93625 World Bank Group Support to Health Financing World Bank Group Support to Health Financing July 2014 © 2014 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this work is subject to copyright. Because The World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e- mail: pubrights@worldbank.org. ii Contents ABBREVIATIONS .................................................................................................................................. V ACKNOWLEDGMENTS ........................................................................................................................ VI OVERVIEW........................................................................................................................................... VII MANAGEMENT RESPONSE ............................................................................................................. XVI MANAGEMENT ACTION RECORD ................................................................................................... XXI CHAIRPERSON’S SUMMARY: COMMITTEE ON DEVELOPMENT EFFECTIVENESS ................. XXV 1. BANK GROUP SUPPORT TO HEALTH FINANCING ....................................................................... 1 Health Financing Influences Health System Performance ........................................................................................ 2 What Has the World Bank Group Been Doing in Health Financing? ........................................................................ 5 Objective of the Evaluation ..................................................................................................................................... 10 2. REVENUE COLLECTION FOR HEALTH................................................................................ 14 Challenges .............................................................................................................................................................. 14 Bank Group Support to Revenue Collection for Health ........................................................................................... 17 Effectiveness of World Bank Group Support to Revenue Collection ...................................................................... 19 3. POOLING HEALTH FUNDS AND RISKS ............................................................................... 26 Challenges .............................................................................................................................................................. 27 Bank Group Support to Risk Pooling ...................................................................................................................... 29 Effectiveness of World Bank Group Support to Pooling .......................................................................................... 31 4. PURCHASING ......................................................................................................................... 42 Challenges .............................................................................................................................................................. 42 Bank Group Support to Purchasing ........................................................................................................................ 45 Effectiveness of Bank Group Support to Purchasing .............................................................................................. 49 5. FACTORS IN SUCCESSFUL BANK GROUP SUPPORT ...................................................... 60 Government Commitment, Technical and Information Capacity, and Flexibility ..................................................... 61 Depth and Relevance of Analytical Work ................................................................................................................ 63 Capabilities and Collaboration in Health Financing ................................................................................................. 65 Integrating All Health Financing Functions .............................................................................................................. 66 Monitoring and Evaluation in Health Financing Projects ......................................................................................... 68 6. CONCLUSIONS AND RECOMMENDATIONS........................................................................ 71 BOXES Box 1.1. What Constitutes a Health System?.......................................................................................... 2 Box 1.2. Health Financing in World Development Reports...................................................................... 6 iii CONTENTS Box 3.1. Risk Pooling Arrangements ..................................................................................................... 27 Box 3.2. Financial Incentives May Endanger the Sustainability of Risk Pools ....................................... 28 Box 3.3. Bank Analysis Informed Risk Pooling in China and Vietnam ................................................... 35 Box 4.1. What Is the Effect of Performance-Based Payments in OECD Countries? ............................. 44 Box 4.2. Integrated Approach to Purchasing ......................................................................................... 45 Box 4.3. Financing Based on Performance or Results in Argentina and Brazil...................................... 48 TABLES Table 1.1. Health Financing Indicators to Measure Progress toward Outcomes ..................................... 5 Table 2.1. Incentives and Challenges in Revenue Collection ................................................................ 16 Table 3.1. Incentives and Challenges in Different Risk Pooling Arrangements ..................................... 28 Table 4.1. Provider Payment Methods and Related Incentives and Challenges.................................... 43 Table 5.1. Poverty Assessments with Relevant Health Financing Analysis, 2007–2012 ....................... 64 Table 5.2. Health Financing Indicators in HNP Strategy and Health Financing Projects ....................... 69 FIGURES Figure 1.1. Bank Group Support to Health Financing Influences Health System Outcomes ................... 3 Figure 1.2. Number of Bank Operations with Health Financing Activities by Sector Board...................... 8 Figure 1.3. Health Financing Interventions in Bank Projects, FY03–12 ................................................... 9 Figure 1.4. Objectives in HNP Health Financing Operations, FY03–12 ................................................... 9 Figure 2.1. Share of Total Health Revenues in Low- and Middle-Income Countries in 2011 ................. 15 Figure 4.1. Provider Payment Methods Supported by Bank Projects, FY03–12 .................................... 46 APPENDIXES A. Data Sources, Methods, and Limitations B. World Bank Projects with Health Financing Approved in FY03–12 C. Health Financing Projects in Portfolio Review D. IFC Health Financing Approved in FY03–12 E. Summaries of Country Case Studies BIBLIOGRAPHY The appendixes and Bibliography are available online at https://ieg.worldbankgroup.org/Data/reports/health_finance_evaluation.pdf iv Abbreviations AfGH Action for Global Health CBHI community-based health insurance DAH development assistance to health DRG diagnosis-related group GIZ Deutsche Gesellschaft für Internationale Zusammenarbeit GPOBA Global Program on Output-Based Aid HCFP Health Care Fund for the Poor HEF health equity fund HMO health maintenance organization HNP Health, Nutrition, and Population IEG Independent Evaluation Group IFC International Finance Corporation MDG Millennium Development Goal M&E monitoring and evaluation MIP Medical Insurance Program NCMS New Cooperative Medical Scheme NGO nongovernmental organization NHA National Health Account NHIS National Health Insurance System OECD Organisation for Economic Co-operation and Development OOP out of pocket PER Public Expenditure Review PREM Poverty Reduction and Economic Management RBF results-based financing SHI social health insurance SP Social Protection SWAp sectorwide approach USAID U.S. Agency for International Development WDR World Development Report All dollar amounts are in U.S. dollars unless otherwise indicated. v Acknowledgments This evaluation is a product of the Independent Evaluation Group (IEG). The report was prepared by Pia Schneider with contributions from a team of evaluators and analysts. Management oversight was provided by Caroline Heider, Emmanuel Jimenez, and Mark Sundberg. Viktoriya Yevsyeyeva was responsible for all administrative aspects. Barbara Rice and Cheryl Toksoz provided editorial support. Ann Flanagan conducted the analysis of the International Finance Corporation (IFC) portfolio and wrote appendix D. Xue Li, Segen Moges, and Moritz Piatti conducted the World Bank portfolio analysis and reviewed the poverty assessments and impact evaluations. Key-informant interviews with international health financing experts were conducted by Pia Schneider and Judy Twigg. The 16 country case studies were conducted by Manjiri Bhawalkar, Erik Bloom, Bjoern Eckman, Judy Gaubatz, Basil Kavalsky, Xue Li, Segen Moges, Moritz Piatti, Ana