Crook and community. Important program innovations and Manor (1998) showed that decentralization included simplified procurement and disburse- yielded good results only where the following ment procedures, which have allowed the direct existed: transfer of funds to communities. • Strong government ownership. A similar program has been implemented in • Appropriate legal, administrative, and fiscal Northeast Brazil, but without being part of the arrangements. intergovernmental fiscal system. A recent rigor- • Local elections. ous evaluation shows that the program is well • Sufficient and reliable funding. targeted to poor communities and has led to im- • Freedom for communities to choose projects. proved water supply and electrification, reduced Community participation and decentraliza- infant mortality and incidences of communica- tion were often introduced in countries inde- ble diseases, and sustainable social community pendent of one another, but they clearly needed capital. to be integrated to improve sustainability and The National Program of Independent Com- maximize synergies. In the 1990s, major proj- munity Empowerment in Indonesia, a commu- ects in Mexico, Brazil, and Indonesia adopted nity-driven development program, started as this approach. a Kecamatan Development Project in 1998­—­a In Mexico, rural municipalities obtained time of political upheaval and financial crisis. It unconditional grants based on a formula target- began just as Indonesia started its massive de- ing poor municipalities. These funds were ear- centralization program, also triggered by the marked only for investment projects identified political and economic crisis. The program aims and executed by communities within the mu- to improve local governance and reduce poverty nicipality; they could not be used for recurrent and corruption by channeling funds directly to municipal budgets. Project priorities were set by local government and community institutions communities, which contributed 20 to 40 per- in Indonesia’s poorest rural communities. The cent of project costs. Municipal Development program provides block grants of approximately Councils headed by elected mayors selected the US$50,000 to US$150,000 to subdistricts (Ke- projects and supervised the execution and fund camatan) depending on population size. Vil- management by the communities. Municipal lagers engage in a participatory planning and funds used a learning-by-doing approach across decisionmaking process where they discuss pri- all levels of operation: federal, state, municipal, orities, choose from an open menu (meaning

Accountability in Public Services in South Africa 109 they largely create their own topics), and propose and downward accountability for funds and the projects for the Kecamatan. The Kecamatans associated procurement practices. Six new ap- then choose the best projects, ask the communi- proaches and mechanisms were pioneered in ties to contribute part of the funds, and provide Mexico, then tested in many countries, and pro- matching grants for community implementa- gressively improved in other countries such as tion. The Kecamatan Development Project em- Brazil and Indonesia (box 8.5). phasizes transparency and information-sharing In most sectors in South Africa, the trans- throughout the project cycle. Decisionmaking fer of funds directly to communities often meets and financial management are open and shared with widespread resistance from the govern- with the community. ment, and even NGOs. The main argument Practitioners of community empowerment against this direct transfer is that South African and decentralization approaches have identified communities lack the capacity and social capi- five main governance transformation dimensions tal to undertake the types of community projects of local and community-driven development: that have worked elsewhere in the world. Of- • Empowering communities. ficials also often believe that transfer of public • Building capacity. funds to entities outside government is illegal. • Empowering local governments. Even in explicitly community-driven programs­ • Realigning the center away from delivery of —­such as the land redistribution program­—­this all services. interpretation often results in government of- • Improving accountability. ficials tendering on behalf of communities for simple inputs such as seeds for a land reform Accountability mechanisms project or for simple services such as plowing. All three legs of the long and short route to Officials are practically running these projects, accountability are internalized into the commu- making them very difficult to scale up. This- in nity. A simplified upward accountability to local terpretation also saps power from the beneficia- government remains. Adherence to a number of ries, leading to dependency and to poor impact long- and short-route accountability design prin- on communities. ciples enhances successful outcomes (box 8.4). The problem with capacity and social capital Direct disbursement of funds requires a rev- arguments is that even if South African com- olution in the way countries, the World Bank, munities had such shortcomings, keeping com- and other donors manage upward, horizontal, munities from engaging in community-driven development precludes the development of these Box 8.4 Design principles for community-driven development capabilities through learning-by-doing­—­the ar- • Establish an enabling environment through relevant institution and policy gument becomes a self-fulfilling prophecy. The reform. financial argument stems from an incomplete • Make investment responsive to informed demand by providing knowledge reading, or biased interpretation, of the relevant about options and requiring communities to contribute to investment and re- financial acts: the Public Financial Management current costs. Act, 1999, Article 38 (1) establishes the princi- • Build participatory mechanisms for community control and stakeholder in- ples for the transfer of funds to communities; the volvement by providing community groups with knowledge, control, and au- thority over decisions and resources. Municipal Finance Management Act No. 56, • Ensure social and gender inclusion. 2003, Article 67, sets out the guidelines on how • Invest in capacity-building for community-based organizations. to transfer funds to communities. The relevant • Facilitate community access to information. legal text is in annex 2. • Develop simple rules and strong incentives, supported by monitoring and evaluation. Toward a model of coproduction of • Maintain flexibility in design of arrangements. services • Design for scaling up. Community participation and empowerment • Invest in an exit strategy that establishes project sustainability, including per- manent institutional and fiscally affordable financing arrangements. does not take place in a vacuum: it is affected by the local government development programs Source: Binswanger-Mkhize, de Regt, and Spector 2010. and the sector development programs of national governments. The emerging consensus is that

110 Accountability in Public Services in South Africa Box 8.5 Six innovations in financial accountability for community project management

Legal ownership of the funds. Funds transferred to communities Purchase of technical support by the community. The community were considered matching grants and thus became the property of can select any capable supplier and use a portion of the matching the communities rather than of the program executing agency. As grant, usually 8 percent, to pay for technical services. with credit from a bank, choice in spending the funds became the privilege and responsibility of the owner­—­the community. Transparency at the community level. Communities elect finance committees, which are in charge of day-to-day spending. Checks Replacing detailed accounting for funds with a community contract. must be signed by at least two members of the finance committee. The “expediente technico” for a small community project specifies The committee must present all accounts to the general assembly, how funds will be used and the technical details of the project. It which often also elects a committee to audit accounts, purchases, is a four- to six-page contract between the executing agency, the stocks, and their uses. community, and sometimes a facilitator or technical agent. Following implementation, the community signs a certificate of completion or Local shopping for both goods and services. Putting aside the tra- a handover of the project, which certifies that the project has been ditional distinction between goods and services in community pro- properly executed and the funds accounted for. The completion cer- curement rules, communities mainly shop locally to procure small tificate serves as a “receipt” for accounting purposes for the financing contracts and small amounts of supplies. (Competitive bidding is agency. Detailed accounts in the local language and receipts, which still required if the community enters into larger contracts.) Local remain with the community, can be inspected in random audits. shopping rules require that the community obtains offers from three suppliers and that its finance or management committee chooses Direct transfer of matching grants into community accounts. This is from these offers. usually done in tranches, the first following the signing of the con- tract and the second and third depending on demonstrating project Source: Binswanger-Mkhize, de Regt, and Spector 2010. progress benchmarks. successful services depend on integrating local Box 8.6 Linking sector programs, local government, and communities government development programs, sector pro- Three alternative approaches to local development­—­decentralized sector pro- grams, and community empowerment (box 8.6). grams, local government approaches, and community support approaches­—­ All of the approaches aim to provide public emphasize many of the same principles: empowerment of the poor and other facilities and services and to promote local de- marginalized groups, responsiveness to beneficiary demand, autonomy of local velopment. Each addresses the same challenge institutions, greater downward accountability, and enhancement of local capacities. from a different entry point: The three approaches organize their efforts differently: • Sector approaches, because they develop • The sector approach is organized by the services to be provided. • The local government approach is organized around a territorial jurisdiction through functional specialization, tend to be with some autonomy. better at mobilizing technical capacity but • The community support approach is organized around social groups that make less responsive to local demand and condi- collective decisions. tions and cross-sector considerations. Each approach has generated a distinct body of theory and practice. Many • Local government approaches, because of their countries use all of them, though this can lead to confusion and overlap. Local institutional territorial governance perspec- and community-driven development is a new framework that seeks to link all three tive, ensure clear formal autonomy and ac- organizational structures and sees them as coproducers of local development. It countability of local decisionmakers but are promises to improve coordination, synergy, efficiency, and responsiveness in local development processes. often more politicized and less effective in

managing service provision. Approaches to local development • Direct community-support approaches en- hance empowerment and responsiveness Local Community to local priorities and conditions, but their government support entry point through community struc- approaches approaches Linked tures and processes often complicates co- approaches ordination with public sector organiza-

tions needed to sustain service delivery and Decentralized infrastructure. sectoral approaches Linking the approaches can capitalize on the comparative advantages of each, while comple- Source: World Bank 2004b. mentary methods drawn from other approaches

Accountability in Public Services in South Africa 111 the legal, fiscal, can be used. Sector experience in organizing, important civil society sector exists in which and policy managing, and delivering local services is linked blacks and women play important roles. The civil environment to decentralized governance systems associated society sector weakened immediately after 1994, with local governments: multisector planning, however, and a “passive recipient” perspective set exists for resource mobilization and management, and in, which remains a barrier to recent participa- community-driven mechanisms of democratic accountability, for tive planning. development, but instance. Public sector approaches are comple- Local governments often act more as gate- mented by community support approaches pro- keepers and controllers than as facilitators for it is underused moting empowerment, participation and social community control of resources and alloca- capital for governance, collective action, and tion.4 Communities have yet to be significantly infrastructure and service coproduction. Insti- involved in planning, project identification, or tutional arrangements and capacities that build widespread participation in implementing local on the strengths of each approach can improve development programs and service delivery. governance, public services, and the welfare of Community-based development remains far households and communities. more widespread than community-driven devel- Scaling up a number of services follow- opment, in which communities have control over ing the coproduction model holds considerable project identification, project resources, and im- potential in South Africa. Further evaluation plementation. Community participation remains may be needed to understand how well cur- at the consultation level during planning, with rent sectoral, local, and community systems are subsequent involvement in running and main- aligned for coproduction and to identify con- taining projects. Few central government pro- straints and develop strategies for successful co- grams have moved toward community-driven production and scaling up. development; provincial and local governments and even NGOs are not yet using this approach. The status of and prospects for local Reasons for the difference between inten- and community-driven development tions and reality include: Case studies of South Africa since 1994 sug- • Local governments use ward committees to gest that the legal, fiscal, and policy environment involve communities, but their status is not exists for community-driven development but yet well defined and their functionality is far that it is underused (Everatt and Gwagwa 2005). from balanced. They often operate as exten- Promising community-driven development ini- sions of local government rather than as in- tiatives during the 1990s, such as by Mvula Trust dependent community structures. in water provision, have been abandoned. Where • Budgets for ward-level initiatives such as a South Africa employs participatory approaches, community fund do not exist. community consultation and facilitation models • The National Development Agency is too are used rather than the community empower- centralized and sets high barriers for NGOs ment model that characterizes community- and communities to receive any funding. driven development. At the time of their study, Few options exist for communities to receive Everatt and Gwagwa could find almost no proj- small amounts of money to start building ects where implementation responsibility and capacity through learning-by-doing. Funds funds were transferred to communities. But the from the Equitable Share grant could be few case studies of community-driven approaches used for this purpose. in South Africa show positive outcomes with no • Government agencies at all levels are re- instances of financial mismanagement, consistent luctant to “let go” and transfer control to with international findings of strong community communities. financial management (Binswanger-Mkhize, • Local development is linked to an elaborate de Regt, and Spector 2010). three-year development planning process­—­ The struggle against apartheid left a pow- the Integrated Development Plan­—­in which erful legacy of community-based organizing communities supposedly play an important and mobilizing. Everatt and Gwagwa show role. Neither local governments nor com- that in South Africa a large and economically munities have the capacity for such elaborate

112 Accountability in Public Services in South Africa planning and therefore rarely overcome hur- review of the Integrated Development Plan, too much emphasis dles to accessing funds. Planning and project implementing and reviewing performance is placed on identification is driven by consultants rather management systems and performance out- onerous planning than by communities. International experi- comes, and consultation on service strate- ence shows that decentralization alone can- gies. The municipality is obliged to build and upward not lead to systematic bottom-up planning. the capacity of the community, staff, and accountability Annual budgets are the primary planning councilors to aid participation (Municipal tool at local levels and incentives and guide- Systems Act Chapter 4). The Act explicitly lines are used to align these budgets with na- allows municipalities to transfer funds to or- tional priorities. ganizations and bodies outside government, • Efforts to ensure accountability are focused subject to the prescriptive conditions listed upward rather than horizontally or down- in annex 2. Section 67 of the Municipal Fi- ward. Both reflecting and reinforcing this nance Management Act also allows munici- upward emphasis, disbursement and audit- palities to transfer funds to organizations ing practices have not adapted to a decen- and bodies outside government. tralized, community-driven mode of opera- • In line with these provisions, the Khanya- tion (even though it is legally possible). As African Institute for Community-Driven a consequence, accounting officers are averse Development (Khanya-aicdd) has developed to risk. a comprehensive approach to community- • Surprisingly, not even NGOs are using based planning, which has been field tested community-driven approaches. They remain and evaluated in eight municipalities (both stuck in a service delivery model that keeps rural and metropolitan). Dozens of wards the communities they are supposed to serve have already demonstrated strong impacts on from being empowered. This may be a con- social capital, local plans, and even service sequence of onerous rules and planning re- delivery. The resulting comprehensive man- quirements that govern access to funding. ual, published and accredited by South Af- The paper concludes that too much emphasis rica’s State Information Technology Agency, is placed on onerous planning, upward account- is being distributed across several provinces ability, and capacity-building preceding action. with European Union support. The wards No substantive opportunities exist for learning- that participated in community-based plan- by-doing, and officials are not willing to cede ning were allocated a discretionary grant of critical identification and implementation func- ZAR50,000 that could be used for start-up tions or control over money to communities. The activities. They were encouraged to seek col- paper recommends broader use of community- laboration and resources from a wide range driven development, tempered by the reminder of institutions. that it is not a blueprint but rather an option to • The Department for Cooperative Gover- be assessed. nance and Traditional Affairs, with the sup- It is now clear that constitutional provisions, port of Khanya-acidd, has developed a pro- the decentralization framework, and the Batho posal for a community development grant Pele principles are conducive to scaling up local coordinated by municipalities and planned and community-driven development. The fol- and executed at the ward level (box 8.7). lowing four factors also support this position: It could be rolled out in stages across the • The Municipal Systems Act Section 16 (1) re- country. quires the municipality to develop “a culture • Stemming from the Second Economy of municipal governance that complements Strategy approved by the Cabinet in 2009, formal representative government with a the Presidency has initiated a Community system of participatory governance.” It states Works Program that guarantees two days of that “the objective of a ward committee is employment to those ready to work and that to enhance participatory democracy in local is highly decentralized, involving munici- government.” Participation is seen to include pal governments and intensive community budgets, preparation, implementation, and participation.