Milena Aguilar Rivera, Pia Schneider, Hjalte Sederlof, and Judy Twigg. Ann Flanagan provided private sector input to all country studies. The studies were peer reviewed by Cheryl Cashin and Nancy Pielemeier. Issue notes based on the country studies were prepared by Manjiri Bhawalkar, Cheryl Cashin, and Bjoern Eckman. The team is grateful for comments received from Robert K. Yin. The evaluation greatly benefited from financial support from the Norwegian Agency for Development and Cooperation. Norwegian funds helped support consultants’ time and field visits for country case studies. The Norwegian support greatly enhanced the quality and depth of the data collection and evaluation analysis. IEG management and colleagues provided helpful guidance and comments including Geeta Batra, Kenneth M. Chomitz, Marie Gaarder, Marvin Taylor- Dormond, Stoyan Tenev, and Nicholas York. Many World Bank and IFC managers and staff provided useful comments and support during the evaluation. The team is grateful for the support received by the representatives of the governments of Benin, Cambodia, Ghana, Kenya, Mexico, Nicaragua, Rwanda, Tanzania, Turkey, and Vietnam where field-based country studies were conducted. We also thank the World Bank and IFC offices in these countries for mission support as well as donors who offered their time for interviews. The team is grateful for comments received from Jack Harlow and Bruce Ross-Larson. Peer reviewers were Francois Diop (Abt Associates/U.S. Agency for International Development, Senegal), Randall P. Ellis (Boston University), Pablo Gottret (Social Protection, South Asia Region, World Bank), Joseph Kutzin (World Health Organization, Geneva), and Juan Pablo Uribe (Fundación Santa Fe de Bogotá, Colombia). The many thoughtful and thorough comments from reviewers improved the report and are much appreciated. VI Overview Highlights The way countries finance health care influences how well a health system performs and achieves its expected outcomes, including how equitable and efficient it is. Countries decide how to mobilize revenues from different sources for financing health care, how to pool revenues in public and private insurance and in a national health system with automatic coverage (risk pooling), and how to purchase care from health care providers. The World Bank has implemented health financing activities in 68 countries during FY03–12. Health financing interventions are found in about 40 percent of the Bank’s Health, Nutrition, and Population portfolio. Most projects include interventions on revenue collection from public sources. Almost half of the projects support public health insurance and automatic coverage. More recently, results-based financing (RBF) operations became more prominent. The International Finance Corporation (IFC) delivered a small program in health financing. This is the first evaluation by the Independent Evaluation Group of the World Bank Group’s support to countries trying to address health financing issues. While much remains to be learned about the health benefits, equity in service use and finance, and the financial protection value of health financing reforms supported by the Bank Group, this evaluation has been able to draw the following four major conclusions: First, there have been some notable successes of Bank support to health financing. Bank support was more successful when the Health and Public Sector teams drew on a variety of skills across sectors and where government commitment to reforms was strong. Second, Bank support has helped raise or protect public revenues for health. Equity in pooling increased where the Bank assisted governments in subsidizing compulsory contributions to various health insurance for low-income groups. However, increased pooling did not always lead to pro-poor spending, improved equity in service use, or greater financial protection. Support to reduce user payments was limited. Third, the Bank has increased its focus on activity- or results-based payments supported by RBF projects. Little attention was paid to the impact on costs and broader effects on the public sector. Fourth, an integrated approach that links health financing with public sector reforms is likely to be more effective than single-issue interventions because it builds the institutions that are needed for sustainability. This includes equitable revenue instruments, taking into account the overall public finance situation, moving toward compulsory pooling in insurance and national health systems, focusing on strategic purchasing, and giving attention to adverse effects in a broader public sector context. The linking of health financing to public finance requires strong collaboration across the Bank Group to facilitate the dialogue at all government levels. The evaluation makes five main recommendations: support government commitment and build technical and information capacity; address health financing as a cross-cutting issue at the country level; focus on health financing as a core comparative advantage; integrate all health financing functions; and strengthen monitoring and evaluation in Bank and IFC projects. vii OVERVIEW This evaluation examines support from Bank Group Support to Health the World Bank and the International Financing Finance Corporation (IFC) to health Improving health outcomes and financing through lending, investment, protecting households against the policy dialogue, and analytical work. financial consequences of ill health are Over FY03–12, the World Bank top priorities to reduce poverty and supported health financing reforms sustain growth. However, poor through 188 operations in 68 countries. individuals often forgo care when it is Health financing interventions have needed because they cannot afford to been included in about 40 percent of the pay user fees. They also report worse Bank’s Health, Nutrition, and health outcomes, which can keep them Population (HNP) portfolio. This period trapped in poverty. How health care is saw a marked decline in Bank support financed thus influences who has to pay to interventions related to public how much for care (financial risk revenue collection for health, whereas protection), how much of the health support to purchasing care from funds are spent on different forms of providers increased substantially. The health care, how equitably health IFC delivered a small program in health revenues are collected from public and financing with six investments and nine private sources and distributed (equity advisory services, and funded two in finance), and how effectively health output-based aid operations to health care costs are managed (efficiency). financing. Accompanying Bank lending operations is a large body of analytical The Bank Group’s role in health and advisory work, knowledge financing should be seen in a context of products, technical assistance, and the changing nature of international training programs including the development assistance. The Bank flagship course organized by the World Group’s share of global development Bank Institute. assistance for health is small and has decreased since 1998 from almost 20 The evaluation recognizes that reforms percent to about 6 percent in 2013. in health financing alone are insufficient Partly in response to this trend, in 2007, and that additional investments are the Bank’s health strategy emphasized needed to ensure the supply of health selectivity and a greater focus on the care. But health financing decisions are Bank’s comparative advantage. Because necessary to influence the provision and of the Bank’s involvement in both core use of health care and ensure financial economic as well as sector issues, health protection. finance was seen as a principal focus Four evaluation questions are area, a perception shared by other addressed: development partners. viii OVERVIEW  What is the evidence that Bank Two-thirds of the Bank’s health Group support to revenue financing portfolio has interventions collection for health leads to related to public revenue collection for improved equity in health health. Depending on the country financing and service use, context, the Bank advised governments financial protection, and to increase their budgets for health, efficiency? protect health spending during the  What is the evidence that Bank economic crisis, and introduce excise and IFC support to pooling health taxes to create fiscal space. In countries funds and risks leads to improved with social health insurance, the Bank equity in health financing and supported improvements in tax service use, financial protection, collection administration and the and efficiency? payroll-tax take. It supported subsidies  What is the evidence that Bank to finance contributions to risk pools for Group support to purchasing low-income groups and helped leads to improved equity in governments introduce explicit health financing and service use, targeting of subsidies. In only a financial protection, and scattering of countries did the Bank help efficiency? institutionalize monitoring and  What are the factors in successful evaluation (M&E) to examine the level Bank Group support to health and flow of health funds to providers financing reforms? from public and private sources. Health and Public Sector teams Revenue Collection for Health emphasized strong institutions and monitoring and evaluation through The main challenge for governments in public expenditure reviews and tracking financing their health care systems is surveys. While this type of support has raising revenues efficiently and been decreasing over time, there are equitably to provide individuals with some notable successes. Several lower- essential health services and financial income countries increased their health protection against unpredictable budgets based on Bank advice, although catastrophic financial losses caused by these increases were not always ill health. Where government revenue- sustained. Bank advice also helped raise raising capacity is weak, countries rely tobacco taxes in some middle-income more on revenues from user fees, countries. It also helped increase insurance payments, and development revenues for health by subsidizing assistance. High user payments have contribution payments to various health raised concerns about the financial insurers. consequences for poor households and the negative effect on service use. ix OVERVIEW User payments are the most important service use, and financial protection and revenue source for health sectors in low- health outcomes in a few countries. income countries, and reducing these Knowledge work informed payments has fiscal and equity governments about consolidating implications. Bank advice and a few fragmented risk pools, mainly in operations have supported governments middle-income countries. In some which have tried to lower user countries, the Bank could have taken a payments as a source of revenue, but more active approach with the evidence is limited that Bank support to government to address weaknesses, reduce copayments for patients has including in targeting the poor. IFC- improved service use and financial supported investments and advisory protection. services include health insurance in India and a few African countries. Pooling Health Funds and Risks Reaching the poor requires commitment by governments. Equity in fund pooling With the exception of user payments, all improved where the Bank helped revenues for health are pooled in public subsidize enrollment of the poor. But and private health insurance and in expanded coverage did not always lead central and local government budgets, to pro-poor spending, improved service and then transferred to providers. As use, or financial protection. The reasons countries grow economically, pooled for ineffective coverage include health financing comes to dominate inadequate funding for services covered revenues from user fees. The objective of in the pool, insufficient information pooling of health funds and risk is to about benefits, and inadequate quality ensure financial protection and equity in in service delivery. The Bank helped service use for members. But managing address fragmented pooling, but the health revenues to ensure equitable and topic remains an issue in several efficient pooling is a major challenge. countries and can reduce efficiency. About 40 percent of the Bank’s health There is little evidence of the effect of financing operations supports pooling IFC’s support to health financing on of public funds through automatic improved service use, equity, or financial coverage in national health systems or protection because of the newness of the mandatory health insurance. The Bank projects and scarcity of data. also helped build institutional, management, and technical capacity to Purchasing manage fund pooling at government units and in health insurance. Bank The public policy objective of analytical work discussed the impact of purchasing is for providers to deliver risk pooling on adverse selection, quality care efficiently to individuals x OVERVIEW who need it. Formulating purchasing An RBF program typically supports a policy is challenged by the financial cash payment or non-monetary transfer incentives of various provider payment made to a national or sub-national methods and by the paucity of government, manager, provider, payer, information on providers’ reactions to or consumer of health services after these methods. Payment incentives may predefined results have been attained encourage providers to change the and verified. Payment is conditional on number of services, manage costs, and undertaking measurable actions. RBF improve quality of care, all of which can operations thus directly influence the affect efficiency. Whether these provider payment method in a country. incentives lead to the desired outcome The Bank is conducting an increasing depends on the institutional context for number of impact evaluations on providers and how they react to them. provider payment reforms supported by Most countries have moved to paying RBF projects. providers based on their activities, which has led to increased service use Bank support to purchasing has and higher costs. A few Organisation for strengthened institutions, including Economic Co-operation and management and information systems. Development countries have introduced Availability of care has increased where performance-based payment to countries moved from line-item budgets incentivize better quality and efficiency, to activity- or performance-based though the evidence for better outcomes payments. Limited evidence suggests is slight. Transparent information and that higher public spending on health peer pressure may also affect provider and performance-based payments have behavior. similar effects on service use. Performance payments mainly An increasing share of Bank health increased utilization of services that had financing projects supports higher unit payments and that governments and insurers in providers could more easily control for; purchasing. Some 60 percent of provider they had no impact on other rewarded payment methods supported by the services. Bank include a performance- or results- based component, often on a piloted Where Bank support to purchasing was basis. Most are introduced in health integrated with other health financing systems with automatic coverage in functions and linked to the public low-income settings. These projects use finance context rather than limited to the government as the purchaser. The narrowly defined payment methods, it majority of them run with the support of the Bank’s results-based financing has been relatively effective. This is (RBF) program to support policy and because it addressed broader investment lending. institutional reforms which in turn support sustainability. Bank RBF xi OVERVIEW support to provider payments without  Depth and relevance in measures to reduce user fees and analytical work. improve risk pooling is unlikely to  Capabilities and collaboration. improve equity in service use and  Integration of all health financial protection. This points to the financing functions. need to strengthen the linkage of RBF  Sound monitoring and interventions to the overall financing of evaluation. health systems. Mounting political commitment by Administrative costs and the financial governments has ensured important implications for the payer are major health financing reforms. The Bank has sustainability concerns when helped build technical and information introducing activity- and results-based capacity that is instrumental in payments, which the Bank did not implementing reforms. Yet insufficient sufficiently address. Adverse effects of financial commitment and capacity payment reforms on sector efficiency constraints are still limiting reform were not examined in Bank analysis. sustainability in low-income countries. The Bank did not examine spillover effects on public sector wages. Bank analytical support to health financing and the policy dialogue with These factors have led to uncertainty governments contribute to informing over the financial sustainability of Bank health financing reforms. Monitoring support to results-based payments as and evaluation of Bank support through shown in the country case studies. Most the relevant health financing indicators governments have not assumed is essential to analyze progress toward financing responsibility in their achieving strategic objectives. recurrent budget for the cost of these programs, and even programs The Bank’s 2007 health strategy sees considered effective have not been taken health financing as having a over by governments. comparative advantage for the Bank. Health financing requires a different skill set from that of the general health Factors in Successful Bank Group specialist. To fully use the Bank’s Support capabilities in health financing, collaboration across the new Global Common success factors include: Practices and the IFC is needed.  