Accountability in Public Services in South Africa 113 Box 8.7 Proposed program of community development grants Internationally, the pendulum has swung Building on wide international experience with local and community-driven de- back toward decentralization. A consensus is now velopment, the Department of Cooperative Governance and Traditional Affairs emerging around the potential for local and com- (CoGTA)­—­with support from Khanya-aicdd­—­has proposed grants of ZAR200,000 munity-driven development to deliver a range of to all wards using community-based planning. Wards would also receive a function- services. Independent evaluations show this can ing grant of ZAR50,000, which has been sorely missing. The total cost of such a be done successfully and scaled cost-effectively. program would be ZAR1.3 billion annually­—­small by comparison with the construc- Decentralization and a number of accountability tion of stadiums and the Gautrain. The combined community-based planning and community development grant innovations are critically important. Participa- program would be integrated into existing decentralization structures to support tion and community-driven development need the effective functioning of ward committees, foster the development of social and not slow delivery­—­they can actually expedite it. administrative capital, and assist in the integration of departments. Funds would be For participation to be scaled up and incorporated managed by ward vote in municipal accounts. CoGTA envisions a phased imple- into massive programs without interrupting them, mentation over three or four years, the first of which would be devoted to diagnosis it is important to understand the complex train- and national, provincial, and local pre-program development. ing and facilitation logistics needed for scaling up Unlike the Community Works Program and community-based planning, this cost-effectively. Other innovations in financial -ac proposal still awaits wider review and approval by the Cabinet. Implementing it would start to fill the main local and community-driven development gaps. countability mechanisms are needed as well. Realizing the government’s vision of popular Source: Goldman 2010. democracy, participation, and local government “at the coalface of service delivery” requires a Many of the elements are in place to make fundamental transformation toward a coproduc- local and community-driven development a re- tion model of service delivery. This model must ality. But even where good experience exists, be based on empowering local governments and scaling it up is daunting. Scaling up requires communities, building capacity, realigning cen- that all communities, wards, and municipalities tral government toward facilitation, and im- be reached, creating perhaps hundreds of thou- proving accountability for progressive transfer sands of community projects. The programs dis- of decisionmaking and public resources to local cussed above will face this challenge. governments and communities. International experience suggests that while In this model, both local governments and challenges are formidable, successful scaling is communities need to be fully empowered to possible. Success depends on adopting a sys- play their part as systems move from commu- tematic approach starting with diagnosis of all nity consultation and facilitation to community coproduction systems, followed by piloting, empowerment models. Line departments must developing comprehensive user manuals, scal- move from direct service provision to policy, fa- ing across the country, implementing monitor- cilitation, and monitoring and evaluation. ing and adaptation processes, and consolidat- Fortunately, many of these elements are al- ing programs. Such an approach is described ready in place. Several case studies of commu- in Binswanger-Mkhize, de Regt, and Spector nity-driven development in South Africa have (2010), including a step-by-step guide and many shown positive outcomes and good financial design elements to facilitate scale up. Critical management. Further study is needed to un- elements of the scaling process are described in derstand how sectoral, local, and community annex 3. systems can be better aligned for coproduction, and strategies must be developed to identify and Conclusion overcome constraints to scaling. This chapter suggests that South African devel- opment experience with services for the poor Notes since 1994 resembles that of many other coun- 1. Its vision is encapsulated in Community- tries: despite a strong commitment to decentral- Driven Development in Africa (World Bank ization and citizen participation, many programs 2001). became increasingly centralized and sector- 2. The study was conducted by Hans driven, with decisionmaking moving away from Binswanger, Keith McLean, Graham Kerr, communities and local governments. Andrew Parker, Suzanne Piriou-Sall, Johan

114 Accountability in Public Services in South Africa van Zyl, and Melissa Williams. Outputs are • Which level of government pays the sal- summarized in McLean and others (1998); aries of staff in the smallest administra- Aiyar (1995); and Aiyar, McLean, and tive unit? Piriou-Sall (1996). • What proportion of sector expendi- 3. The study constructed a decentralization tures of the smallest administrative index ranging from 0 to 10 based on data unit is derived from the budgets of local collected from World Bank sector specialists governments? who had worked intensively in the respective • What proportion of sector expenditures countries/states/provinces. It was based on of the smallest administrative unit are answers to the following questions: derived from user charges, in-kind con- • Where is the smallest management unit tributions, and other beneficiary cost-re- for rural sector service delivery physically covery schemes? located? • Who determines the budget of the • Which level of government is responsible smallest sector management unit? for conditions of service of civil servants 4. See for instance Williams (2005, p. 19), who in the smallest management unit? concludes that evidence from Cape Town • How important are elected bodies in over 1994–2004 shows that “community (a) service delivery, (b) policy formula- participation has been largely rhetorical and tion, and (c) funding of each sector? not substantive.”

Accountability in Public Services in South Africa 115

9 Approaches and Tools for Effecting Change

Strengthening service accountability relation- model (Takeuchi, Osono, and Shimizu 2008), ships requires fundamental transformations that whose application has spread beyond automobiles must be approached with humility­—­no single, to other industries, private sector companies, and magic reform can make it simple. organizations such as hospitals and postal services. A comprehensive analysis of social services in Toyota’s recent failures notwithstanding, its Peru warns that many reforms are often reversed approach to quality control, lean manufacturing, after a short time because “corporations of pro- and accountability systems made the company viders” push the system back to an “equilibrium the world’s premier automobile manufacturer trap” of low expectations (Cotlear 2006, pp. 2–3). with an unsurpassed reputation for quality. Cotlear argues that when a system finds itself in The lean production model maximizes learn- this kind of low-level equilibrium trap, a set of ing by forcing errors. To understand this coun- multifaceted reforms are needed to “shock” the terintuitive idea, compare the more familiar Ford system. In Peru, he suggests establishing simple mass production line, which divides production and clear standards of qualitative and quantita- into a sequence of assembly activities. The over- tive goals, rigorous definition of accountability all goal is to keep the line going to maximize relationships, investment in capacity, and adop- delivery­—­that is, output (a finished car)­—­at the tion of measures that make citizen-users more end of the line. The line risks breakdown at every aware of their potential to influence providers. point from many possible errors­—­overall design Such a multifaceted approach seems promis- errors, problems with specific parts, how some ing for South Africa. A private sector transfor- parts fit with others, and so on. Ford’s solution mation model­—­quality control and the “lean to speed delivery was to inventory enough spare production” model now used widely in the au- parts at each assembly point to quickly replace tomobile industry­—­might also be useful to con- failed parts; the model purposefully created ex- sider for development. The model’s mantra is that cess capacity to ensure delivery at scale without a set of simple reforms can lead to systemwide interruptions. It was so successful that it became change if the system and incentives are geared virtually synonymous with efficient production toward continuing learning and outcomes. technology characterized by economies of scale. But the Ford model has two main prob- Quality control and lean production in lems: the cost of inventory along the produc- automobile manufacturing tion line and the fact that a manager standing at Toyota’s success is widely attributed to its highly the end of the production line can evaluate the innovative quality control and lean production quantity of cars produced but nothing else­—­so

Accountability in Public Services in South Africa 117 Toyota created an the system precludes “learning.” Say a design improve. The agreements specified rules for cost- entire production problem was causing battery problems, result- and profit-sharing and for how disagreements system—­ not­ just a ing in errors even though the battery itself was would be resolved. For instance, team decisions fine. Occasionally the power supply would trip, would be taken by consensus; only when the few mechanisms­ and to keep the line moving the battery would team could not reach consensus would higher —focused­ on quickly be replaced. No opportunity would arise levels of management be asked to arbitrate. continual for the design expert to work together with, say, The automobile sector has embraced many of the electrical specialist to create a better design. the lessons of designing and implementing ac- improvement Of course, good managers would notice how countability institutions discussed in this report. and outcomes frequently batteries were being replaced, but the Simple and continual communication strives for real cause of the problem could only surface after “everybody to know everything.” It also creates special evaluations were conducted off the as- many different social networks throughout the sembly line. The Ford model “delivered” cars, organization—­ horizontal,­ vertical, cross-func- but little instantaneous, automatic, and system- tional, geographic, and across hierarchies. atic learning from production errors took place The point is that Toyota created an entire Toyota’s new production model took the car production system­—­not just a few mechanisms­ industry by storm. American car makers ini- —­focused on continual improvement and out- tially resisted the idea, arguing that a unique comes. Many public sector organizations­—­ Japanese character trait explained the success including the World Bank and the government of the model. However, when Toyota started to of South Africa­—­are not organized in this way use it successfully in American car factories, the and thus are not efficient at learning, innovating, American industry relented­—­Ford now uses the and focusing on outcomes. This is a sizable de- Toyota model. fect; the last half century has taught that no easy, The model is deceptively simple: take away one-size-fits-all mass production solution exists­ all inventories, force the errors, and stop the pro- —­nor a “magic bullet” for change. The Toyota duction line. Once halted, a team of specialists­ story teaches that a systemwide reform agenda is —­empowered and incentivized to find and im- needed to replace the old Ford model. plement the solution and drawn from the various technical departments­—­is sent in to solve the World Bank and South African problem. The production line restarts and ev- experience eryone awaits the next error. Learning and im- Take the World Bank in the 1970s and 1980s: provements take place continually. projects emphasized the design phase, before The pioneers of the model believed that the “production line” even started to roll and frontline staff were not just “implementers”­—­or, when the least was known about what would in the language of this report, “service deliverers”­ work and what would not. Staff were organized —­but knowledge workers uniquely positioned to in technical silos and rewarded for taking proj- accumulate “chi”­—­the wisdom of experience. ect designs to the Board. Independent project Toyota designed the system to maximize knowl- evaluation was typically done three years after edge gathering; not much attention was paid to project completion, when lessons would be too the “organogram” and to lines of communication late to be implemented. In the 1990s, the World and hierarchy. Coordination and integration Bank introduced more flexible approaches with took place “on the ground and on the job” using more emphasis on monitoring and evaluation, multidisciplinary teams incentivized to succeed. but challenges and parts of the old model persist. In a sense, the team members became part of a The situation in the South African govern- coproduction agreement. ment is similar. So far little investment has been When developing a new model or a new made in improving the learning-by-doing of assembly technique, the new model is not delivery systems. Few impact evaluations have yet known, so optimal designs, and contracts been conducted, either after or as part of pro- with suppliers, are difficult. Instead, Toyota gram implementation (the preferable approach). reached open-ended coproduction agreements­ Implementation policies are designed by cen- —­not contracts­—­with suppliers to innovate and tral departmental units before program rollout.

118 Accountability in Public Services in South Africa These designs­—­typically not involving imple- Tracking Surveys, which may identify leakage Improving service menting managers, operational staff, or end- of funds, but which part of the leakage is due accountability users­—­are performed as desk exercises by staff to corruption and which part represents a legiti- requires focusing of centralized “systems and procedures” units. mate reallocation of the budget? Once programs start, continually updated op- The Presidency has supported better moni- on outcomes while erational manuals are rarely used to reflect the toring and evaluation systems for a number of tracking inputs valuable experience gained. Similarly, no joint years. The new Department of Performance and outputs manuals are produced describing the responsi- Monitoring and Evaluation is expanding to de- bilities of different coproducers and how those sign a central, national system that could then be responsibilities should be carried out. cascaded down to municipalities.1 As program rollout proceeds, the imple- Even a centralized system must be able to menting teams are almost never empowered collect basic administrative data from the de- to solve problems and update policies based on partments involved in service delivery, however, practical experience and participation by ben- and such basic administrative data is often hard eficiaries and other stakeholders. Changes can to come by. The departmental monitoring and usually be made only after formal review and evaluation units are usually separated from im- approval by senior managers, often at the high- plementation to ensure their independence, but est levels­—­so programs change slowly, if at all. this also separates the units from vital informa- Also, practical experience is often only “stored” tion flows. Some departments have attempted in the individuals who are implementing the to automate information flows, but few incen- programs, not in operational manuals; experi- tives exist for implementation officials to input enced staff take their implementation knowledge data into the system. A vicious cycle is created in with them when they leave, creating “capacity which data are not current, further undermin- problems” with new staff. ing the system. Without functional, automated Many departments know very little about databases, many departments and their monitor- which policies and programs are working and ing and evaluation units find it hard to report which are not, and why. Many departments keep basic data. The analytical work of monitoring adding new “products” and new organizations in and evaluation units is often based on incom- search of a “magic bullet,” instead of gradually plete data and not fed continually into updated improving and consolidating programs and proj- program manuals. ects by building on the lessons of failure. Integrated financial management systems are a possible solution to ensure that basic data are Monitoring inputs and outputs collected for further analysis and dissemination. Improving service accountability requires focus- These systems are designed to automatically link ing on outcomes while tracking inputs and out- inputs (mostly budgets) to outputs. (Outcomes puts, the vision South Africa’s new Department are often not immediately available.) In Latin of Performance Monitoring and Evaluation America, integrated financial management sys- has for development sectors. Examples of some tems, implemented as part of wider public ex- of the tools needed include client satisfaction penditure reform, have been among the most surveys­—­such as citizen report cards­—­and Pub- successful public sector management reforms. lic Expenditure Tracking Surveys. Amin, Das, Guatemala’s Integrated Financial Manage- and Goldstein (2008) provide a comprehensive ment System, for example, eliminated wasteful overview of the existing “toolbox” for measuring and unauthorized spending by establishing a de- service delivery. centralized, web-based procurement system. The Measuring services (coverage, quality, ef- system reduced disbursement and other costs ficiency, targeting, and impact, among others) and dramatically improved transparency by in- is not straightforward. Take client satisfaction, troducing electronic payments. The system also which may not be related to objective mea- improved information flow to legislators, private sures of quality, or which may be biased by cli- citizens, nongovernmental organizations, and ent expectations formed perhaps by listening to other stakeholders by providing Internet access a political speech. Or take Public Expenditure to national budgets and expenditures, including

Accountability in Public Services in South Africa 119 An integrated physical and financial performance indicators. health, and nutrition in Mexico was part of a financial The internal government compact and the short category of programs using “conditional cash management route to accountability were strengthened, and transfers.” Money was transferred to families, the innovations are now being implemented in conditional on family members having health system can be municipalities (Fiszbein 2005). checkups, mothers participating in hygiene and an important South Africa’s National Treasury has also nutrition information sessions, and children at- reform element, successfully introduced a public finance manage- tending school. While stretching the limited ment system. As in Latin America, this system is fiscal resources of the country, the program but it should being used as a springboard to wider public ex- was intended to assist the poor in investing in not be seen as a penditure management reforms, particularly by human capital, recognizing the interdependen- “magic bullet” linking fiscal resources more clearly with outputs cies among health, education, and nutrition. and outcomes. A word of caution is appropriate, A rigorous impact evaluation was conducted however: in many countries, reforms seem to to measure whether the program was attaining have stopped midway through the process. Bo- its objectives. The results were dramatic: by the livia, Brazil, Chile, Colombia, Mexico, and Peru end of 2002, 21 million people had benefited have all embarked on reforms to develop a re- from the program, or about a fifth of the Mexi- sults-oriented management framework, but only can population. Girls’ enrollment in middle Chile has been able to design a fully functioning school rose from 67 percent to around 75 per- performance management system. cent, and boys’ from 73 percent to 78 percent, An integrated financial management sys- while labor force participation decreased about tem can be an important reform element, but 20 percent for boys. Annual mean growth in it should not be seen as a “magic bullet.” It is height was 16 percent for children covered by more important perhaps­—­as the Toyota model the program, while average height increased by strongly suggests­—­to ensure that services focus 1–4 percent and weight by 3.5 percent. Visits by on outcomes, which can be linked confidently pregnant women in their first trimester rose 8 with inputs and outputs. percent, keeping babies and mothers healthier. Illnesses dropped 25 percent among newborns Impact evaluation and 20 percent among children under age five. Basic administrative data can demonstrate a The prevalence of anemia in children two to four great deal about inputs and outputs but usually years old declined 19 percent; adult health im- very little about impact. Impact evaluation asks proved too. a more fundamental, seemingly impossible ques- The widely publicized, clearly positive results tion: What would have happened to the users had another important impact: the program of a particular service had they not received the survived a major political transition. Observers service? say that political interference would in all likeli- Impact evaluation methodologies try to con- hood have stopped the program had it not been struct a “counterfactual” or a “control group,” evaluated. which measures the impact­—­say, on incomes or So, by showing how a service is perform- living standards­—­of those who received the ser- ing, impact evaluation provides an important, vice compared with a group of exactly the same objective input into both the policy debate be- kind of people who did not receive the service. tween citizens and politicians (the long route to The methodology will not be discussed in-depth accountability) and the implementation debate here, but it is important to understand both the between users and providers (the short route to major advantage and major drawback of impact accountability). evaluation. But impact evaluation also has a major The advantage of impact evaluation is often drawback. The impact of a service is defined illustrated by the PROGRESA story in Latin by the outcome of a range of variables, with America. Its success spawned a series of innova- the various accountability relationships being tions in impact evaluation, helping to establish one subset. To evaluate which variables, or in- it in the mainstream of development methodol- teraction of variables, cause the positive impact ogies. The PROGRESA program on education, would require more complicated methodologies.

120 Accountability in Public Services in South Africa Governments should not, therefore, count on Note impact evaluation alone to determine how to 1. Observers argue that the attempts to estab- start reforming service systems. This is where lish this system have not had much impact. approaches such as the Toyota model can be a “Unless a simplified system that is under- practical and powerful complement to scientific stood by all stakeholders is introduced, mon- impact evaluations: governments should try to itoring and evaluation will continue to gen- build evaluation and learning into the system erate large quantities of [poor quality] data by design, creating momentum for continual that only the specialists and consultants will improvements. read and use, with little democratic or effi- ciency benefits” (Pieterse and others 2008, p. 9).