Government commitment and Synergies in collaboration with other technical and information organizations can be leveraged to raise capacity. the quality of the health financing dialogue. xii OVERVIEW An integrated approach that links health evaluation finds that key elements of the financing to multiple public sector strategy have proven elusive (e.g., better reforms is likely to be more effective integration and M&E). The reasons than single-issue interventions. This is mainly revolve around capabilities and the Bank’s and IFC’s comparative constraints to cross-sector collaboration, advantage as described in the 2007 which are areas for further reflection for health strategy. An integrated approach the Global Practices. Addressing these to health financing would entail efficient would allow the Bank Group to “punch and equitable revenue instruments (tax at (or even above) its weight class” in an and nontax) for health, taking into area where it has a comparative account the overall public finance advantage. situation. It also includes moving toward compulsory pooling, reducing The evaluation showed that the Bank fragmentation in pooling, and focusing and IFC do not have a joint strategic on strategic purchasing. And it approach to health financing -- there are considers potential adverse effects in a no explicitly held positions about the public sector context. Linking health mix of public and private insurance, financing reforms to public sector which population groups they should reforms requires strong collaboration insure, and how to prevent and address between the IFC and the Bank’s , risk selection in multiple-insurance Health, and Fiscal Management teams contexts. The Bank Group did not take to help facilitate the dialogue on health an ideological stance in its work in financing at all government levels. health financing; rather, it worked flexibly in different country contexts. In This evaluation may be missing some line with the Bank’s health strategy, the successful Bank and IFC engagement in Bank did promote a focus on improved health financing because of weak M&E results and performance in health in health projects. Although the HNP facilities by helping governments and strategy stipulates that the Bank insurers change the way they pay monitor how health financing affects providers. equity in service use, risk pooling, and financial protection, this information is The evaluation finds that evidence is rarely collected in health financing thin on the effect of Bank and IFC operations. operations and programs on ultimate outcomes, and much remains to be learned about the health benefits, equity Conclusions and Recommendations in service use and finance, and the financial protection value of public The Bank’s 2007 health strategy remains spending, pooling, and purchasing valid to guide support to health supported by the Bank Group. financing reforms. However, the xiii OVERVIEW The four main conclusions of the institutional reforms to allow evaluation are: providers to react to financial incentives and to demand-side  There have been some notable barriers including user fees, and successes of Bank support to all how to tackle these in a fiscally three health financing functions. sustainable manner. These have occurred when  An integrated approach that Health and other Public Sector links health financing including teams drew on a variety of skills RBF with public sector reforms across sectors and where is likely to be more effective than government commitment to single-issue interventions in reforms was strong. establishing the relevant  Bank support has helped raise or institutions that are needed to protect public revenues for sustain reforms. health against budget cuts during economic crisis. Equity in The evaluation makes five pooling increased where the recommendations to guide the Bank Bank assisted governments in Group’s future work on health subsidizing compulsory financing: contributions to various health insurance plans for low-income Support government commitment and groups. However, increased build technical and information pooling did not always lead to capacity to be able to inform health pro-poor spending, improved priorities and spending by: equity in service use, or greater  Supporting countries through financial protection. Support to capacity building in reduce user payments was standardized monitoring of total limited, and evidence is missing health expenditures (e.g., that it improved equity in National Health Accounts), with service use and financial attention to serving the needs of protection. This type of support the poor; and often lacked the necessary fiscal  Expanding training in client and equity analysis. countries in collaboration with  The Bank has been shifting its local institutions to build focus on health financing to knowledge and technical performance- or results-based capacity through health payments supported by RBF financing learning platforms. projects. Little attention was given to the impact on costs, broader public sector xiv OVERVIEW Address health financing as a cross-  Integrating results-based cutting issue at the country level by: financing interventions with other health financing functions  Ensuring analysis of equity in and the broader public finance health service use and finance, context at the country level to financial protection, and address sustainability and financial sustainability prevent distortions; and consistent with the aim of  Developing a joint strategic promoting Universal Health approach between IFC and the Care coverage. Bank and complementary implementation on the ground Have Global Practices focus on health toward health insurance, financing as a core comparative including mandatory and advantage of the Bank by: voluntary coverage.  Building and expanding Strengthen M&E in Bank and IFC technical capacity among staff projects by: working on health financing in different Global Practices  Improving appropriate M&E (including Health, Macro and frameworks in Bank and IFC Fiscal Management, projects to put in place Governance, Poverty, and Social mechanisms to collect and Protection) to ensure that staff monitor relevant indicators; capacity is adequate to respond and to country demand; and  Monitoring distributional  Having a clearly identified focal indicators, including on point on health financing for the access and outcomes, World Bank Group. consistent with benchmarking and tracking Integrate all health financing functions progress toward Universal by: Health Care coverage. xv Management Response The World Bank Group thanks the Independent Evaluation Group (IEG) for undertaking this evaluation. Management welcomes the opportunity to review and comment on IEG’s report on World Bank Group Support to Health Financing for Improving Health System Performance. This evaluation is timely as we embark on a One World Bank Group model encompassing the Global Practices and Cross Cutting Solution Areas and reevaluating our areas of strengths and space to enhance the performance in health financing. The IEG report generally provides a balanced commentary on most topics regarding the support of the World Bank Group to health financing and covers a wide terrain. Management also commends IEG for the way it engaged with management in a consultative process during the drafting of the report. Broad Concurrence with Conclusions and Recommendations. Management broadly concurs with the conclusions and recommendations of this evaluation. Management welcomes IEG’s call for effective collaboration across the new Global Practices and IFC as well as the need for synergies in collaboration with external partners, as this will be critical in improving future World Bank Group support to health financing. IEG’s recommendation to develop a joint Bank and IFC approach to health financing is also timely and could not be over emphasized. The findings of the evaluation have broad relevance across the organization. COMMENTS SPECIFIC TO WORLD BANK OPERATIONS GENERAL COMMENTS While the IEG report covers a significant amount of ground in terms of World Bank interventions on health financing, it could be more inclusive of the big picture in terms of the context in which interventions in health financing impact on our client countries. For example, management notes that in many emerging economies, while the private sector is not yet bigger than the public, the private sector is growing at a much faster rate. If that trend continues, over the period of a decade, the public sector’s weight will be reduced from perhaps half of the total to a small fraction of the total. Other players in global health have also grown and have a large weight relative to the direct role of the Bank in health financing. The report covers the work of the World Bank Group only for the period of FY03–12. While it could be seen as beyond the scope of the report, it would be helpful to the reader to understand the larger historical perspective and note that the Bank’s health financing work has evolved over time (e.g., advocate for user fees in the 1980s, analytic work on voluntary insurance in the 1990s) to its current state. It could also xvi MANAGEMENT RESPONSE note that client demands have changed over the years (e.g., helping countries in Europe and Central Asia and Latin America and the Caribbean pivot away from tax- based health finance in the 1990s). More context could help explain the current state. The report underemphasizes the Bank’s knowledge program role in supporting health financing, focusing mostly on lending. However, it could more explicitly recognize that much of Bank support to health financing reform is through technical assistance (often as an outcome of analytic and advisory activities, or AAA) rather than through lending. It may not have large monetary value (which is perhaps why work led by the Poverty Reduction and Economic Management Vice Presidency through development policy loans, or DPLs, comes across as so important in the overall support), but this does not make the technical assistance any less important to improving health system performance. The report has an implicit focus on the Bank’s normative view of health financing. It indirectly suggests that one of the Bank's strengths is its recognition of the many different ways to finance health and that there is not a “one-size-fits-all” prescription for clients. The report could benefit from recognizing this strength more explicitly. RESULTS-BASED FINANCING The portrayal of results-based financing (RBF) in the report could better reflect the reality of how Bank RBF projects in low- and middle-income countries operate. The country cases that were chosen are not the most representative. There are many RBF programs in the Health, Nutrition, and Population (HNP) portfolio that are more mature and have been under implementation for some time enabling a more in depth analysis of the impact over time. While the definition of RBF on page xv and paragraph 4.14 is accurate, in the rest of the report RBF is understood mostly as Pay-for-Performance (P4P), used in the OECD countries. The report often uses different yardsticks to evaluate the effectiveness and credibility of RBF. As many Bank-funded impact evaluations of RBF are still ongoing, the report prematurely draws several negative conclusions on RBF and minimizes positive findings. The report describes RBF as a costly intervention and attention is drawn on financial sustainability. In most low-resource settings that the RBF operates in, key issues include poor utilization and low quality of services. Introducing RBF has resulted in large increases in service utilization and provision of quality interventions. By improving productivity and better leveraging the resources already invested, RBF payments form the incremental unit cost of providing the resultant service levels and quality standards. The small incentives used by Argentina’s Plan Nacer, one of the RBF programs discussed in the report, (2 to 4 percent of the provincial public health budget) have successfully leveraged the existing resources for health in the country. Impact evaluation results for the Nacer xvii MANAGEMENT RESPONSE Plan show that the performance incentives are enormously cost-effective: the cost of a Disability Adjusted Life Year (DALY) saved was US$1,115, compared to a gross domestic product per capita of US$6,075. In Rwanda, facilities paid based on performance yielded better results in service provision and quality of care compared to facilities which received equivalent input-based budgets. Government commitment is crucial for sustainability and is shown, among other things, by the financial contributions made by countries as diverse as Cameroon, Zimbabwe, and Armenia committing US$2 million, US$3 million and US$4 million, respectively, from their government budget to RBF. Further, RBF has supported the process of aligning and harmonizing donor inputs with government budgets. Burundi scaled up a virtual pooling system enabling the Government of Burundi, the Bank, and ten other development partners to jointly finance a comprehensive package of services. In Benin, a joint-basket fund supported by the Bank, GAVI and the Global Fund to Fight AIDS, Tuberculosis, and Malaria is managed by the Ministry of Health and used to pay for the RBF results in the health facilities in the country. The report does not recognize the range of benefits that RBF provides. The report focuses on health financing, and portrays RBF merely as a provider payment mechanism. This is inaccurate because RBF is used as a platform to improve providers’ autonomy, strengthen monitoring, increase supervision, boost utilization and quality of care, and overall improve accountability and transparency in the health system. Impact evaluation studies (Basinga et al 2011; Gertler, Paul; Vermeersch, Christel. 2012) have demonstrated that paying for performance increased prenatal and postnatal care quality in addition to boosting service provision and that these effects translated into large and significant improvements in child health outcomes. The core concept of RBF is to promote a results-orientation by linking financing to desired outputs and encouraging entrepreneurial behavior by staff and managers. Further, unlike the typical provider payment methods (capitation, DRG1, case based), RBF payments do not reflect real service production costs, but aim at investing in front line services and modifying behavior, while leveraging existing resources in the health system. The IEG review, by focusing on a relatively narrow subset of country cases2 with an explicit health financing lens and drawing generic lessons, does not recognize the comprehensive nature of the RBF approach and what it has to offer. Moreover, the World Bank Group aims to 1 DRG-Diagnostic Related Groups. 2 The 16 country cases in the IEG study that form the basis of the opinion on RBF include four countries where RBF is implemented with Bank support; out of which two countries (namely Benin and Kenya) were at the early stages of RBF implementation at the time of the study and two were more advanced (namely Rwanda and Afghanistan). xviii MANAGEMENT RESPONSE continue integrating RBF with the other health financing functions it is delivering to create a more comprehensive, systems approach. EVIDENCE AND MONITORING AND EVALUATION Finding evidence for IEG’s research questions is difficult. Most of the health financing reforms supported by the Bank are implemented nationwide, therefore making it difficult to use an experimental design (e.g., randomization) as an outcome identification strategy. Therefore, it should be recognized that the "limited evidence" of the Bank’s support to health financing reforms is also the result of the difficulties in producing rigorous impact evaluations that are implemented nationwide. A more nuanced treatment of the monitoring and evaluation discussion would add value. Paragraph 5.23 states that “evaluation may be missing some successful Bank and IFC engagement in health financing because of weak M&E in health projects.” A more nuanced approach to this could be helpful. In cases where the Bank supported national reforms (e.g., through DPLs or AAA), there are no counterfactual or control groups to assess the impact. Analyses of the effects of health financing can be plagued by endogeneity (e.g., in the case of pooling, this could be the fact that insurance is a choice) that is difficult to overcome statistically and quantify without carefully designed impact evaluation and big data requirements. Evaluating these effects properly would require big financial investments by the Bank and, quite possibly, convincing clients to roll out health financing reforms in an "evaluable” way (e.g., phased or partial) — which may not be desirable for a number of reasons. METHODOLOGY Health financing is one of several building blocks of a health system. The IEG conceptual framework for health financing would benefit from situating it as one of several "building blocks" of a health system (as the World Health Organization conceptualizes it). In several instances, there is a jump from health financing to health outcomes without putting other building blocks -- such as service delivery -- in complementary context. MISSING PRODUCTS FROM THE WORLD BANK Missing references to the Global Expert Team (GET) on Health Financing and Insurance. This was one of the few GETs in the Bank, and it would have been expected to help strengthen the Bank's contributions in health financing and linkages across countries/regions. Its establishment attests to the priority given to xix MANAGEMENT RESPONSE the Bank's role in health financing. The evaluation did not mention this initiative, and did not comment on what mechanisms could have better ensured effective action in each health financing engagement. Narrow representation of lending and non-lending Health Financing tasks. The report appears to have excluded projects where the Bank worked with clients to improve the allocative and technical efficiency of public expenditure, as most Sector- wide Approaches (SWAps) did explicitly (in South Asia, this would include the Bangladesh SWAp). In addition, the evaluation could have included operations which aimed to improve accountability of public expenditure and efficiency through contracting (such as the Uttar Pradesh Health Systems Strengthening Project) and as well as projects where the Bank supported efforts to pursue fiscal decentralization in health (as in the Sri Lanka Health Sector Development Project). COMMENTS SPECIFIC TO THE INTERNATIONAL FINANCE CORPORATION The International Finance Corporation’s (IFC) experience in the health financing space is relatively limited. Over the FY03-12 review period, IFC committed $161 million in this subsector, representing less than 1 percent of IFC total commitment volume across all sectors. According to IEG, the six investments and nine advisory services projects covered in the report already represent 100 percent of IFC interventions during the review period. IFC was more optimistic in health financing when it formulated its health sector strategy in 2002. As indicated in IFC’s Management Response3 to a different but related IEG evaluation of the World Bank Group’s support to Health, Nutrition and Population in 2009, IFC learned that the business case for direct investment in stand- alone private health insurance does not exist to the extent IFC has envisaged it. Going forward, recognizing that in many emerging economies, the private sector is now growing at a much faster rate than the public sector, IFC anticipates greater opportunities for the World Bank Group to support private sector development in health financing. 