Accountability in Public Services in South Africa 121

10

Conclusions

This chapter summarizes what has been learned The main finding of a number of assessments, from this study and makes policy recommen- and of this report, is that despite the Batho Pele dations based on the hypotheses, conjectures, principles and best intentions to choose partici- and inferences developed. As has been stressed patory approaches, implementation has sapped throughout this report, the international experi- power from citizens by emphasizing delivery ence with services teaches that no “magic bul- through supply-driven sector programs. This is let” and no one-size-fits-all solution exist for the primarily the consequence of neglect of the short delivery of critical services. Local context, long- route to accountability­—­client power over the and short-route accountability relationships, service provider. Though not expressed in these and the technology of the particular service all terms, the South African government, the Pub- interact in a way that precludes any simple pre- lic Service Commission, the African Peer Re- scriptions. It is important to create a system in view Mechanism, and many policy analysts have which citizens, politicians, and service provid- come to the same diagnosis. ers have the right incentives and hold each other Citizens feel this lack of power acutely, as accountable, while working together to continu- shown by the participatory assessment in chapter ally improve services by learning from experi- 5, but the same assessment also showed that this ence. This is much easier said than done­—­so lack of power was not perceived as a problem by how is such a system established? the local politicians and service providers. South Africa’s strengths and achievements Why this difference in perspective? It seems can serve as the foundation for such a system: that intense delivery pressures and a focus on • Democracy. quantitative output and input targets are partly­—­ • The Constitution, including the constitu- albeit inadvertently­—­responsible for undermin- tional mandates for basic services. ing the intentions of the Constitution, the Batho • The framework for decentralization and the Pele vision, the sector strategies and policies, and intergovernmental fiscal system. the promise of decentralization. Participation has • The Batho Pele principles. often been treated as a time- and resource-inten- • The sector-specific legal and policy sive activity that would slow implementation and frameworks. achievement of ambitious targets. • A vibrant civil society. However, international experience shows But it is widely agreed that implementation that participation can be scaled up and in- today falls short of the aspirations of the citizens corporated into massive delivery programs and of the goals and targets of the government. without slowing them. This requires a better

Accountability in Public Services in South Africa 123 a broader reform understanding of the complex training and facil- and outputs, resulting in dispersed, inac- agenda than itation logistics needed to scale up participation cessible, poor, and untimely basic data on that envisaged cost-effectively (see annex 2). Scaling up partici- services. pation also requires a number of other innova- • Underdeveloped feedback mechanisms have by the current tions in financial accountability mechanisms. no systematic approach to correct errors government must Recurring problems associated with the quickly. be implemented short route to accountability include: • Lack of follow-up and sanctioning­—­both • Participation is often diluted to mere com- exacerbated by poor data­—­when manager to improve munity consultations and treated as some- performance is unsatisfactory, including in services thing to check off on the official’s checklist the area of financial management. of “his” project. • Incomplete decentralization due to a lack • Little or no choice of providers by users, of consistent application of the subsidiarity combined with low or absent user fees and principle. copayments. • Overlapping mandates and responsibilities • Underdeveloped or absent user committees among central government, provinces, dis- that supervise service providers. (Even where tricts, municipalities, and service providers, they exist they have few powers over service and a lack of coproduction agreements that providers.) define the functions and responsibilities of • No systematic, institutionalized use of client each actor. feedback mechanisms such as citizen report • Excessively complex Integrated Develop- cards. ment Plan process, the purpose interpreted • No systematic, institutionalized reporting differently by politicians/policymakers than by service providers and local governments by citizen-users. to the citizens about coverage and quality of • Municipalities’ preference to deliver ser- services and use of resources. vices themselves rather than through pub- • Insufficient use of local languages in infor- lic-private or other partnerships, due to the mation provision and feedback mechanisms. onerous requirements for contracting service • Ineffective complaints, redress, and sanc- provision. tioning/reward mechanisms. These problems have led to breakdowns in The underdevelopment of the short route to both the short and long route to accountability accountability is further aggravated by problems and to a state unable to respond to citizens, in- in the long route to accountability: cluding an inability to provide the quantity and • Compacts focused on inputs and outputs, quality of services demanded by the population. not on outcomes. This has led to a series of service delivery pro- • Services organized as centrally driven sec- tests across South Africa and more recently has tor silos, which are difficult to coordinate as contributed to xenophobia. demonstrated internationally and in South As reflected in the lists above, a broader re- Africa. form agenda than that envisaged by the cur- • Complex procedures, some still stemming rent government must be implemented to im- from apartheid, developed in central offices prove services. The multifaceted agenda first rather than in the field through participa- requires a change in mindset that prioritizes tory methods and after field testing prior to accountability and is focused on the long route, adoption. coproduction, real empowerment, and contin- • Lack of practical and comprehensive opera- ual monitoring and learning. Specific reforms tional manuals for implementation. will differ by level of government, across sec- • Inadequate monitoring, evaluation, feed- tors, and within civil society. The Presidency, back, and learning processes treat monitor- the National Treasury, and the Department ing and evaluation as ancillary rather than of Provincial and Local Government are al- integral to the service delivery process. ready leading the improvement of services, but • Financial management and information sys- they will likely need to broaden their reform tems not integrated with the data on inputs agenda.

124 Accountability in Public Services in South Africa Broad recommendations • Establishing or strengthening the sector- The Presidency The Presidency must spearhead reforms by pro- specific monitoring and evaluation systems, must spearhead viding the vision and political will: citizen-user dialogue, and supervisory user reforms by • The Presidency could propose a clear vision committees to give these systems real power of accountability­—­for service providers and over service providers. providing the policymakers, but especially for citizen- Changes to be driven by local governments vision and users­—­based on the Constitution and the include: political will Batho Pele principles. This would require a • Changing the mindset of officials in local sustained campaign to change the collective governments. mindset toward accountability. • Strengthening the roles of councilors and • The Presidency must display the political ward committees. will to complete the decentralization pro- • Designing and implementing planning, cess. South Africa needs a development ap- budgeting, and services using approaches proach that empowers local governments, that make local government accountable to communities, and civil society while align- citizen-users. ing the sectors to achieve this vision of de- • Rolling out community-based planning, centralization, community empowerment, community development grants, and the and accountability to citizen-users. Community Works Program (broadly local The Treasury and the Department of Coop- and community-driven development activi- erative Governance and Traditional Affairs can ties and programs) once these programs are drive reforms to complete the decentralization devolved to local government. process by: Changes to be driven by nongovernmental • Rolling out the existing community-based organizations and civil society would emphasize: planning process across the wards in all • Changing the mindset from a community municipalities. consultation model to a model that effec- • Developing the community development tively empowers communities. grant proposals and scaling up the Commu- • Becoming active and practical facilitators in nity Works Program. coproduction of services. • Developing, reviewing, and improving sim- • Strengthening community capacity to hold ple guidelines to municipalities on how to service providers and community leaders devolve power and resources to communities. accountable. • Providing guidelines for better alignment • Strengthening community capacity to man- of the sectors toward decentralization and age projects and project funds. community empowerment. • Simplifying processes, rules, and procedures­ Sector-specific recommendations —including­ fiduciary systems—­ and­ imple- menting better accountability mechanisms Education to citizen-users. • Strengthen citizen voice by ensuring that • Developing and implementing accessible outcome information is available and dis- systems to combine input, output, and finan- seminated so that progress can be tracked cial data into a single system. and parents can know how schools are • Improving data collection and developing performing. simple, timely, and publicly accessible moni- • Strengthen the short route to accountabil- toring and evaluation systems. ity in most South African schools, especially Sector administration reforms include: among those serving the poorest communi- • Changing the mindset of sector administra- ties, ensuring that parents are empowered to tion staff and reviving the spirit and practice hold schools accountable. of participatory policies and implementation. • Explore options for increased parental • Realigning sector administration for decen- decisionmaking and decentralize power tralization, community empowerment, and to schools. Provide schools with increased accountability to citizen-users. autonomy in decisionmaking in financial

Accountability in Public Services in South Africa 125 Accountability operations, procurement of goods and ser- these relationships­—­while the relationship be- must be rooted vices, and hiring and firing educators. Cou- tween citizens and municipal service providers locally, on the pled with increased information on service is given less attention. Primary responsibility for delivery, the combined reforms can create an holding service providers accountable should not ground, and at environment that empowers frontline pro- lay with a remote regulator. Such an approach the interface viders while making them accountable to places even greater emphasis on the relation- between citizens their clients. ship between government and providers, further • Strengthen the compact between govern- marginalizing citizens. and service ment and frontline providers by reviewing Citizens need direct, short-route account- providers the operation of the paper budget in non- ability mechanisms. Accountability must be and elected Article 21 schools and reconsider the restric- rooted locally, on the ground, and at the inter- tions for funds flowing to Article 21 schools. face between citizens and service providers and representatives Both measures could reduce delays in pro- elected representatives. Citizens should have re- curement and uncertainties about remaining course through the regulator, but primary en- budgets, which keep schools from planning gagement and response must be local, where effectively. More generally, understanding problems are felt and where immediate interven- of the actual financial operations of schools tion is usually required. to further hone the reform process could be Three measures would make a substantial improved. impact: • Reconsider teacher pay structures that at- • Dedicate significant resources in every mu- tract those with poorer abilities or endow- nicipality to build public understanding of ments to the profession­—­where they will how water and sanitation services work and earn more than in the private sector­—­and help citizens understand their rights, roles, push those with more favorable characteris- and shared responsibilities. Public education tics into the private sector or other parts of would begin to build mutual accountability the public sector. relationships for service provision. • Strengthen public accountability in the se- Water lection of service providers, preferably Enormous water services achievements in the through revising Section 78 of the Munici- first decade of democracy in South Africa may be pal Systems Act to permit more integrated compromised in the second; measures must soon and inclusive assessment of provider options. be implemented to mitigate the consequences This could include greater use of service of the current supply-driven approach to service partnerships with local small and medium- provision. South Africa’s state-centric approach size enterprises or community-based organi- is top-down, target-driven, supply-led, and zations, particularly in rural areas where mu- grant-funded. This approach skews infrastruc- nicipal service providers struggle to provide ture investment decisions, overestimates the skills support and management oversight. available to operate and maintain services, risks • Strengthen regulatory tools and capacity high service failures, and limits citizen power. to incentivize improvements in poor per- Public accountability in the water services forming municipalities. This may include sector is focused primarily on the relation- a greater emphasis on conditional require- ships among spheres of government­—­this study ments for grant funding and on sound asset makes recommendations for improvement in management.

126 Accountability in Public Services in South Africa Annex 1

Policy Matrix

By leg of accountability

Main areas Key issues Consequences Policy recommendations Long route to accountability, leg 1: citizen • Some municipal councils do not have ward • The formal democratic system does not yet • Clarify role of ward committees. voice (political accountability) committees adequately serve citizens. • Make establishment and elections of ward • Ward committees’ role is still unclear • Citizen voice is weak. committees mandatory. • Lack of capacity of elected members. • More voice mechanisms than elections are needed. • Provide ward committees with well-defined rules • Insufficient communication between the councilor, • Ward committees are failing in participatory local and a small budget for their functioning. the ward committee, and community. governance. • Ensure participation in choice of services and how • Councilors often do not reside in their communities • Civil society is often excluded from governance they are provided. and are not available to constituents. processes. • Strengthen links with grassroots organizations. • Ward committees find it difficult to communicate • Communities find alternative ways to express • Ensure simple and accessible mechanisms for with their councilor because they have no budget to dissatisfaction (protests). communication and dialogue. make phone calls or travel.

Long route to accountability, leg 2: the compact • Compacts. Compacts focus too much on inputs and • Participation is undermined and it is difficult to • Focus on outcomes and include citizen-user (state capacity and accountability) outputs, instead of outcomes. link the strategy and its outputs and outcomes to participation in the evaluation of outcomes and the budget. processes.

• Procedures. Administrative procedures are often not • Nontransparent administration inaccessible to the • Rigorously simplify central and local administrative understandable and practicable to the average staff majority of the population. procedures and communicate them in simple ways, member and the average user. Some complexities including in local languages. from the apartheid bureaucracy survive.

• Financial management systems. Current information • Sectors have a hard time linking finances, outputs, • Develop systems to combine outcomes, outputs, management systems do not combine outcomes, and outcomes. and financial data into a single system. outputs, and financial data into a single system. • While such reforms achieve better input/output • Technocratic public sector management reforms management, more emphasis is placed on the pace in a context of weak voice and accountability are of delivery and expenditure than on value for money usually largely ineffective. and the quality of outcomes.

• Silos. Services reach the local sphere in “silos” • Very difficult to coordinate programs. Service • Decentralize and unify budgets following subsidiarity implemented and funded by single sector programs. providers rewarded for the delivery of outputs principle. and inputs in “their” sector, instead of for good • Define roles and responsibilities of coproducers. outcomes based on coproduction. • Reward sectors for successful coproduction of outcomes.

• Contracting works best for services that are easy • Legal framework seems biased toward in-house • Review experience with contracting to define to monitor, are nondiscretionary, and can be provision. improvements needed. contracted-in for relatively short time periods. • Provide support to municipalities to improve decision to contract and design and supervision of contracts.

Decentralization • Some local government service assignments are • Overlapping mandates and responsibilities between • Provide the political will to complete the underdeveloped. center, provinces, districts and municipalities, and decentralization, to move to LCDD, and to align • Services that could be executed directly by local their service providers. the sectors with decentralization and community government are not (housing, education, and health, • Lack of precision in the allocation functions to empowerment. for instance). different levels. • Design and implement programs for local • Incomplete application of subsidiarity principle in • Underdevelopment of local and community-driven governments to hold themselves accountable to devolution of responsibilities to municipalities and development (LCDD) approach. citizen-users. communities. • Poor understanding of coproduction.

Accountability in Public Services in South Africa 127 Main areas Key issues Consequences Policy recommendations Short route to accountability, leg 3: • Communities do not have the mechanisms to hold • Lack of accountability to users reduces the incentive • Develop accountability mechanisms to users in all client/user/citizen power service providers accountable. to improve service quality. Users do not know their services. • Information without client power rarely improves rights and quality and delivery standards and are • Introduce the use of an “annual citizens’ report.” accountability. alienated by a lack of involvement. • Strengthen complaints mechanisms. • Information does not flow or is not easily accessible. • They express their frustrations by alternative means • Implement good practice for redress. such as service delivery protests. • Develop targets or standards for redress. • Benchmark redress performance and measure it. • Ensure accountability for complaints handling systems.

• Absenteeism. Citizen supervisory committees exist • Service coverage and quality deteriorates and • Establish or strengthen the sector-specific only in schools. While parent-teacher associations citizens feel powerless and alienated. citizen‑user supervisory committees and give them have voice, they often lack sanctioning power. Thus, real powers. absenteeism persists in schools. • Develop mechanisms that report on service provider absenteeism and link such reports to incentives. • Monitor absenteeism at more aggregate levels.

• Provider choice and user fees or copayments. • Citizens cannot vote with their feet. • Review options for more provider choice in all Provider choice and user fees can improve • With free provision, users will demand the services. accountability in services, if properly implemented. highest level of service irrespective of costs and • Review the efficacy of current arrangements for sustainability. funding free basic services.

• Participatory budgeting and development planning. • The IDP poorly reflects the public’s priorities. • Follow up the existing initiatives to simplify planning The Integrated Development Plan (IDP) process is • Citizens and civic organizations miss out on an processes. very complex. opportunity to influence the behavior of service • Scale up the already tested approach, community- • Knowledge about IDPs among the public is very low. providers. based planning. • Difference in interpretation between government • Complexities of planning may lead to consultant- and citizen-users about what an IDP is. driven processes. • Civil society organizations are absent from the IDP process. • Participation is consultative only.

• Community-driven development (CDD) has largely • Local development initiatives are lacking the cost- • Fully develop and scale up decentralized CDD disappeared from South Africa. effective and empowering CDD tool. programs such as community-based planning, • The constitutional, legislative, and decentralization community development grants, and the Community framework is conducive to CDD. Works Program. • A number of approaches and program proposals • Experiment with LCDD approaches for other using CDD are ready for scale up. services for which it is appropriate.

By sphere of government

Sphere of government Actions to drive policy recommendations The Presidency • Maintain and further strengthen leadership in accountability. • Articulate a vision for enhanced accountability to citizen-users. • Initiate and lead a sustained campaign to change the collective South African mindset to include accountability to users. • Ensure that the Department of Performance Monitoring and Evaluation translates this vision into its monitoring and evaluation activities. • Provide the political will to complete the decentralization process, move to LCDD, and align the sectors with decentralization and community empowerment. • Champion community-based planning, community development grants, the Community Works Program, and similar initiatives.

National Treasury • Simplify central rules and procedures, including fiduciary systems. • Develop systems to combine input, output, and financial data into a single system. • Review the efficacy of current instruments for funding free basic services. • Strengthen public accountability in the selection of service providers by reviewing Section 78 of the Municipal Systems Act to promote a more integrated and inclusive approach.

Department of Cooperative Governance • Complete the decentralization process. and Traditional Affairs • Provide the guidelines for the alignment of the sectors to decentralization. • Make the establishment of ward committees and their election mandatory, following well-defined rules, and with a clear role and a small budget for their functioning. • Make community development managers more accountable to ward committees. • Clearly define ward committees to involve communities and allow them to operate as independent community structures, rather than as extensions of local government. • Provide budgets for ward level initiatives, such as a community fund.

Individual sectors and line ministries • Change the mindset of the staff in their sectors. • Revive the participatory strategies and policies. • Manage their own realignment with decentralization, community empowerment, and accountability to citizen-users. • Establish or strengthen the sector-specific citizen-user supervisory committees and give them real powers over service providers. • Experiment with mechanisms that report on service provider absenteeism and link such reports to incentive mechanisms. • Introduce the use of an “annual citizens’ report,” which in simple (and possibly local) language explains what a department does, how it has met its service standards, and how it intends to improve its services. • Strengthen regulatory tools for oversight and compliance enforcement. • Implement the model of good practice for redress based on the steps outlined in the Batho Pele Handbook. • Develop targets or standards and tools for redress mechanisms.

Local governments • Change the mindset of local governments. • Strengthen roles of councilors. • Become active players in coproduction. • Strengthen community capacity to hold service provider and community leaders accountable. • Implement LCDD programs.

128 Accountability in Public Services in South Africa By sector

Education

“Technology:” Proposed changes in allocation monitorability, technology, and Current assignment of main and Currently used accountability of primary and coproduction Proposed changes in skills intensity Key issues coproduction responsibilities instruments and mechanisms responsibility instruments and mechanisms Education decisions take place • The efficiency of public • National level: sets policy • Participation in regional • Increase school-level • Ensure frequent production and numerous times at numerous expenditures in producing framework and budget comparable learning decisionmaking capacity and dissemination about school locations­—­meaning that they are learning outcomes is low­—­ • Department level: holds primary assessment; national “systemic authority. quality. hard to monitor. especially for the poor. responsibility for delivery of evaluation” assessment. • Ensure a teacher pay structure • Provinces control much of the education services, transfers • School Governing Bodies de jure that encourages highly skilled funding and procurement, which money to Article 21 schools, and oversee activities at school level. people to become teachers. may result in delays in funds and procures goods and services on But they do not appear to play • Devolve some power to schools in access to education materials behalf of non-Article 21 schools an effective role in most schools. in the hiring and firing of as well as circumscribed use of (typically those serving the teachers. funds at the school level. poorest communities). • Ensure that parents are aware • The pay structure for teachers • District level: “middle- of their rights, have adequate does not encourage hard work. management.” Support and information, and are empowered • The accountability system for monitor education practices in to hold schools to account. teacher oversight is weak. schools. • Increased decisionmaking at the • Many schools lack the autonomy • School level: responsible for school level. (and may need more capacity) delivering education but has to make appropriate local no authority to hire and fire decisions. educators.

Water supply and sanitation

“Technology:” Proposed changes in allocation monitorability, technology, and Current assignment of main and Currently used accountability of primary and coproduction Proposed changes in skills intensity. Key issues coproduction responsibilities instruments and mechanisms responsibility instruments and mechanisms Urban • Supply-driven infrastructure • Municipality has full • Most municipal service provision • Where current provider • Dedicate significant resources to • Reticulated water and sanitation delivery target, funded by capital responsibility for provision. A is in-house. performance is poor, consolidate building public understanding of networks are highly capital- grants. minority of services have been • A small minority of service and strengthen available how services work and helping and skills-intensive. They • Severe shortage of technical outsourced to specialist water providers are regulated by capacity into service utilities that citizens understand their rights, require extensive preventive skills to operate and manage entities in the public or private formal contracts. can achieve scale economies for roles, and shared responsibilities. maintenance and prompt services, poor asset sector. • Very few short-leg mechanisms service provision. • Strengthen regulatory tools reactive maintenance. management, and a growing • The Department of Water for citizens. Residents mostly and capacity to incentivize • Comprehensive monitoring is cost recovery gap threatens Affairs’ regulatory functions and take their complaints to ward improvements in poorly difficult without specialized sustainability. powers are still evolving. committees or councilors. performing municipalities. equipment. • Networked monopolies require • Councilor refers the complaints • Strengthen public accountability • Large-scale water and close oversight to protect the to relevant service provider. in the selection of service wastewater treatment is highly rights of citizen-users. • Makes citizens reliant on providers, preferably through specialized. responsive councilors. revising Section 78 of the Municipal Systems Act. • Review the efficacy of current measures for funding free basic services. • Review the conditions associated with accessing Municipal Infrastructure Grant funding, to provide better incentives to providers. • Build capacity to strengthen contract management and oversight

Rural • Service deficiencies are • As above, plus: • As above, plus: • Promote greater use of • Identify incentives for service • Localized water networks are most severe in areas where • Local water supply operators • Local civic participation service partnerships between partnerships between mostly not highly complex municipalities are youngest and are generally employed by and is usually greater in rural municipalities and local municipalities and local and are well suited to local least equipped. accountable to the municipality; areas, offering a valuable small, medium, and micro small, medium, and micro management if there is recourse • Poor performance in many local management is often poor. basis for strengthening local enterprises or community-based enterprises or community-based to technical support on demand. schemes. • Very few community-based accountability mechanisms. organizations. organizations. • Persistent service backlogs, organizations are still involved • Promote greater use of locally • Establish local user platforms especially for sanitation. in service management or based service providers who for citizens, providers, and • Supply-driven approach has oversight. can be held accountable by local municipalities to promote alienated and marginalized local civic structures. dialogue and joint problem residents. • Provide technical and other solving. • Failure of full ventilated support to rural settlements • Revive demand-driven improved pit toilets raises to reduce the frequency and approaches to basic sanitation questions around how best to duration of water outages. For improvement, with instruments respond to a complex second example, appoint a technical to offer citizens more choice and generation backlog. support agency (public or private “ownership.” sector) that can offer prompt • Halt grant funding for immovable call-down support. ventilated improved pit toilets.

Accountability in Public Services in South Africa 129

Annex 2 South Africa’s Legal Framework for Community-Driven Development Grant Funding Mechanisms

Contrary to perceptions, the legal framework government finances on sustainable footing to en- for local governments to establish community- sure that it can meet its constitutional obligations. driven development grant funding mechanisms The MFMA also explicitly allows for mu- already exists in South Africa. The Municipal nicipalities to transfer funds to organizations Systems Act (MSA) requires a municipality to and bodies outside government, subject to cer- develop a culture of community participation tain prescriptive conditions. First, no political to ensure that the community provides input in structure or office bearer of a municipality may municipal affairs. However, the municipality set up a relief fund, charitable fund, trust fund, must ensure that community involvement does or other fund of whatever description, except in not result in the community rendering services the name of the municipality. Only the munici- on behalf of the municipality. pal manager may be the accounting officer of any Section 16 of the MSA stipulates as follows: such fund. Second, grant funding can be allo- “A municipality must develop a culture of cated only if it is accommodated in the annual municipal governance that complements for- budget. And third, Section 67 of the MFMA mal representative government with a system of allows for municipalities to transfer funds to or- participatory governance, and must for this pur- ganizations and bodies outside government. The pose . . . encourage, and create conditions for, relevant section reads as follows: the local community to participate in the affairs 1. Before transferring funds of the municipal- of the municipality . . .” ity to an organization or body outside any The areas that focus on community partici- sphere of government otherwise than in pation are: compliance with a commercial or other busi- • Preparation, implementation, and review of ness transaction, the accounting officer must the Integrated Development Plan. be satisfied that the organization or body­—­ • Establishment, implementation, and review a. has the capacity and has agreed­—­ of the performance management system. i. to comply with any agreement with • Monitoring and review of performance. the municipality; • Preparation of the annual budget. ii. for the period of the agreement to • Strategic decisions for the provision of mu- comply with all reporting, financial nicipal service. management and auditing require- The main aim of the Municipal Finance ments as maybe stipulated in the Management Act (MFMA) is to place local agreement;

Accountability in Public Services in South Africa 131 iii. to report at least monthly to the ac- the benefit of the transferred funds; counting officer on actual expendi- and ture against such transfer; and ii. certifies to the Auditor-General that iv. to submit its audited financial state- compliance by that organization or ments for its financial year to the ac- body with subsection (1)(a) is uneco- counting officer promptly; nomical or unreasonable. b. implements effective, efficient and trans- In applying the above section of the MFMA, parent financial management and in- the municipality: ternal control systems to guard against • May transfer funds to nongovernmental fraud, theft and financial mismanage- organizations and community-based orga- ment; and nizations to further social and community c. has in respect of previous similar trans- development. fers complied with all the requirements • Must ensure that the funds are not used to of this section. undertake municipal functions. 2. If there has been a failure by an organization • Should ensure that the funds are not used for or body to comply with the requirements of political party affiliates. subsection (1) in respect of a previous trans- In considering Section 67 of the MFMA, fer, the municipality may despite subsection the municipality may be exposed to the follow- (1)(c) make a further transfer to that organi- ing risks: zation or body provided that­—­ • The mandate of the organization does not a. subsection (1)(a) and (b) is complied align with the objectives of the municipality. with; and • Allegations of bias with regard to the b. the relevant provincial treasury has ap- allocations. proved the transfer. • Funds are misappropriated and not used for 3. The accounting officer must through -con their intended purpose. tractual and other appropriate mechanisms • The organization is not legally registered and enforce compliance with subsection (1). does not have the equivalent of an account- 4. Subsection (1)(a) does not apply to an orga- ing officer. nization or body serving the poor or used by • The organization lacks the capacity to spend government as an agency to serve the poor, the funds. provided­—­ • Lack of monitoring capacity and systems in a. that the transfer does not exceed a pre- the municipality. scribed limit; and • The municipality is unaware that there may b. that the accounting officer­—­ be a duplication of funding where the orga- i. takes all reasonable steps to ensure nization has successfully secured funds from that the targeted beneficiaries receive another funder.

132 Accountability in Public Services in South Africa Annex 3 Scaling Up Local and Community‑Driven Development

Analyzing prior successes will identify innova- out as far as possible by the existing structures tions and processes that can be applied to local at the municipal and community levels and in and community-driven development. A diag- the sectors. But these structures alone cannot nostic tool for a rapid and participatory diag- invent and document new assignments, take nostic phase was developed by Heemskerk and up difficulties with higher levels for resolution, Baltissen (2005). Once the diagnosis is com- document the process, and compile it in an op- plete, policymakers from the national, provin- erational manual. These are the tasks of a project cial, and local spheres, as well as from the sectors design team­—­a facilitator, not the driver of the involved, must work with stakeholders and com- process. munity members to review these findings and A strong information, education, and com- define what is to be scaled up, how the identified munication program must be developed, im- bottlenecks can be overcome, and how labor will plemented, and documented. It should aim to be divided among coproducers. The outcome of empower communities to reflect; identify their this consensus building must be widely dissemi- needs, challenges, and resources; extend the nated to foster the necessary support. voices of the poor for participation in public dia- Even where there is good pilot experience, logue; facilitate education and learning about local and community-driven development has sectoral and multisectoral topics for behavior rarely been scaled up to an entire municipality change toward sustainable development and or district. A critical step is to select one mu- empowerment; and facilitate community access nicipality or district for scaling up in at least one to market information, access, and values. The jurisdiction. During this process, all the con- Malawi Social Action Fund, for instance, com- straints will become obvious­—­such as difficul- municates messages to all stakeholders in the ties in transferring funds to municipalities and community-project cycle through radio plays communities­—­and have to be resolved. The ex- and television dramas. Other broadcasted mes- perience can then be summarized in a compre- sages explain the importance of principles such hensive operational manual­—­developed in the as accountability and transparency and offer in- field rather than in an office­—­to transfer knowl- struction on technical issues such as procure- edge, processes, and tools to other jurisdictions ment and contracting. To meet process monitor- in a province or country. ing needs, an effective communication plan must Scaling up in a specific jurisdiction should focus on instituting multiway communication, be entrusted to a local development committee monitoring, evaluation, and feedback chan- associated with the local council and be carried nels among coproducing agencies. Moreover,

Accountability in Public Services in South Africa 133 the information generated and disseminated by and projects of nongovernmental organizations these systems is central to enhancing all forms of (NGO) or local sector offices; measures for local accountability. resource mobilization (from local revenues, cost A critical step is clearly assigning and fully recoveries, and other sources); monitoring and describing the specific program functions to evaluation of the local development program participants at each level. These assignments are and the performance of implementers; and re- best worked out at the local level by the stake- porting to the local and central authorities and holders who will have to carry them out, since the population at large. Participatory monitoring only they have the detailed knowledge of the and evaluation by communities can significantly actors, systems, processes, and relationships on reduce the progress monitoring functions of the which the program will need to build. The pro- district development committee. cess usually involves one or several broad stake- The project design team assists the local de- holder workshops. Depending on context and velopment committee in program development existing governance systems, the assignments by: can be chosen from a range of institutional op- • Assembling operational procedures, manu- tions. Despite variations, however, certain basic als, and tools from existing programs and structures, as shown in figure A1, will be neces- approaches. sary to meet the management and coordination • Integrating materials into a single set of op- needs of the emerging program. erational and training manuals and tools. The role of the district development com- • Developing components for training mittee is to coordinate elements of the local and materials. community-driven development program, in- • Revising materials as experience grows. cluding: the initial and subsequent information • Assisting the district in the development of campaign and Independent Electoral Commis- its IEC strategy and tools. sion (IEC) program; supervision of the train- • Helping the secretary and staff of the devel- ing, facilitation, and community and local plan- opment committees prepare the sessions and ning process; integration of development plans reports of the development committees. of subdistricts; approval of subdistrict and dis- • Reporting on the progress of the develop- trict level projects; recommendation for approval ment program at regular intervals. at higher levels for large district projects; final • Assisting in the design and implementation no-objection or approval of community projects of the monitoring and evaluation program, processes, tools, and reports. Figure A1 Basic institutional requirements Participatory development planning and ex- ecution requires enormous amounts of training and facilitation of communities and coproduc- Community development ers. Skills include participatory assessment and committee planning, setting up or strengthening of the • Coordinate, supervise, and • Recommend local-level projects administer subdistrict activities and community subprojects community development committee and its sub- • IEC, training, facilitation, for approval monitoring and evaluation, • Report committees, procurement and financial manage- and so on Subdistrict ment, planning and community project prepara- development committee tion, auditing, and participatory monitoring and

• Coordinate and supervise overall • Submit local-level projects and evaluation. In addition, technical training will district program community subprojects for have to be provided for all but the simplest in- • IEC, training, facilitation, approval monitoring and evaluation, and so on • Report and feedback frastructure community projects. Each commu- • Approve local level projects and community subprojects (or nity training program will include between four recommend large subprojects for approval) and eight people to ensure that members of dif-

Elected local ferent genders and age groups, minority groups, Impact evaluation government or Local program team local development support consultants and community technical specialists receive committee training. If a district has 300 communities, for

Source: Binswanger and Nguyen 2005. example, this would mean that 1,200 to 2,400 people would be trained. Training also has to