3 IEG’s 2009 Report Improving Effectiveness and Outcomes for the Poor in Health, Nutrition, and Population: An Evaluation of the World Bank Group Support Since 1997. xx Management Action Record IEG Findings IEG Recommendations Acceptance by Management Response Management Whether Bank support to health 1.Support government commitment WB: Agree The Health, Nutrition, and Population financing reforms is sustained and build technical and information Global Practice (HNP GP) will depends on government capacity to be able to inform health continue and expand support to build commitments and local technical priorities and spending by: capacity to monitor public and capacity. • Supporting countries through private sector spending, and to Technical capacity has facilitated capacity building in standardized prioritize the use of public subsidies understanding for health financing monitoring of total health based on evidence. reforms and can be built in expenditures (e.g., National Health Training in the HOW and the WHAT collaboration with local Institutions. Accounts), with attention to serving of health financing will be expanded the needs of the poor; and through various learning and • Expanding training in client knowledge management vehicles. countries in collaboration with local Institutions to build knowledge and technical capacity through health financing learning platforms. xxi MANAGEMENT ACTION RECORD IEG Findings IEG Recommendations Acceptance by Management Response Management The Bank has produced an array of 2. Address health financing as a WB: Agree A Universal Health Care monitoring analytical work on health financing, cross-cutting issue at the country framework has been co-produced including health financing analysis level by: with the World Health Organization in PERs, poverty assessments, fiscal • Ensuring analysis of equity in to monitor access to essential space analyses, and a growing body health service use and finance, services, the level of financial of impact evaluations. financial protection, and financial protection granted to the population Bank reports do not necessarily sustainability consistent with the and equity in health care. Moving examine the poverty and equity aim of promoting Universal Health forward, attention to financial effect of health financing. Care coverage. sustainability will become an even greater focus of attention. Health financing requires a different 3. Have Global Practices focus on WB: Agree skill set from that of the general health financing as a core health specialist. comparative advantage of the Bank The HNP GP will build staff capacity To fully use its capabilities, the by: in health financing. It will work with Bank Group should use multi-sector • Building and expanding colleagues from IFC Global Industry teams that draw on expertise from technical capacity among staff Groups, other global practices and Health and other sector experts and working on health financing in with the International Monetary Fund work across the new Global different Global Practices (including to systematically discuss and Practices and the IFC. Health, Macro and Fiscal operationalize financial sustainability The Bank’s capabilities in health Management, Governance, Social of health programs, including with financing affect partnerships with Protection) to ensure that staff the use of MTEF instruments. other bodies. capacity is adequate to respond to country demand; and . The HNP GP will have a focal point • Having a clearly identified for health financing to provide WBG focal point for health quality assurance for the World Bank financing. Group’s work in health financing. xxii MANAGEMENT ACTION RECORD IEG Findings IEG Recommendations Acceptance by Management Response Management The Bank can add value by 4. Integrate all health financing stressing its comparative advantage functions by: via linking health financing with • Integrating results-based WB: Agree The second phase of the RBF will public finance and working across financing interventions with other include an explicit emphasis on teams, as suggested in the 2007 health financing functions and the fiscal limits and sustainability. HNP strategy. broader public finance context at the An integrated approach that links country level to address health financing, including RBF, sustainability and prevent with public sector reforms is likely distortions; and WB, IFC: The HNP GP will be working on to be more effective than single- • Developing a joint strategic Agree developing a joint strategic approach issue interventions in establishing approach between IFC and the Bank with IFC toward health insurance. the relevant institutions that are and complementary implementation needed to sustain reforms. on the ground, toward health The health financing collaboration insurance, including mandatory and between the IFC and the Bank has voluntary coverage. been limited so far. xxiii MANAGEMENT ACTION RECORD IEG Findings IEG Recommendations Acceptance by Management Response Management Evidence is scant on the effect of 5. Strengthen M&E in Bank and IFC WB, IFC: Agree A Universal Health Care monitoring Bank and IFC operations and health financing projects by: framework has been co-produced programs on final outcomes. • Improving appropriate M&E with WHO to monitor access to The quality of project M&E is weak frameworks in Bank and IFC essential services, the level of in Bank and IFC health projects. projects to put in place mechanisms financial protection granted to the However, there is an increase in the to collect and monitor relevant population and equity in health care. number of impact evaluations. indicators; and Given IFC's limited exposure in the The HNP strategy stipulates that the health financing space, IFC Bank monitor how health financing Management generally agrees with affects equity in service use, risk the IEG's recommendation to pooling, and financial protection, strengthen M&E frameworks and but this information is rarely will track indicators that measure collected in health financing effectiveness when investing in this operations. type of projects or conducting impact evaluations in the future. The GP will continue to improve • Monitoring distributional WB: Agree the monitoring of health financing indicators, including on access and projects. outcomes, consistent with benchmarking and tracking progress towards Universal Health Care coverage. xxiv Chairperson’s Summary: Committee on Development Effectiveness The Committee on Development Effectiveness (the Committee) met to consider the Independent Evaluation Group (IEG’s) report entitled World Bank Group Support to Health Financing for Improving Health System Performance FY03-FY12 and Draft Management Response. The Committee welcomed the timeliness of the evaluation and was encouraged that management broadly concurred with the report’s recommendations. Members recognized that health is inextricably correlated with the World Bank Group corporate strategies and the mandate of poverty reduction. Members appreciated the timeliness of the evaluation, coinciding with the implementation of Global Practices, Cross Cutting Solution Areas and the “One-WBG” model. Members agreed this was an opportunity for the institution to assess comparative advantages to further enhance its development effectiveness including in the health sector. Members noted that the evaluation covered the period from 2003 to 2012 and that the Bank's approach to the financing of the health sector has evolved significantly. They expressed strong interest in seeing the Global Practices lead to a consistent and cohesive strategy to financing across the Regions, while still appreciating the country context. Members agreed with the importance of a more holistic approach to health financing and service delivery, augmented cross-sector coordination, and increased public-private interface and partnerships, including in particular with the International Finance Corporation (IFC). In this respect, members were encouraged that the Bank and IFC are joining forces to develop a road map for collaboration on health financing as well as on broader assistance in health systems and reforms. They