134 Accountability in Public Services in South Africa be provided to all other program participants numbers require that facilitators too be recruited engaged in coproduction tasks; that is, sub- from local populations and trained. Members of district and district development committees, subdistrict training committees can sometimes staff and volunteers from participating NGOs become facilitators. They are reactivated when- and sector agencies, and district and subdistrict ever training or facilitation needs arise. administrators. Specialized training and facilitation for spe- The large training volume requires careful cific subsector community projects should be attention to the logistics of the process and the primarily demand-driven, based on community cost of the people involved in training. Hiring development plans. Communities should be able professional trainers and facilitators for these to spend part of the money they receive to fi- activities is likely to be prohibitively expensive, nance or cofinance these services. The precise irrespective of whether the trainers and facili- source and logistics of providing these services tators come from the private sector, NGOs, or need to be worked out locally by the respective government agencies. Instead, latent local ca- sectors in coordination with the district devel- pacities could be mobilized by recruiting quali- opment council. Some district-level projects can fied locals, such as retired people, educated finance some of the basic costs of NGOs, private spouses, educated youth, and village elders, firms, or local sector offices of the government. who may have been teachers, health practitio- For fiduciary and disbursement purposes, ners, agricultural extension workers, and the a “community project” can either be an annual like. Again, South African experience is quite community budget comprising several commu- well developed here. An alternative or comple- nity projects or individual community projects. mentary approach is to organize specialized The first option is far more empowering, espe- training and facilitation through community- cially if combined with disbursement against to-community extension, which relies on and the budget in two tranches­—­or a maximum of strengthens latent community capacities and three. Funding is typically accompanied by a set thus saves money. In either case, a few profes- of rules and corresponding training that: sional trainers and facilitators are necessary to • Ensure broad local participation. facilitate the training and management of all • Promote transparency and accountability. the activities. Cost-effective training therefore • Prevent fraud and misuse. can be organized in training camps at the dis- • Avoid elite capture and social exclusion. trict and subdistrict levels. In an HIV/AIDS • Ensure that the community can manage prevention program in Burkina Faso, the villag- and maintain the asset after project com- ers (one woman, one man, one male youth, and pletion through local resource generation one female youth) are assembled at the subdis- mechanisms. trict level for two week-long training workshops Disbursements can be in tranches based on that cover the elements of the project cycle, fi- statements of expenditures. Verifying the proper nancial management, and basic knowledge of use of the money is the primary responsibility HIV/AIDS and its prevention. A locally re- of community finance and audit committees cruited training team operating under the su- and the general assembly of community mem- pervision of a district training team supplies bers. Auditing hundreds or thousands of small training to surrounding communities from the community accounts is not cost-effective, so it subdistrict or local market town. No program should be part of the monitoring and evaluation participant travels on a route other than to the plan and carried out on a random sample basis. market or local subdistrict center, sharply re- Audits should include the financial records, ducing transport logistics and costs. the decisionmaking processes, and the quality Participatory planning (and monitoring) of output produced. Whether in cash, labor, or processes at all levels must be facilitated prop- materials, such community contributions may erly by outside trainers and facilitators. Unlike constitute between 10 percent and 40 percent of training, the facilitation must be done separately total project cost. in each community, as well as for subdistrict Options for allocating money to communi- and district development committees. The large ties include a district/municipal development

Accountability in Public Services in South Africa 135 fund with a special window for communities or The planning process involves preparing detailed separate district/municipal and community de- budgets and a financing plan involving all pos- velopment funds. The advantage of having sepa- sible sources of funds: the national government, rate funds is that it better ensures empowerment local governments, communities, and external and learning-by-doing at these lower levels than donors. Financing norms and the pace of scal- does a process that maintains all the allocation ing up may have to be adjusted to fit the avail- power at the district level. The disadvantage is able financing to which the different partners that sequencing and economy of scale consider- are willing to commit. Budgeting and financ- ations are more difficult to accommodate than ing options include collapsing similar ongoing in a single, fully fungible budget at the district programs into the new program; allocating ad- level. Allocation to the communities from the ditional tax bases to local governments and com- community window or fund could be based on munities, which can be used to cofinance the proposals from these levels or on ex ante allo- program; allocating existing national revenues to cation to each community and subdistrict based the program through the intergovernmental fis- on a formula. Other rules can also be devised cal system; and raising additional taxes for fund- to ensure an equitable allocation of resources. ing shortfalls. Mexico’s Decentralizaiton and Regional Devel- It will also be important that all participat- opment projects, for example, adopted the prin- ing stakeholders identify gaps in the national ciple that at least half of the resources had to be and local institutional frameworks and remain- used for rural communities within a munici- ing bureaucratic hurdles and define time-bound pality, rather than the central municipal town, plans to overcome them. Issues to be discussed and that it should preferably target the poorest may include further aligning sectors with the communities. national decentralization framework, empow- A sound monitoring and evaluation system ering the community, strengthening the central has four parts: participation by communities and institutions in charge of decentralization and users for implementation monitoring; financial local governments, further simplifying proce- and accounting systems for financial monitor- dures seen as causing delays or bottlenecks, and ing; a management information system for prog- addressing weaknesses in national statistical sys- ress monitoring; and impact evaluation. At the tems and monitoring. local level, monitoring and evaluation is the re- Based on the experience in the planning sponsibility of the District Local Development phase, the national system for managing the Committee, the Sub-District Development program has to be designed in detail, particu- Committee, and the Community Development larly for the central program office and its sub- Committees. Implementing these improvements sidiary branches and their capacities, terms of has to start in the small municipality first chosen reference, and staffing. Planners will also need for scaling up and be integrated into the result- to consider phasing. Two or three phases may ing operational manual. be needed to reach national coverage. As in in- Once the program is running properly in dividual districts, scaling up across districts or one or a few districts, the operational manual, provinces will also require a detailed logistical training materials, costs elements, monitoring plan that attends to the costs of managing the and evaluation reports, initial impact evaluation program. Cost-minimizing approaches involve results, and other relevant instruments need to mobilizing existing structures and latent capaci- be properly reviewed, integrated, and presented. ties to manage the program rather than develop- These tools are subject to final revisions at a -na ing new ones or hiring special staff and consul- tional stakeholder workshop, which can also be tants to do the job. used to consolidate the political will among the Implementing the national communica- provincial and national governments to scale tion strategy is critical. The national monitoring up the program to national levels. The mate- and evaluation system and the impact evalua- rials can then be translated into the major na- tion program need to be put in place to ensure tional languages. The materials are an essential regular and speedy feedback during scale up. input to the planning of the national scale up. Learning-by-doing never stops, so operational

136 Accountability in Public Services in South Africa manuals and training materials should not be shift the program from basic infrastructure static, but subject to at least annual revisions to and services to economic development and so- incorporate lessons from the scaling up phase. cial protection. It should also focus on the con- The final step is a program launch workshop in comitant development of the fiscal base of local which representatives from all the key copro- governments and communities. With the basic ducers and from the next batch of lead districts structure in place, emphasis and resources can should participate. also go toward strengthening accountability After each phase of scaling up, a similar re- mechanisms, improving technical and organi- view and further refinements and adaptations to zational capability, and expanding targeted pro- the institutional frameworks need to be carried grams to tackle issues that communities may out. The consolidation phase should gradually have neglected.

Accountability in Public Services in South Africa 137

Annex 4 Land Reform in South Africa: Legacies, Limits, and Innovations

Meaningful land reform in South Africa is the Two recent experiences in land management key to attacking the acute inequalities that still show why land reform in South Africa is moving inhibit economic and social development. With- so slowly. But they also point to some underex- out it, poor communities are likely to remain ploited opportunities that may enable landown- locked in the geographic legacy of apartheid, ers, low-income communities, and the govern- which consigned the nonwhite majority to sep- ment to work more constructively to tackle the arate locations and severely limited their eco- legal, regulatory, and cultural barriers. An effort nomic opportunities. to create a mixed-use settlement outside Johan- So far, change has been frustratingly slow. nesburg, to be known as “Ethembalethu,” or More than a decade after South Africa em- “Our Hope,” is the first experience to be ana- braced democracy and a highly progressive Con- lyzed. The second involves an area-based land stitution, nearly 80 percent of farmland is owned reform and enterprise development program in by white South Africans. Low-income black or a beef-producing area subject to the KwaZulu- mixed-race citizens, meanwhile, have little ac- Natal Provincial Land Reform Office. This so- cess to land, even when they have financial re- called Qedusizi/Besters initiative has shown sources to acquire it. The scarcity of land where some results not only in restructuring local land poorer South Africans can build houses and ownership but also in mobilizing income-gen- support small-scale agricultural and commercial erating activities for a low-income community. activities makes it all but impossible for many low-income communities to improve their living Ethembalethu: a hope, an idea, an standards. opportunity to learn Land management faces centralized con- Ethembalethu is the name of a village that has trols that afford only limited involvement for yet to come into being but that remains a hope small farmers, beneficiaries, civil society orga- for the families that have saved, collaborated, nizations, and municipalities. For communities and negotiated to make it a reality. seeking land, the regulations are prohibitively The architects and potential beneficiaries of complex and rigid. The absence of a pro-poor the project have been seeking to buy land and policy for the many peri-urban areas that sur- establish a mixed-use settlement in a peri-urban round major cities is another gap that needs to area west of Johannesburg. Despite 10 years of be addressed. Finally, resistance to sharing land effort, the dream still is not realized, underscor- resources with low-income communities persists ing the lack of supportive policies for peri-urban among many wealthier South Africans. areas. Peri-urban areas, formerly rural localities

Accountability in Public Services in South Africa 139 that now occupy the edges of expanding urban Bureaucratic categories became significant. centers, are home to millions of black and Though MHTF envisioned a low-density vil- mixed-race South Africans. lage, with spaces for recreation and grazing, the For poorer South Africans, peri-urban areas category of “township” was viewed as preferable, are ideal for mixed-use settlements where they because its legal status ensured access to certain can build homes and benefit from fairly easy potentially important benefits, including hous- access to urban schools, health facilities, and ing subsidies. jobs, while having space for managing farm- Progress was stymied again in late 1998 ing and small-enterprise activities. For oth- when, with 82 percent of the land price paid, ers, these areas beckon as potential real estate the sale was cancelled. The association learned for building gated housing communities for that the NIMBY syndrome persisted: the same wealthy South Africans. The resulting tension­—­ neighbors who had scuttled the first attempt at and the absence of clear policies for peri-urban acquiring land were spoiling the second. The communities­—partly­ explains the frustrations seller offered to repay ZAR126,000 already paid experienced by the Ethembalethu planners, who toward the purchase price of ZAR154,000. The despite having the financial resources to acquire MHTF hired lawyers to protect the associa- and develop the land, have so far been snared tion’s contractual rights to the land. The seller, by a complex bureaucratic tangle of government meanwhile, charged the association with mis- policies, development programs, and environ- representation and accused MHTF members mental requirements. of trespassing. In March 1999, 47 community members were arrested in what the South Af- A story of delay and frustration rican Police Service subsequently acknowledged The Ethembalethu campaign for a new settle- were wrongful arrests. ment is a story of frustration but also of com- The dispute persisted for another two years, munity persistence. Beginning with about 250 until the High Court took the case in Novem- families living in the Muldersdrift area, the ber 2000. Once it became clear that there was community formed a nonprofit organization, the no legal basis for cancelling the sale, the land- Muldersdrift Home Trust Foundation (MHTF), owner and the white neighbors opposing the and saved about ZAR126,000 to purchase land plan offered a settlement: full return of the for a village with houses, farming plots, a grazing ZAR126,000 already paid by the community, area, and common recreation spaces. Families plus another ZAR270,000 paid by the neigh- contributed ZAR50 per year, later increasing this bors to be used to finance an alternative land commitment to ZAR100 per month, to acquire purchase so long as it lay outside a land belt pre- the necessary land. scribed by the white neighbors. The first reversal came in 1997. The MHTF Combining funds from the settlement made an offer to purchase 8.5 hectares in Mul- with their savings, the community located a dersdrift. Although a deed of sale was concluded, larger land parcel for the village, which cost the sale was aborted after a group of neighbor- ZAR650,000­—­well above the association’s ing landowners made a successful counter offer. available resources. The group appealed to the Affluent landowners have tended to block low- Gauteng Provincial Department of Housing, the income communities from acquiring land for National Department of Land Affairs, and Mo- residential and commercial activities­—­the gale City to explore options. The NIMBY syn- shorthand “Not In My Back Yard” (NIMBY) drome was reignited, however, with yet another aptly describes this pervasive attitude. group of white neighbors raising objections, this The MHTF made a second offer on an -ad time using the argument that an environmental jacent parcel of land and entered into a purchase impact assessment was missing. Although this agreement. The MHTF sought a commercial assessment is not required for a land purchase, bank loan and eventually agreed on an install- the dispute continued. ment plan, with ZAR50,000 to be paid up front In March 2002, the National Department and the balance to be paid in 15 monthly install- of Land Affairs purchased the land for the full ments of ZAR7,000. price of ZAR650,000, placing it in the name of

140 Accountability in Public Services in South Africa the municipality but under a complex arrange- shifts that support the efforts of low-income ment that did not stipulate how the land would communities to secure land. The following be used once it was transferred. The municipal- changes and innovations would help: ity then entered an agreement with the MHTF • New government policies for urban edges. under which individuals would obtain title to in- The current regulations create a dichotomy dividual plots, with businesses following a sepa- between residential and agricultural land, to rate case-by-case process. the detriment of low-income communities. The purchase final, the MHTF was re- • More willingness among municipali- quired to embark on two separate legal pro- ties to assist the settlement of low-income cesses: a “scoping environmental impact assess- families. ment” under the Environment Conservation • Integrated Development Plans that are ex- Act and a Development Facilitation Applica- plicitly pro-poor. When low-income com- tion under the Development Facilitation Act. munities cannot establish ownership, they The environmental impact assessment process tend to rent, forgoing the benefits of asset ac- provided a platform for neighbors to express cumulation, or to buy in remote areas with opposition, and the Development Facilitation limited access to jobs and services. Many re- Application ran afoul of various procedural and sort to informal settlements, with their in- technical snarls. Consultant payments and legal herent insecurity. fees were mounting. Disputes erupted over who • A restructured land market that, among would pay the various specialists and how they other things, would make more small land would be paid. Because of complications in- parcels available in peri-urban areas, includ- volving peri-urban areas, the environment ap- ing land that is presently underused. plication was rejected and the development fa- • A single planning process consolidating cilitation paper languished at the Development the environmental impact assessment and a Tribunal. township development. A final obstacle was a complication in a • A single agency responsible for land develop- UNESCO World Heritage Site near the pro- ment and land use, lessening the uncertainty posed settlement, requiring a buffer zone. Both for beneficiaries and saving time. the environmental and the development applica- • A simpler set of guidelines for building tions resided with different parts of the Gauteng projects, expressed clearly in manuals avail- government, and after another rejection in 2007, able to the public. The government could more reports were requested. also support capacity-building programs to In sum, the community mobilized its mem- support public sector lawyers and program bers, saved the necessary resources, and devised managers. a plan for acquiring land to improve lives and livelihoods. It has been thwarted by a confluence Accelerating Equitable Land Reform in of obstacles ranging from landowner attitudes to the Qedusizi/Besters Cluster Project government regulations. Ethembalethu—­ “Our­ Despite the slow pace of land reform, some new Hope”­—­is still not a reality, but it remains a mechanisms to accelerate the redevelopment of hope. land have yielded positive results. The Qedusizi/ Besters land reform project in KwaZulu-Natal Drawing the lessons shows the potential for an area-based approach The saga of Ethembalethu shines a sometimes to land reform that builds on the participation harsh light on the obstacles that block opportu- of the full range of actors­—­white farmers, com- nity not only for the 250 families backing the munities seeking to acquire land, central and project but for millions of low-income South local government, and the private sector all Africans living a tenuous existence in enclaves participate. with limited access to income-generating oppor- The key players included the Department tunities and basic services. of Land Affairs, a group of white farmers, and To dismantle the geographic legacy of apart- the Besters Farmers Association. Together they heid, South Africa needs policy and regulatory identified the land for purchase and the people

Accountability in Public Services in South Africa 141 to be settled there, negotiated a land sale, and activist leaders in the project have helped accel- launched a resettlement program. erate the progress, but in the future there will The beneficiary community consisted of 199 be a need for stronger institutions to support muzis, or extended family households, represent- profitable cattle production, small-scale farm- ing 993 adults and a total population of about ing activities, and programs to advance enter- 2,500. These households acquired 14 farms com- prises led by women and young people in the prising 14,728 hectares. Each farm consists of an community. average of 14 muzis that have been formed into There have been some gaps in participation Communal Property Associations (CPAs), each by beneficiary families. Planners didn’t fully as- with its own constitution and bylaws. Each CPA sess the needs and aspirations of the beneficia- was assigned a communally owned commercial ries at the start, particularly those of women and beef herd and specified grazing rights­—­not only youth. The white farmers and mentors designed for cattle but for goats, horses, and sheep­—­that the key elements of the project, including the could be allocated to householders and passed on farming and enterprise models, the management to future generations. system for farm machinery, and the acquisition Effective leadership was crucial to this proj- of a dairy farm. Beneficiaries would gain from ect. Champions for the Qedusizi/Besters pro- more training in team-building and negotia- gram managed to articulate a convincing vision tion, recordkeeping, and financial management. and establish processes and institutions for mak- The project did not incorporate a formal plan for ing it a reality. Committed champions from the monitoring and evaluation and would benefit Department of Land Affairs, the white farmers from a baseline study of black and white house- group, and the Besters Farmers Association were holds, the CPAs, and enterprises. critical to building consensus and managing ad- Finally, the process was slowed by the bu- ministrative hurdles. reaucratic complexities hampering land reform Several supportive arrangements helped. in South Africa. From project conception to First, the stakeholders formed a company that the launch of the first farming enterprises, the guided the development, implementation, and process took 28 months. Faster approval of the management of the land reform effort. Built planning grant would have advanced the process into the structure was a network of mentors to significantly. assist the households in managing their herds The government could ease the way for fu- and farms. A grant of ZAR23.3 million for ture land development projects by instating an “land redistribution for agricultural develop- integrated regulatory process for buying, trans- ment” covered the cost of the farms and proj- ferring, and settling land, along with the sub- ect preparation work and helped establish the sequent development of enterprises. The district herds. municipality is the natural level for coordination. Early results are encouraging: each CPA es- tablished a commercially viable herd, and some Conclusion participants generated additional revenue by The experiences of the Ethembalethu commu- leasing pastureland or planting maize and soy- nity and the cattle farmers of Qedusizi/Besters beans. The more significant revenue would come make clear that South Africa’s land reform will with the sale of the herds’ first weaned cattle. not have a single formula for success. Rather it New land continues to be made available for ex- will likely generate a panoply of innovations and pansion and additional beneficiaries. homegrown solutions over time. The need for flexibility is clear too. The goals Remaining questions and gaps of the Ethembalethu community were thwarted Some difficulties and uncertainties remain. repeatedly by regulatory processes that proved Farmers in the new settlements still need to to be overly rigid and unresponsive to local ini- supplement their income by working on neigh- tiatives. The backers of these two land-reform boring farms owned by white landowners, efforts­—­and countless others not covered in which cuts into the time spent on their own this analysis­—­would benefit from consolidat- herds and commercial activities. A number of ing and coordinating government functions,

142 Accountability in Public Services in South Africa particularly if a single, one-stop shop can be es- establish a mixed-use village such as Ethembale- tablished for land sales and transfer, settlement, thu does not need to relive every frustration that and development. impeded this group’s dream of a secure and pro- Learning and disseminating experience ductive future in a village appropriately named will ensure that the next community seeking to “Our Hope.”

Accountability in Public Services in South Africa 143

Annex 5 Project Consolidate: Hands‑On Local Government Support and Engagement Program

Capacity constraints of local governments are The DPLG then designed Project Consoli- often seen as a major impediment to meeting date, a national program based on cooperation ambitious service delivery targets. In addition, across sectors and spheres of government to ad- although not all the services are the competency dress the main capacity barriers to service deliv- of local government, municipal mayors, council- ery with a focus on the weakest municipalities. ors, and officials increasingly have become the Project Consolidate was conceived as a joint pro- target of protests aimed at forcing the state to gram with specific roles and responsibilities for recognize and address service deficiencies. One each sphere of government, for civil society, and response to the growing protests was a capacity for the private sector. The main objectives were audit carried out in 2004 in 284 municipalities by to install additional capacity, including by cre- the Department of Provincial and Local Govern- ating opportunities for skills transfer for critical ment (DPLG). This audit found that almost half technical and financial skills and operationaliz- of South African municipalities lacked key finan- ing the mechanisms laid out in policy for pub- cial and technical skills to deliver basic services. lic participation. The United States Agency for The audit identified the nation’s weakest munici- International Development, the UK’s Depart- palities. Some of the data in the survey showed ment for International Development, the Ger- that backlogs in service coverage and quality were man Technical Cooperation Agency, Old Mu- not distributed evenly across the municipalities: tual, and Eskom have contributed resources to • In 135 municipalities, unemployment is the project. greater than 35 percent of those able and A supplemental program, Siyenza Manje (we willing to work. are doing it now), has been housed at the De- • In 182 municipalities, less than 60 percent velopment Bank of Southern Africa, and with of households have access to refuse removal. the National Treasury, the DPLG, and the • In 203 municipalities, less than 60 percent South African Local Government Association, of households have access to sanitation (flush has deployed expertise in engineering, finance, toilet, chemical toilet, or septic tank). and town planning and project management. • In 122 municipalities, less than 60 percent of The purpose of the program is to complement households have access to electricity. Project Consolidate with additional capacity- • In 155 municipalities, less than 60 percent building and technical skills transfer at the mu- of households have access to clean water nicipal level. As of December 2006, under Siy- (DPLG 2004). enza Manje 81 municipalities have benefited and

Accountability in Public Services in South Africa 145 received support from 281 experts. In the June These achievements are significant given 2008 budget vote speech, the DPLG minister the limited time that Project Consolidate has shared the accomplishments of Project Consoli- been in place. However, the weaknesses of date since its inception. These include a: Project Consolidate still cause for concern in • 68 percent increase in the number of munici- mid-2007 included high vacancy rates in es- palities with 60 percent or more households sential management positions in the strug- having access to formal housing. gling municipalities targeted by the program. • 31 percent increase in the number of mu- As a result, the intended skills transfers by fi- nicipalities with 60 percent or more house- nancial and technical specialists do not hap- holds having access to water in their homes pen, and the scarce and vital skills are unde- or yards. rused as the specialists take on administrative • 48 percent increase in the number of munici- tasks (CDE 2007). In addition, the data from palities with 60 percent or more households all the complementary initiatives have not having accessing to electricity, at least for been consolidated to enable a comprehensive lighting. stock-taking of the investments and the im- • 66 percent increase in the number of munici- pact. Finally, the short timeframe in which the palities with 60 percent or more households program has been under way means that sig- having access to sanitation at the standard of nificant investment must continue for benefits a flush toilet, septic tank, or chemical toilet. to be sustainable.

146 Accountability in Public Services in South Africa References

2030 Water Resources Group. 2010. “Charting ANC (African National Congress). 1994. Recon- Our Water Future. Economic Frameworks struction and Development Programme. Johan- to inform decision-making.” McKinsey and nesburg: Raven Press. Company. www.mckinsey.com/.../Water/ ANC Constitutional Committee and Depart- Charting_Our_Water_Future_Full_Re- ment of Local and Regional Government port_001.pdf. and Housing. 1992. ANC Regional Policy. Abadzi, H. 2009. “Instructional Time Loss in Draft. The Centre for Development Studies, Developing Countries: Concepts, Measure- University of the Western Cape, Cape Town. ment, and Implications.” World Bank Research Andrabi, T., J. Das, and A. Khwaja. 2009. “Report Observer 24 (2): 267–90. Cards: The Impact of Providing School and Aguero, J., M. Carter, and J. May. 2008. “Poverty Child Test Scores on Educational Markets.” and Inequality in the First Decade of South Mimeo. World Bank, Washington, DC. Africa’s Democracy: What can be Learnt APRM (African Peer Review Mechanism). 2007. from Panel Data from KwaZulu-Natal?” “Country Report No. 4: South Africa.” The Journal of African Economies 16 (5): 782–812. New Partnership for African Development. Aiyar, S. 1995. “How Well Has Decentralization Ardington, C., D. Lam, M. Leibbrandt, and M. Worked for Rural Development?” AGR Dis- Welch. 2005. “The Sensitivity to Key Data semination Notes 2, World Bank, Washing- Imputations of Recent Estimates of Income ton, DC. Poverty and Inequality in South Africa.” Eco- ———. 1996. “Has Decentralization Aided Bio- nomic Modelling 23 (5): 822–35. diversity Conservation?” AGR Dissemination Asian Development Bank. 2005. “Progress Re- Notes 11, World Bank, Washington, DC. port on the Program Loan Cluster, Phase I, Aiyar, S., K. McLean, and S. Piriou-Sall, ed. 1996. of the Program Loan to the Socialist Repub- “The Political Economy of Democratic De- lic of Vietnam to Support Implementation of centralization.” AGR Dissemination Notes 9, the Public Administration Reform Master World Bank, Washington, DC. Program in Vietnam.” http://www.adb.org/ Ajama, T., and J. Aron. 2008. “Fiscal Renaissance Documents/Reports/PLC/sf-vie-1992.pdf. in a Democratic South Africa.” Journal of Af- Balfour, F. 2008. Personal communication, Jef- rican Economies 16 (5): 745–81. fares & Green (Pty) Ltd., Pietermaritzburg, Amin, S., J. Das, and M. Goldstein. 2007. Are You South Africa. June 6. Being Served? New Tools for Measuring Service Banerjee, A., R. Banerji, E. Duflo, R. Glenne- Delivery. Washington, DC: World Bank. rster, and S. Khemani. 2008. “Pitfalls of

Accountability in Public Services in South Africa 147 Participatory Programs: Evidence from a Confluence of the Water Industry.” Sun City, Randomized Evaluation in Education in South Africa, May 21. India.” Policy Research Working Paper 4584, Bhorat, H., and R. Kanbur. ed. 2005. “Poverty World Bank, Washington, DC. and Well-being in Post-Apartheid South Af- Banerjee, A., A. Deaton, and E. Duflo. 2004. rica: An Overview of Data, Outcomes and “Health Care Delivery in Rural Rajasthan.” Policy.” Development Policy Research Unit Economic and Political Weekly 39 (9): 944–49. Working Paper 101, School of Economics, Banerjee, A., and E. Duflo. 2006. “Addressing University of Cape Town, Cape Town. Absence.” Journal of Economic Perspectives 20 Binswanger-Mkhize, H.P., J. de Regt, and S. (1): 117–32. Spector. 2010. Local and Community-driven Banglapedia. 2005. “The Comilla Model.” Dhaka, Development: Moving to Scale in Theory and Bangladesh. Practice, Washington, DC: World Bank. Bardhan, P., and D. Mookherjee. 2000. “Cap- Binswanger, H.P, and T.V. Nguyen. 2005. “A ture and Governance at Local and National Step by Step Guide to Scale Up Commu- Levels.” American Economic Review 90 (2): nity Driven Development.” Paper presented 135–39. at the Water Management Institute interna- Barichievy, K., L. Piper, and B. Parker. 2005. “As- tional workshop, “African Water Laws: Plu- sessing ‘Participatory Governance’ in Local ral Legislative Frameworks for Rural Water Government: A Case of Two South African Management in Africa,” Johannesburg, Janu- Cities.” Politikon 24 (3): 370–93. ary 26–28. Barron, P., C. Day, F. Monticelli, K. Vermaak, O. Björkman, M. 2007. “Does Money Matter Okorafor, K. Moodley, and T. Doherty. 2006. for Student Performance? Evidence from The District Health Barometer 2005/06. Dur- a Grant Program in Uganda.” Innocenzo ban, South Africa: Health Systems Trust. Gasparini Institute for Economic Re- Bekker, S., and A. Leildé. 2003. “Residents’ Per- search Working Paper 326, Bocconi Uni- ceptions of Developmental Local Govern- versity, Milan. ftp://ftp.igier.uni-bocconi.it/ ment: Exit, Voice and Loyalty in South Afri- wp/2007/326.pdf. can Towns.” Politikon 22 (1):144–65. Björkman, M., R. Reinikka, and J. Svensson. Belli, P., M. Finger, and A. Ballivian. 1993. 2006. “Local Accountability.” Institute for “South Africa: A Review of Trade Policies.” International Economic Studies Seminar Informal Discussion Papers on Aspects of the Paper 749, Stockholm University, Stockholm. South African Economy 4. Southern Africa Björkman, M., and J. Svensson. 2007. “Power to Department, World Bank, Washington, DC. the People: Evidence from a Randomized Benson, F. 2009. “Overview of the Cholera Out- Experiment of a Citizen Report Card Project break.” Briefing to the Parliamentary Portfo- in Uganda.” Centre for Economic Policy Re- lio Committee on Water Affairs. South Af- search Working Paper 6344, Centre for Eco- rica Department of Health. http://www.pmg. nomic Policy Research, London. org.za/report/20090204-department-health Blecher, M., and S. Harrison. 2006. “Health Care -department-water-affairs-forestry-responses Financing.” In South African Health Review -cholera. 2006, ed. P. Ijumba and A. Padarath, 31–64. Berrisford, S., D. DeGroot, M. Kihato, N. Mar- Durban, South Africa: Health Systems Trust. rengane, Z. Mhlanga, and R. van den Brink. Booysen, S. 2007. “With the Ballot and the 2008. “In Search of Land and Housing in the Brick: The Politics of Attaining Service De- New South Africa: The Case of Ethembale- livery.” Progress in Development Studies 7 (1): thu.” Working Paper 130, Office of the Pub- 21–32. lisher, World Bank, Washington, DC. Boyd, L.A. 2004. “Development of Regulations Bhagwan, J., K. Wall, and O. Ive. 2008. “Unlock- for Water Care Works and Process Person- ing the Potential of Franchising Partnerships nel.” Water SA (30) 5: 155–57. to Improve Water Services Compliance.” Buccus, I., D. Hemson, J. Hicks, and L. Piper. Unpublished paper presented at the Water 2008. “Community Development and En- Institute of South Africa conference, “The gagement with Local Governance in South

148 Accountability in Public Services in South Africa Africa.” Community Development Journal 43 articles/attachments/10214_schools_that_ (3): 297–311. work_ministerial_committee_report.pdf. Buckley, C., K.M. Foxon, N. Rodda, C.J. Brouck- Cloete, J.J.N., and C. Thornhil. 2005. South Af- aert, S. Mantovanelli, and M. Mnguni. 2008. rican Municipal Government and Adminis- “Research into Urine Diversion, Ventilated tration: A New Dispensation. Johannesburg: Improved Double Pit Toilets (UD /VIDP) Dotsquare Publishing. in eThekwini.” Research Report 1629/1/08, Coates, D. 2003. “Education Production Func- Water Research Commission, Pretoria, South tions Using Instructional Time as an Input.” Africa. Education Economics 11 (3): 273–92. Burger, R. 2007. “Policy Brief: How Pro-poor CoGTA (Department of Cooperative Gover- is the South African Health System?” Eco- nance and Traditional Affairs). 2009a. The nomic Working Paper 6/07, Department of State of Local Government Report. Pretoria, Economics, University of Stellenbosch, Stel- South Africa: Government Printers. lenbosch, South Africa. ———. 2009b. Local Government Turnaround Burger, R., and C. Swanepoel. 2006. “Have Pro- Strategy. Working Together, Turning the Tide poor Health Policies Improved the Target- in Local Government. Pretoria, South Africa: ing of Spending and the Effective Delivery Government Printers. of Health Care in South Africa?” Economic ———. 2010. “Turning Around Local Govern- Working Paper 12/06, Department of Eco- ment.” Presentation to the Adhoc Commit- nomics, University of Stellenbosch, Stellen- tee on Service Delivery, February 2. ww.pmg. bosch, South Africa. org.za/docs/2010/100202cogta.ppt. Carrim, Y. 2009. “The National Local Govern- Collins, D., V. Brijlal, B. Makan, and P. Cross. ment Turnaround Strategy Framework: 2000. “Why do some South African Provin- Some Issues.” Lecture, Development Bank cial Governments Spend More on Health of South Africa Knowledge Week. Pretoria, Care than Others?” Mimeo. Equity project South Africa, November 4. www.thedplg.gov. discussion paper, Department of Health, Pre- za/index.php/component/content/96.html?. toria, South Africa. Case, A., and A. Deaton. 1999. “School Inputs Cotlear, D. 2006. “Improving Education, Health and Educational Outcomes in South Af- Care, and Social Assistance for the Poor.” In rica.” Quarterly Journal of Economics 114 (3): A New Social Contract for Peru: An Agenda for 1047–84. Improving Education, Health Care, and the So- Cashdan, B. 2002. “Local Government and Pov- cial Safety Net, ed. D. Cotlear, 1–37. A World erty in South Africa.” In Democratising Local Bank Country Study. Washington, DC: The Government: The South African Experiment, International Bank for Reconstruction and ed. S. Parnell, E. Pieterse, M. Swilling, and D. Development. Wooldridge. Cape Town: Juta and Company. Crook, R.C., and J. Manor. 1998. Democracy and Chaudhury, N., J. Hammer, M. Kremer, K. Mu- Decentralisation in South Asia and West Africa. ralidharan, and F. Halsey Rogers. 2006. Cambridge: Cambridge University Press. “Missing in Action: Teacher and Health Crouch, L. 1996. “Public Education Equity Worker Absence in Developing Coun- and Efficiency in South Africa: Lessons for tries.” Journal of Economic Perspectives. 20 (1): Other Countries.” Economics of Education Re- 91–116. view 15 (2): 125–37 Chisholm, L., U. Hoadley, M. wa Kivulu, K. CSIR (Centre for Scientific and Industrial Re- Brookes, C. Prinsloo, A. Kgobe, D. Mosia, search). 2006. Database on Tracking Voting H. Narsee, and S. Rule. 2005. Educator Work- Changes in the South African Electorate. load in South Africa. Pretoria, South Africa: Pretoria, South Africa. Human Sciences Research Council. Daily Dispatch. 2009. http://blogs.dispatch. Christie, P., B. Butler, and M. Potterson. 2007. co.za/dispatchnow/2009/07/21/crisis-in “Schools That Work: Report to the Minister -mthatha. of Education.” South Africa Department of Das, J., S. Dercon, J. Habyarimana, and P. Krish- Education. http://llnw.creamermedia.co.za/ nan. 2005. “Teacher Shocks and Student