underscored the need for synergies in collaboration with external partners to further improve future World Bank Group support to health financing. Members emphasized the importance of monitoring and evaluation and encouraged management to focus on cost-effective means of evaluation of the relevant policies and operations. Some members commented that the emphasis of health financing should be equally placed on financial sustainability and equity, in order to improve poverty alleviation effects and strengthen equity of health systems. Members noted the difficulty in evaluating the Bank’s knowledge role, including technical assistance, in health financing and welcomed the clarification that there have been various knowledge initiatives underway to improve knowledge development and sharing, including designing a prospective impact evaluation at the inception of Bank projects. Members recognized that the divergent views with respect to results-based xxv MANAGEMENT ACTION RECORD financing (RBF) reflect primarily the fact that in 2012 the RBF portfolio was relatively new. They supported management’s plan to continue integrating RBF with other health financing functions in order to create a more comprehensive approach. Juan Jose Bravo Chairperson xxvi 1. Bank Group Support to Health Financing Highlights  The way that health services are financed affects human welfare because it influences how health systems perform in improving health outcomes, and more directly, it affects the income and consumption of the poor.  Health financing affects health outcomes and poverty through three main functions: revenues collected for health, risk pooling, and purchasing.  Almost half of the World Bank’s health operations support countries in improving the way these three functions perform. The topic is nascent at the International Finance Corporation. Most Bank projects support revenues collected from public sources, but this support has declined over time, whereas Bank support to purchasing has increased substantially.  Evaluating Bank Group support is timely because of its relevance to the institution’s newly articulated poverty goals and its ability to inform the post–Millennium Development Goals 2015 agenda. Also, health finance is a central part of the health strategy implemented by the new Global Practices. The way health systems are financed can directly affect growth and human welfare (Box 1.1). Ill health can lead to financial hardship among low-income households that have to pay fees for health services: they may have to sell assets and incur debts to pay for care, and may fall into poverty or deeper into poverty. As a result, the poor often forgo care when it is needed and report worse health outcomes. Their ill health can keep them trapped in poverty and negatively affect a country’s growth prospects. Improving health outcomes and protecting households against the financial consequences of ill health are top priorities to reduce poverty and sustain growth. Continuous growth depends on a healthy and productive labor force. Good health helps to increase education and the level of human capital. A healthy population also has a fiscal impact as it frees up government resources that can be used for alternative investments. These health outcomes are determined both by household behavior and by the level and quality of health care services. How revenues for health are raised, managed, and then allocated to health care providers may also create different financial incentives for insurers, providers, and consumers, which will affect their behavior and use of resources for service delivery. This affects the type of care patients receive, including the quantity and quality of services and efficiency in service delivery. 1 CHAPTER 1 BANK GROUP SUPPORT TO HEALTH FINANCING Box 1.1. What Constitutes a Health System? There are diverse views as to what should constitute a health system. To date, 41 different conceptual frameworks have been developed to describe health systems, offering diverse perspectives in terms of focus, scope, taxonomy, linguistics, usability, and other features (Hoffman et al. 2012). Common elements are found across the different definitions. These include the need to support health system performance measured by improved equity in access, quality, and efficiency of care, independent of the patients’ diseases. The World Bank has embraced strengthening health systems in its operational work. This approach was articulated in the 2007 Health Nutrition and Population Strategy (World Bank 2007). It says, “Health systems encompass all country activities, organizations, governance arrangements, and resources (public and private) dedicated primarily to improving, maintaining, or restoring the health of individuals and populations and preventing households from falling into poverty (or becoming further impoverished) as a result of illness.” As countries become richer, they make tremendous progress in achieving better health outcomes. Yet substantial inequities in health remain across population groups because health systems in low-income settings often fail to respond to the needs of the population. A major problem is that poor individuals often do not receive needed care because they cannot afford to pay user fees charged by health care providers (Gottret et al. 2008). In addition, patient surveys and citizen scorecards point to public dissatisfaction with low-quality care, informal payments charged by providers to patients, absentee health workers, and unavailability of pharmaceuticals in underfunded health facilities (WHO 2000). Countries are responding to these challenges by ensuring that the way they finance health care is efficient and equitable. The World Bank Group has supported these efforts through a combination of financial assistance, policy advice, and technical assistance. This is the first evaluation by the Independent Evaluation Group (IEG) of the World Bank Group’s support to clients seeking to design their health financing functions. The evaluation is timely because of both its relevance to the Bank’s newly articulated poverty goals and the need to inform the post–Millennium Development Goals 2015 agenda. Also, health finance is a central part of the health strategy to be implemented by the Bank’s new Global Practices. Health Financing Influences Health System Performance Health financing systems consists of three main functions: raising revenues to finance health, pooling health funds and risks, and purchasing health care (Figure 1.1). These functions are designed differently across countries, and no single health financing model is supported by the Bank Group. Instead, the World Bank and the International Finance Corporation (IFC) have taken a flexible approach toward 2 CHAPTER 1 BANK GROUP SUPPORT TO HEALTH FINANCING advice and support to health financing functions through tailored interventions, depending on country context. Figure 1.1. Bank Group Support to Health Financing Influences Health System Outcomes Health Financing System General Labor Taxes / Voluntary User Raising Revenues for Health Taxes Compulsory Premiums payments Contributions (Private) (Private) Central Local Social Health Voluntary (Private) Risk Pooling Government Insurance Health Insurance / Budget (Mandatory CBHI / NGOs (Automatic Pooling) (Voluntary Pooling) Coverage) Purchasing Public / Private Providers Service Use Financial Protection Efficiency Relative to Need and Equity in Finance Source: Drawn from Hsiao (2007), Kutzin (2013), and OECD et al. (2011). Note: CBHI = community-based health insurance; NGOs = nongovernmental organizations. Revenues to finance health systems are raised from public, private, and external sources. Governments collect revenues through direct and indirect taxes to finance public spending, including that for health care. Some of these taxes can be earmarked for health. These domestic revenues for health are then transferred to the health sector in the form of internal transfers, subsidies, and grants to the budget of the Ministry of Health and to lower levels of government (e.g., regions, states, and municipalities), and as subsidies to public or social health insurance (SHI) to finance contribution payments for groups such as informal sector workers. Compulsory contributions to SHI are paid by employees, employers, and the self-employed. Private revenues for health include voluntary premiums paid by households to private insurance and to other prepayment mechanisms, and user payments made by patients (or out-of-pocket payments) directly to public and private providers. Some private providers and pharmacies only receive revenues from user payments. Additional revenues for health are transfers from external sources including bi- and multilateral donors and nongovernmental institutions (OECD et al. 2011). 