Accountability in Public Services in South Africa 149 Learning: Evidence from Zambia.” Policy ———. 2008d. “National Water Services Regu- Research Working Paper 3602, World Bank, lation Strategy.” Final Draft. Pretoria, South Washington, DC. Africa. Department of Education. 2008. Education Statis- ———. 2009a. Annual Report. 1 April 2008 to tics in South Africa 2006. Pretoria, South Africa. 31 March 2009. Pretoria, South Africa. Department of Health. 2007. A Policy on Quality ———. 2009b. “Under-capacitated municipali- in Health Care in South Africa. Pretoria, South ties as at 30 October 2009.” Unpublished in- Africa. ternal document, Pretoria, South Africa. ———. 2008. “Report on Childhood Deaths, ———. 2010. “Strategic Perspective on Water Uhkahlamba District, Eastern Cape. http://www. Services in South Africa.” Pretoria, South health.org.za/documents/7f3d48106e60e6e5 Africa. d164698624c5efb3.pdf?PHPSESSID= DWAF and CSIR. 2007. “Spot Check Assess- 1a9e0a0951. ment of MIG-funded Water and Sanitation DPLG (Department of Provincial and Local Projects.” www.dwaf.gov.za/Masibambane/ Government). 1998. “White Paper on Local documents/monitoring-evaluation/ Government.” Pretoria, South Africa: Gov- MIGSpotcheckApr07part1.pdf. ernment Printers. Eales, K. 2008. “Water Services in South Africa, ———. 2004. “White Paper on Municipal Ser- 1994–2008.” In Water Policy in South Africa: vice Partnerships.” Pretoria, South Africa: Resources for the Future, ed. B. Schreiner and Government Printers. R. Hassan. Forthcoming. ———. 2008. Provincial and Local Government ———. 2009. Personal communication. Depart- Department Budget Vote 2008/09. Speech ment of Water Affairs, Pretoria. December by the Minister for Provincial and Local 10. Government, Fholisani Sydney Mufamadi. ———. 2010a. “Water Services in South Africa, Duflo, E., and R. Hanna. 2005. “Monitor- 1994–2009.” In A Policy Framework for Trans- ing Works: Getting Teachers to Come to forming Water Management in South Africa, School.” Mimeo. Massachusetts Institute of ed. B. Schreiner and R. Hassan. New York: Technology, Cambridge. Springer. Duflo, E., R. Hanna, and S. Ryan. 2007. “Moni- ———. 2010b. “Water, Sanitation and Waste- toring Works: Getting Teachers to Come to water Management: Some Questions for School.” Revision Invited, American Eco- National Water Security in South Africa.” nomic Review. (See NBER Working Paper In A Policy Framework for Transforming Water 11880, 2005; BREAD Working Paper 103, Management in South Africa, eds. B. Schreiner November 2007.) and R. Hassan. New York: Springer. DWAF (Department of Water Affairs and For- Everatt, D., and L. Gwagwa. 2005. “Commu- estry). 1994. “White Paper on Water Supply nity Driven Development in South Africa, and Sanitation.” Pretoria, South Africa. 1990–2004.” Africa Region Working Paper ———. 2007. “Summative evaluation of the Ma- 92, World Bank, Washington, DC. sibambane II Programme, Project Evaluation Ezpeleta, J., and E. Weiss. 1996. “Las escuelas Report.” Pretoria, South Africa. rurales en zonas de pobreza y sus maestros: ———. 2008a. “Consolidated Water Sector Re- tramas preexistentes y políticas innovadoras.” port, Fourth Quarter 2007/08, Vol. 1.” Un- Revista Mexicana de Investigación Educativa published report. Pretoria, South Africa. 1: 53–69. ———. 2008b. “Water for Growth and Develop- Faguet, J.P. 1997. “Decentralization and Local ment Strategy Discussion Document.” Preto- Government Performance.” Paper presented ria, South Africa. at the FAO Technical Consultation on De- ———. 2008c. “Drinking Water Quality in centralization, FAO, Rome, December 16–18. South Africa.” Unpublished presentation Figlio, D.N., and L.S. Getzler. 2002. “Account- to the Parliamentary Portfolio Committee ability, Ability, and Disability: Gaming the Public Hearings on Water Quality.” June 4. System.” Working Paper 9307, National Bu- www.pmg.org.za/files/docs/080604dwaf.ppt. reau of Economic Research, Cambridge.

150 Accountability in Public Services in South Africa Filmer, D. 2003. “The Incidence of Public Ex- United Kingdom: Intermediate Technology penditures on Health and Education.” Development Group. Background paper to the 2004 World De- Galvin, M., and A. Habib. 2003. “The Politics velopment Report. http://econ.worldbank. of Decentralisation and Donor Funding in org/WBSITE/EXTERNAL/EXTDEC/ South Africa’s Rural Water Sector.” Journal of EXTRESEARCH/EXTWDRS/ Southern African Studies 29 (4): 865–84. EXTWDR2004/0,,ImgPagePK:64202988 GCIS (Government Communication and Infor- ~entityID:000160016_20031020130801~ mation System). 2005. South Africa Year- menuPK:64217940~pagePK:64217930~ book 2004/05. Pretoria, South Africa. piPK:64217936~theSitePK:477688,00.html. ———. 2008. South Africa Yearbook 2007/08. ———. 2005. “Teacher Pay in Indonesia.” Draft. Pretoria, South Africa. Development Research Group, World Bank, Glewwe, P., N. Ilias, and M. Kremer. 2003. Washington, DC. “Teacher Incentives.” Working Paper 9671, Fiske, E.R., and H.F. Ladd. 2004. Elusive Equity: National Bureau of Economic Research, Education Reform in Post-Apartheid South Af- Cambridge. rica. Washington, DC: Brookings Institution Goldman, I. 2009. “Experiences of Commu- Press. nity-based Planning—­ Lessons­ from South Fiszbein, A. 1997. “The Emergence of Local Ca- Africa, Uganda, Zimbabwe, and Ghana. pacity: Lessons from Colombia.” World De- Khanya-African Institute for Community- velopment 25 (7): 1029–43. Driven Development presentation. Fiszbein, A., ed. 2005. Citizens, Politicians and ———. 2010. “Building on the Energy of Our Providers: The Latin American Experience with People: Lessons from the S/E Africa on Service Delivery Reform. Washington, DC: Local and Community-driven Develop- World Bank. ment. Presentation at the Human Sciences Foundation for Contemporary Reserach. 2007. Research Council Seminar Series, Pretoria, “A Handbook on Community Development South Africa, July 6. Workers in South Africa.” http://www.fcr. Gounden, T. 2008. Personal communication. org.za/publications/case-studies/relevant June 5. -literature/cdw-handbook.pdf/view. Government of Tanzania. 2004. “Tanzania Public Francken, N. 2003. “Service Delivery in Public Expenditure Tracking Study. Study of Financial Primary Schools in Madagascar: Results of a and Non-Financial PEDP Flows from Cen- Budget Tracking Survey.” Processed. Mada- tral Government to Schools in 2002 and 2003.” gascar Country Office, Antananarivo, World Ministry of Finance, Dar es Salaam, Tanzania. Bank. Grobler, C., and I. Stuart. 2007. “Health Care Friedman, S. 2005. “South Africa: Globalization Provider Choice.” South African Journal of and the Politics of Redistribution.” In Getting Economics 75 (2): 327–50. Globalization Right: The Dilemmas of Inequal- Groenewald, Y. 2009. “Eastern Cape on the ity, ed. J. Tulchin and G. Bland. Boulder, CO Brink of a Water Crisis.” Mail & Guardian, and London: Lynne Rienner. June 12. ———. 2006a. Participatory Governance and Gustafsson, M. 2007. “Using the Hierarchical Citizen Action in Post-Apartheid South Af- Linear Model to Understand School Produc- rica. Discussion Paper 164/2006, Interna- tion in South Africa.” Economic Working tional Institute for Labour Studies, Geneva. Paper 01/07, Stellenbosch University, Stel- ———. 2006b. “Delivery Fetish Obscures lenbosch, South Africa, and National Bureau Deeper Meaning of Government Service.” of Economic Research, Cambridge. Business Day. March 26. Gustafsson, M., and F. Patel. 2006. “Undoing the Friedman, S., K. Hlela, and P. Thulare. 2005. “A Apartheid Legacy: Pro-poor Spending Shifts Question of Voice: Informality and Pro-poor in the South African Public School System.” Policy in Johannesburg, South Africa.” In Perspectives in Education 24 (2): 65–77. Urban Futures: Economic Growth and Pov- ———. 2008. “Managing the Teacher Pay Sys- erty Reduction, ed. N. Hamdi, 51–68. Rugby, tem: What the Local and International Data

Accountability in Public Services in South Africa 151 are Telling Us.” Economic Working Paper IEC (Independent Electoral Commission). 27/06, University of Stellenbosch, Stellen- 2008. Municipal Elections 2006. http://www. bosch, South Africa. elections.org.za/MunicElections2006.asp. Habib, A. 2009. “South Africa: Conceptualiz- Instituto Apoyo and World Bank. 2002. “Public ing a Politics of Human-Oriented Devel- Expenditure Tracking Survey: The Education opment.” http://www.polity.org.za/article/ Sector in Peru.” World Bank, Washington, south-africa-conceptualizing-a-politics-of DC. -human-oriented-development-2009-08-13. Jacob, B.A., and S.E. Levitt. 2003. “Rotten Ap- Habib, A., and H. Kotzé. 2003. “Civil Society, ples: An Investigation of the Prevalence and Governance and Development in an Era of Predictors of Teacher Cheating.” Quarterly Globalisation: The South African Case.” In Journal of Economics 118 (3): 843–78. Governance in the New South Africa: The Chal- Khanya-aicdd (Khanya-African Institute for lenges of Globalisation, ed. G. Mhone and O. Community-driven Development). 2008. Edigheji, 246–70. Cape Town: University of “Participatory Service Delivery Assessment Cape Town Press. in Six Municipalities in South Africa.” Draft. Hanushek, E. 1996. “Outcomes, Costs, and In- Kremer, M., N. Chaudhury, F.H. Rogers, K. Mu- centives in Schools.” In Improving America’s ralidharan, and J. Hammer. 2005. “Teacher Schools: The Role of Incentives, ed. E.A. Ha- Absence in India: A Snapshot.” Journal of the nushek and D.W. Jorgenson, 29–52. Wash- European Economic Association 3 (2–3): 658–67. ington, DC: National Academy Press. Kremer M., and D. Chen. 2001. “An Interim Re- Heemskerk, W., and G. Baltissen. 2005. “Com- port on a Teacher Attendance Incentive Pro- munity-driven Development: Toolkit for gram in Kenya.” Mimeo. Harvard University, National Stocktaking and Reviews.” Royal Cambridge. Tropical Institute (KIT): Amsterdam. Lavy, V. 2002. “Evaluating the Effect of Teach- http://info.worldbank.org/etools/docs/ ers’ Group Performance Incentives on Pupil library/209137/CDD%20Stocktaking%20 Achievement.” Journal of Political Economy and%20review%20toolkit_update%20dec05. 110: 1286–1317. pdf. Lawless, A. 2007. Numbers and Needs in Local Hendricks, L.B. 2008. Budget Vote 34 of Government. Midrand, South Africa: South 2008/09. Speech, Parliament of South Africa. African Institution of Civil Engineering Cape Town, May 23. http://www.dwaf.gov. ———. 2008. Personal communication. January za/minister/Default_dates06.asp?year=2008. 28. Hernani-Limarino, W. 2005. “Are Teachers ———. 2009. Presentation to the Local Govern- Well Paid in Latin America and the Carib- ment Indaba. Boksburg, South Africa. No- bean? Relative Wage and Structure of Re- vember 29. turns of Teachers.” In Incentives to Improve ———. 2010. Personal communication to K. A. Teaching: Lessons from Latin America, ed. E. Eales, Department of Water Affairs. Pretoria, Vegas, 63–102. Washington, DC: World South Africa. January 22. Bank. Marais, H., D. Everatt, and N. Dube. 2007. “The HSRC (Human Sciences Research Coun- Depth and Quality of Public Participation in cil). 2005. Educator Workload in South Af- the Integrated Development Planning Pro- rica. By L. Chisholm, U. Hoadley, M. wa cess in Gauteng.” Gauteng Provincial De- Kivulu, H. Brookes, C. Prinsloo, A. Kgobe, partment of Local Government. D. Mosia, H. Narsee, and S. Rule. Preto- Masango, R. 2002. “Public Participation: A Crit- ria, South Africa: HSRC Press. http:// ical Ingredient of Good Governance.” Poli- www.hsrcpress.ac.za/product.php? tikon 21 (2): 52–65. productid=2120&cat=8&page=2. McEwan, P.J., and L. Santibañez. 2005. “Teacher ———. 2006. “Consultative Citizens’ Report Incentives in Mexico.” In Beyond Pay: Moti- Card Survey, City of Tshwane, South Africa.” vating Teachers to Raise Student Learning. Les- Survey Report, World Bank, Washington, sons from Latin America, ed. E. Vegas. Wash- DC. ington, DC: World Bank.

152 Accountability in Public Services in South Africa McIntyre, D., D. Muirhead, and L. Gilson. 2002. Research Commission, Pretoria, South “Geographic Patterns of Deprivation and Africa. Health Inequalities in South Africa: Inform- Municipal Demarcation Board. 2004. “Shap- ing Public Resource Allocation Strategies.” ing South Africa: Reflections on the First Health and Policy Planning 17 (1): 30–39. Term of the Municipal Demarcation Board, McKay, V. 2001. How Local Government Works. South Africa 1999–2004.” Pretoria, South Braamfontein: Planact. Africa. McLean, K., G. Kerr, S. Piriou-Sall, and M. Wil- Municipal IQ. 2010. “Service Delivery Protests: liams. 1998. “Decentralization and Rural First Quarter 2010.” http://www.municipaliq. Development: Characterizing Efforts in 19 co.za/press/press_release_service_delivery_ Countries.” Working Paper, World Bank, protests_update_q1_2010.doc. Washington, DC. Muralidharan, K., and V. Sundararaman. 2006. Meth, C. 2006. “Income Poverty in 2004: A “Teacher Incentives in Developing Coun- Second Engagement with the Recent van tries: Experimental Evidence from India.” der Berg et al. Figures.” Working Paper 47, Draft Working Paper 42129. World Bank, School of Development Studies, University Washington, DC. http://www.people.fas.har- of KwaZulu-Natal, South Africa and South- vard.edu/~muralidh/Karthik_Muralidharan_ ern African Labor and Development Re- JMP_Teacher_Incentives_In_Developing_ search Unit, University of Cape Town. Countries.pdf. Millot, B., and J. Lane. 2002. “The Efficient Use Murnane, R.J., and D.K. Cohen. 1986. “Merit of Time in Education.” Education Economics Pay and the Evaluation Problem: Why Most 10 (2): 209–28. Merit Pay Plans Fail and a Few Survive.” Mizala, A., and P. Romaguera. 2003. “Regu- Harvard Educational Review 56 (1): 1–17. lación, Incentivos y Remuneraciones de los Narayan, D. 1995. “The Contribution of Peo- Profesores en Chile.” in Políticas Educacionales ple’s Participation: Evidence from 121 Rural en el Cambio de Siglo: La Reforma del Sistema Water Supply Projects.” Environmentally Escolar de Chile, ed. C. Cox. Santiago: Edito- Sustainable Development Occasional Paper rial Universitaria. Series 1, World Bank, Washington, DC. Momba, M.N.B., C.L.I. Obi, and P. Thompson. National Assembly. 2007. Parliamentary Portfo- 2008. “Improving Disinfection Efficiency in lio Committee on Water Affairs and Forestry Small Drinking Water Treatment Plants.” Portfolio Committee. May 4. Water SA 30 (5): 69–74. National Treasury. 2008. “The 2008 Local Gov- Momba, M.N.B., A.N. Osode, and M. Sibewu. ernment Budgets and Expenditure Review 2006. “The Impact of Inadequate Wastewater 2003/04–2009/10.” Cape Town. Treatment on the Receiving Water Bodies—­ ­ ———. 2009. “Press Release: 2008/09 Munici- Case study: Buffalo City and Nkokonbe Mu- pal Budget Information: Fourth Quarter Fi- nicipalities of the Eastern Cape Province.” nancial Results as at 30 June 2009.” http:// Water SA 32 (5): 687–92. www.treasury.gov.za/comm_media/news_ Monga, C. 1996. The Anthropology of Anger: Civil archive_2009.aspx. Society and Democracy in Africa. Boulder, CO Ndaba, D. 2010. “Now R23bn Needed to Prevent and London: Lynne Rienner. Wastewater Treatment Collapse.” Engineering Motala, S. 2006. “Education Resourcing in Post- News. May 14. http://www.engineeringnews. Apartheid South Africa: The Impact of Fi- co.za/article/about-r23-billion-to-prevent nance Equity Reforms in Public Schooling.” -the-countrys-wastewater-treatment-works Perspectives in Education 24 (2): 79–93. -from-collapsing-2010-05-14. Muller, M. 2007. “Parish Pump Politics: The Pol- Palmer, N. 1999. “Patient Choice of Primary itics of Water Supply in South Africa.” Pro- Care Provider.” In South African Health Re- gressive Development Studies 7 (1): 33–45. view 1999, ed. N. Crisp. Durban: Health ———. 2009. “Towards the Regulation of the System Trust. http://www.hst.org/sahr. Competencies of South Africa’s Water Ser- Palmer, N., A. Mills, H. Wadee, L. Gilson, and vices Managers.” Report TT 401/09. Water H. Scheider. 2002. “A New Face for Private

Accountability in Public Services in South Africa 153 providers in Developing Countries: What Rapoo, T. 2004. “Rating the Effectiveness of Implications for Public Health?” Mimeo. Legislative Oversight Methods and Tech- London, London School of Hygiene and niques at Provincial Level: The Views of Se- Tropical Medicine. nior Public Service Officials.” Policy Brief 34, Pandey, P., S. Goyal, and V. Sundararaman. 2009. Centre for Policy Studies, London. http:// “Community Participation in Public Schools: www.cps.org.za/cps%20pdf/polbrief34.pdf. Impact of Information Campaigns in Three In- Rawlings, L.B., L. Darling Sherburne-Benz, and dian States.” Education Economics 17 (3): 355–75. L. Van Domelen. 2003. Evaluating Social Patel, F., and L. Crouch. 2008. “South African Funds: A Cross-country Analysis of Community Education Investment: A Rights or Skills Investments. Washington, DC: World Bank. Agenda?” In Investment Choices for South Af- Reddy, V. 2006. “Mathematics and Science rican Education, ed. L. Bloch, L. Chisholm, Achievement at South African Schools in and B. Fleisch. Johannesburg: University of TIMSS 2003.” Cape Town: HSRC Press. the Witwatersrand Press. Reinikka, R. 2007. Extracts of speech given as the PCAS (Policy Co-ordination and Advisory Ser- Economic Society of South Africa. September 11. vice), The Presidency. 2003. “Towards a Ten Reinikka, R., and J. Svensson. 2004. “Fighting Year Review.” Synthesis Report on Imple- Corruption to Improve Schooling: Evidence mentation of Government Programmes. Dis- from a Newspaper Campaign in Uganda.” cussion Document. http://www.info.gov.za/ Journal of the European Economic Association 3 otherdocs/2003/10year.pdf. (2–3): 259–67. Philip, K. 2009. “The Community Work Pro- Republic of Kenya. 2005. “Public Expenditure gram in South Africa.” Presentation at the Tracking Survey (PETS) 2004 Preliminary Workshop on the “Second Economy” Strat- Report.” Ministries of Planning and National egy Project, an Initiative of the Presidency of Development, Finance, Health; and Higher South Africa. Education Science and Technology. Pieterse, E., S. Parnell, M. Swilling, and M. van Republic of South Africa. 1996. “The Constitu- Donk. 2008. “Consolidating Developmental tion of the Republic of South Africa, Act 108 Local Government” In Consolidating Develop- of 1996.” Pretoria: Government Printers. mental Local Government: Lessons from the South ———. 1999. “Intergovernmental Fiscal Re- African Experience, ed. M. Van Donk, M. Swill- view.” Department of Finance, September. ing, E. Pieterse, and S. Parnell. An Islanda In- ———. 2007. “Provincial Budgets and Expen- stitute Book Project. Cape Town: UCT Press. diture Review 2003/04–2009/10.” National PSC (Public Service Commission). 2006. “Re- Treasury, September. port on the Evaluation of Performance and Santibañez, L. 2008. “Literature Review on Teacher Compliance with the Batho Pele Principles Contracting: Review of the Theory and Evi- of Redress.” www.psc.gov.za. dence Around the World.” Informal Draft. ———. 2007. “Citizen Satisfaction Survey: Eco- Santibañez, L., J.F. Martínez, A. Datar, P.J. McE- nomic and Infrastructure Services Sector.” wan, C.M. Setodji, and R. Basurto-Dávila. www.psc.co.za. 2007. Breaking Ground: Analysis of the As- ———. 2008a. “A Mid-term Review of Public sessments and Impact of the Carrera Magiste- Service Transformation.” State of the Public rial Program in Mexico. Santa Monica, CA: Service Report. www.psc.gov.za. RAND/MG-141. ———. 2008b. “Report on the Evaluation of the Savage, D., K. Eales, and L. Smith. 2008. “Secur- Implementation of the Batho Pele Principle ing South Africa’s Water Sector to Support of Consultation.” www.psc.gov.za. Sustainable Growth and Development.” Un- ———. 2009. “State of the Public Service Re- published manuscript. April. port.” www.psc.gov.za. Schoeman, G., and N. Mjoli. 2006. “Assess- Rampedi, P. 2010. “Borehole Tender Shame: ment of Training Programmes and Capacity ANC Cadres Profit from Taxpayers’ R25m Needs for the Water Sector.” Water Research but Fail to Fix Limpopo Water Crisis.” City Commission Report K8/664, Pretoria, South Press, June 27. Africa.

154 Accountability in Public Services in South Africa Schoeman, M., and C. Puttergill. 2007. “Voting Be- ———. 2010a. “Non-financial Census of Mu- haviour in the SA Local Government Elections nicipalities for the Year Ended June 2009”. of 2006 with Specific Reference to the Youth.” Report P115, Pretoria, South Africa. Journal of African Elections 6 (1): 152–73. ———. 2010b. General Household Survey 2009. Scott, J. 1987. Weapons of the Weak: Everyday Statistical Release P0318. Pretoria, South Forms of Peasants Resistance. New Haven, CT: Africa. Yale University Press. Steffensen, J., and S. Trollegard. 2000. Fiscal De- Seddon, J. 2008. Systems Thinking in the Public centralization and Subnational Government Sector. The Failure of the Reform Regime . . . and Finance in relation to Infrastructure and Service a Manifesto for a Better Way. Axminster, UK: Provision. Washington, DC: World Bank. Triarchy Press. Takeuchi, H., E. Osono, and N. Shimizu. 2008. Seekings, J. 2007. “Poverty and Inequality after “The Contradictions that Drive Toyota’s Suc- Apartheid.” CSSR Working Paper 200, Cen- cess.” Harvard Business Review. June. http:// tre for Social Science Research, University www.hbsp.harvard.edu/hbsp/hbr/index. of Cape Town. http://www.sarpn.org.za/ jsp?ml_issueid=BR0806&_requestid=264340. documents/d0003024/Poverty_inequality_ Taylor, N. Forthcoming. “Equity, Efficiency and Seekings_Sep2007.pdf the Development of South African Schools.” Shiceka, S. 2010. “Budget Speech to the Na- In International Handbook of School Effective- tional Assembly.” April 22. www.dplg.gov.za/ ness and Improvement, ed. T. Townsend. Dor- ?option=com_content&view=article& drecht, the Netherlands: Springer. id=119:mr-sicelo-shiceka-present-the-budget Tendler, J. 1999. “The Rise of Social Funds: -vote&catid=3:minister-sicelo-shiceka. What are they a Model of?” Unpublished Sigodi Marah Martin. 2007. “Services Barom- manuscript. http://www.ifad.org/rural/ eter Study: The State of Community Con- learningnotes/hsa/3.htm. sultation in the Provision of Water Services.” TIPS (Trade & Industrial Policy Strategies). Research Report K5/1616. Water Research 2009. “2009–10 Community Works Program Commission, Pretoria, South Africa. Report.” Smith, L., and N. Morris. 2008. “Municipal Umansky, I., and E. Vegas. 2007. “Inside Decentral- Entities: A Panacea for Service Delivery ization: How Three Central American School- Constraints.” In Consolidating Developmen- based Management Reforms Affect Student tal Local Government: Lessons from the South Learning Through Teacher Incentives.” The African Experiment, ed. M. Van Donk, M. World Bank Research Observer 22 (2): 197–215. Swilling, E. Pieterse, and S. Parnell. An Is- UNDP (United Nations Development Pro- landa Institute Book Project. Cape Town: gramme). 2002. “Case Studies on Sustain­ UCT Press. ability in Local Governance.” Environmental Smith, L., and A. Vawda. 2003. “Citizen vs. Evaluation of the University of Cape Town, Customer­—­Different Approaches to Pub- UNDP, South Africa. lic Participation in Service Delivery in Cape Van den Brink, R., G.S. Thomas, and H.P. Town.” Urban Forum 14 (1): 26–52. Binswanger. 2007. “Agricultural Land Redis- ———. 2007. Community Survey 2007. Statis- tribution in South Africa: Towards Acceler- tical Release P0301. Pretoria, South Africa. ated Implementation.” In: The Land Question Sonjica, B. 2009. “National Assembly: Ques- in South Africa: The Challenge of Transforma- tion 186 for Written Reply.” http://www. tion and Redistribution, ed. L. Ntsebeza and dwaf.gov.za/communications/Q&A/2009/ R. Hall, 152–201. Cape Town: HSRC Press. Q186June09.pdf. Van der Berg, S. 2005. “Fiscal Expenditure In- Stats SA (Statistics South Africa). 2005. “Achiev- cidence in South Africa, 1995 and 2000: A ing a Better Life for All: Progress between Report to the National Treasury.” Unpub- Census ’96 and Census 2001.” Report 02-02- lished manuscript. http://www.sarpn.org.za/ 16, Pretoria, South Africa. documents/d0001113/index.php. ———. 2007. Community Survey 2007. Statis- ———. 2006a. “Public Spending and the Poor tical Release P0301. Pretoria, South Africa. since the Transition to Democracy.” In

Accountability in Public Services in South Africa 155 Poverty and policy in South Africa, ed. H. Bho- Developments 2000–2004.” Budget Brief rat and R. Kanbur, 201–31. Pretoria, South 149, Budget Information Service, IDASA, Africa: HSRC Press. Johannesburg. ———. 2006b. “The Targeting of Public Spend- Williams, J.J. 2005. “Community Participation ing on School Education, 1995 and 2000.” and Democratic Practice in Post-Apartheid Perspectives in Education 24 (2): 49–64. South Africa: Rhetoric vs. Reality.” Critical Van der Berg, S., and R. Burger. 2002. “The Sto- Dialogue 2 (1): 19–27. ries Behind the Numbers: An Investigation Wilson, Z., M. Malakoana, and T. Gounden. of Efforts to Deliver Services to the South 2008. “Trusting Consumers: Involving Com- African Poor.” Background study for World munities in Municipal Water Service Deci- Development Report 2004, World Bank, sion Making in Durban, South Africa.” Washington, DC. http://econ.worldbank. Water SA 34 (2): 141–46. http://www.wrc. org/files/28003_van_der_berg.pdf. org.za. ———. 2005. “Social Delivery in South Africa.” Winkler, D. 2005. “Increasing Accountability in In The Next Decade: Perspectives on South Af- Education in Paraná State, Brazil.” Equip2 rica’s Growth and Development, ed. A. Bern- Policy Brief, Academy for Educational De- stein and S. Johnston. Johannesburg: Centre velopment. Washington, DC. http://www. for Development and Enterprise. equip123.net/docs/e2-Increasing%20 ———. 2007. “Teacher Pay in South Africa.” Accountability%20in%20Parana_PB.pdf. Background paper for this report, World World Bank. 2003. World Development Report Bank, Washington, DC. 2004: Making Services Work for Poor People. Van der Berg, S., and M. Louw. 2007. “Lessons Oxford: Oxford University Press. http://www Learnt from SACMEQII: South African -wds.worldbank.org/external/default/ Student Performance in Regional Context.” WDSContentServer/IW3P/IB/2003/10/ Working Paper 16/2007, Department of 07/000090341_20031007150121/Rendered/ Economics, Stellenbosch University, Stellen- PDF/268950PAPER0WDR02004.pdf. bosch, South Africa. ———. 2004a. “Papua New Guinea: Public Van der Berg, S., R. Burger, M. Louw, and D. Yu. Expenditures and Service Delivery.” World 2005. “Trends in Poverty and Inequality since Bank, Washington, DC. http://siteresources. the Political Transition.” Working Paper worldbank.org/INTPUBSERV/Resources/ 1/2005, Department of Economics, Stellen- PNG.PESD.Education.Final(G).jun.2004. bosch University, Stellenbosch, South Africa. pdf. Van der Merwe, K. 2004. “Service Delivery Mech- ———. 2004b. “Local Development Discussion anisms of Local Government.” Unpublished Paper.” Paper presented at the World Bank presentation to the INCA Summer School. Conference on Local Development, Wash- www.inca.co.za/presentations/2004_Karen ington, DC, June. vdMerwe_ServiceDeliveryMechanisms.ppt. ———. 2005a. “The Effectiveness of World Van Donk, M., M. Swilling, E. Pieterse, and S. Bank Support for Community-based and Parnell. 2008. Consolidating Developmental -driven Development: An OED Evaluation.” Local Government: Lessons from the South Af- World Bank, Washington, DC. rican Experience. An Islanda Institute Book ———. 2005b. “Vietnam Development Report: Project. Cape Town: UCT Press. Governance, Rethinking the Role of Govern- Van Rooyen, J. 2008. Personal communication ment in Vietnam.” World Bank, Washington, to K.A. Eales, Department of Water Affairs. DC. December 12. Pretoria, South Africa. ———. 2006. “Making Services Work for the Weber, M. 1978. Economy and Society: An Outline Poor in Indonesia: Focusing on Achieving of Interpretive Sociology. Berkeley: University Results on the Ground.” World Bank, Wash- of California Press. ington, DC. Wildeman, R. 2004. “The School Funding ———. 2008. “What Do We Know About Norms Five Years On: Service Delivery School-based Management?” Education,

156 Accountability in Public Services in South Africa Human Development Network, World Bank, Beyond. A Report on Ghana’s 2000 Public Washington, DC. http://go.worldbank. Expenditure Tracking Survey in the Sec- org/9EX3ZU6O90. tors of Primary Health and Education.” Af- Ye, X., and S. Canagarajah. 2002. “Efficiency rica Region Working Paper 31, World Bank, of Public Expenditure Distribution and Washington, DC.

Accountability in Public Services in South Africa 157 In 1994, the legacy of apartheid presented the new South Africa with enormous challenges: poverty, inequality, and the immense aspirations for greater access to basic public services. The government immediately began addressing these challenges by enshrining constitutional rights to service access, radically reforming economic and sector policies, and funding ambitious service delivery programs. A new governance and administrative framework for racially diverse provinces and municipalities had to be built­—­and the new structure had to progress quickly in delivering the improved services expected by the people. These challenges have grown in the past 16 years, as South Africa’s population has increased with immigration and the HIV/AIDS crisis has spread.

This study attempts to shed light on service delivery capacity and quality challenges in South Africa by analyzing the institutions that “produce” services. “Institutions” are not defined as “organizations” but as the rules that govern the interaction between service providers and beneficiaries. The study examines the accountability relationships among citizen-users, policymakers, and service providers. These relationships can be seen as the three legs of a triangle: the first leg represents accountability between citizen-users and politicians/­policymakers, the second between politicians/­policymakers and service providers, and the third between service providers and citizen-users.