3 CHAPTER 1 BANK GROUP SUPPORT TO HEALTH FINANCING Pooled financing is money raised through taxes or insurance contributions and premiums that individuals must pay whether or not they need care (Savedoff et al. 2012). Risk pooling is about how to pool financing to share the health risk among pool members. With the exception of user payments, all revenues for health are pooled and then transferred to providers. Depending on the country context, pools can take different forms including the central and local government budget, public and private health insurance, and community-based health insurance, among others. Participation in a pool is compulsory or voluntary. Compulsory pooling of public funds includes (i) automatic coverage of the population (e.g., national health services) and (ii) mandatory participation by law for all or a defined population group in social health insurance, which can be public or private health insurance. Voluntary pooling refers to coverage of individuals at their own discretion through private health insurance and community-based health insurance (Gottret and Schieber 2006; OECD et al. 2011). Governments and health insurers purchase health care benefits on behalf of pool members from public and private providers and nongovernmental organizations (NGOs). Passive purchasing is when providers are simply reimbursed for medical services. Strategic purchasing requires countries to make decisions about how to pay providers and at what price, what benefit package should be purchased for whom, and from which provider. These decisions require information about the behavior of providers and consumers. They also need institutions to govern management in health facilities (Figueras et al. 2005; Gottret and Schieber 2006; Langenbrunner et al. 2009). The way the three health financing functions are designed sets different financial incentives to the government, health insurers, providers, and consumers that will affect the attainment of health system outcomes. It will also influence how much of the health funds are spent on different forms of health care (to ensure service use relative to need); how equitable health revenues are collected from public and private sources and distributed (equity in finance); who is protected against the financial risk of having to pay for care (financial risk protection); and how effectively health care costs are managed (efficiency) (Hsiao 2007). The three main outcomes can be assessed by a set of indicators (Table 1.1). 4 CHAPTER 1 BANK GROUP SUPPORT TO HEALTH FINANCING Table 1.1. Health Financing Indicators to Measure Progress toward Outcomes Outcome Selected indicators Service use relative to Utilization of care relative to need across socioeconomic groups need (equity) Financial protection and Percentage of total health revenues from public funds equity in financing Out-of-pocket spending as percentage of total health revenuesa Percentage of lowest quintile household participating in risk poola Percentage of households with catastrophic health expenditures Percentage of households falling into poverty because of illness Efficiency Percentage of health revenues spent on cost-effective services Percentage of donor funds earmarked Number of risk pools and pool size Quality and productivity in health facilities Source: World Bank (2007). a. Indicator for the World Bank’s 2007 Health, Nutrition, and Population strategy. What Has the World Bank Group Been Doing in Health Financing? World Bank and IFC support to countries’ efforts to improve their health finance systems are guided by clearly articulated strategies. The Bank’s 2007 Health, Nutrition, and Population (HNP) strategy on healthy development sees health financing as a comparative advantage for the Bank because of “its multi-sector nature, its core mandate on sustainable financing, and its fiscal, general economic, and insurance analytical capacity, on regulation, and on demand-side interventions” (World Bank 2007, 51). The strategy aims to prevent poverty as the result of illness by improving financial protection, and strives to improve health outcomes, particularly for the poor and vulnerable. It also aims to improve financial sustainability in health and contribute to sound macroeconomic and fiscal policy, as well as governance, accountability, and transparency in health. The strategy focuses on results and agreements with global partners on a collaborative division of labor in client countries (World Bank 2007). Also, two World Development Reports have brought health financing on the international policy agenda (Box 1.2). To help countries improve their financial protection, the Bank in its HNP strategy commits to provide sound policy advice about the best use of external assistance for health; to remove user fees if the lost revenue can be replaced with alternative resources that reach facilities in a fiscally sustainable manner; and to support effective public financial management systems to document the flow of funds. The Bank also stands to help countries identify options to reduce fragmentation across insurance and public funds, and improve integration with regulatory frameworks for public–private collaboration. Extending risk pooling to the informal sector and the rural population, guided by solid evidence, is a key priority for HNP. The Bank 5 CHAPTER 1 BANK GROUP SUPPORT TO HEALTH FINANCING also commits to support countries in their monitoring and evaluation (M&E), to assess whether arrangements improve financial protection for everyone including for the poor and near-poor (World Bank 2007). Box 1.2. Health Financing in World Development Reports The World Bank’s 1993 World Development Report (WDR) on investing in health argued that countries could reduce their disease burden by, at a minimum, doubling their public spending on cost-effective public health interventions and that external assistance for health should be increased in low-income countries (World Bank 1993). The 2004 WDR on making services work for poor people reasoned that to improve services for the poor, copayments made by patients needed to be retained locally and tied to the performance of providers. They also need to contribute to the income of providers rather than compensate for inadequate public funds. To provide income protection for the poor against the financial risk related to health, the WDR argued that governments should subsidize insurance enrollment or develop specific programs, adjusting subsidies between rich and poor regions in decentralized health systems (World Bank 2004). The HNP strategy strives to improve the financial sustainability of the health sector by helping countries monitor indicators for fiscal sustainability, fiscal space, effects of health financing on labor markets, and country-competitiveness determinants. The Bank commits to help low-income countries address issues of financial sustainability by leveraging household financing to expand risk pooling, attending to volatility in external funding for health, and encouraging governments to adopt pro-poor fiscal policies. In middle-income countries, Bank support aims to help countries dealing with financial sustainability including systemic efficiency problems and the fiscal and labor implications of SHI (World Bank 2007). IFC’s health strategy seeks to contribute to institutional capacity building in client countries. It aims to promote efficiency and innovation within health, while improving health security and expanding financial protection against the impoverishing effects of ill health (IFC 2002). In 2007 the IFC outlined a strategy for engaging in the health sector in Africa (IFC 2007). The 2007 strategy called for combined investment and Advisory Service operations, to assist governments with developing appropriate regulatory frameworks in order to support growth in the private health sector; to increase access to capital, promote quality standards for service delivery, and support risk pooling mechanisms (IFC 2007). Both strategies expected a growing portfolio to focus on private health insurance and to support supplementary insurance that covers services excluded from mandatory coverage. The Bank and IFC do not have a joint strategic approach to health financing. There is no joint strategic direction about the mix of public and private insurance, which population groups they should insure, and how to prevent and address risk- 6 CHAPTER 1 BANK GROUP SUPPORT TO HEALTH FINANCING selection in multiple-insurance contexts. Nor have the two institutions decided on whether and how to separate the financing and the provision of care. HOW HAS THE WORLD BANK GROUP OPERATIONALIZED THESE STRATEGIES? The Bank Group’s role in health must be seen in a context of the changing nature of international development assistance. Its share of total development assistance for health is small and has decreased since 1998 from almost 20 percent to about 6 percent in 2013 (appendix Figure B.3). The largest areas of growth in donor assistance have been in health related global programs (e.g., the GAVI Alliance; Global Fund to Fight AIDS, Tuberculosis, and Malaria; and U.S. President’s Emergency Plan for AIDS Relief) targeted to diseases but typically not addressing health finance and system requirements (IEG 2011a). Partly in response to this trend, in 2007, the Bank’s health strategy emphasized selectivity and a greater focus on the Bank’s comparative advantage. This evaluation conducted a detailed review of the World Bank and IFC support to health financing through lending, investment, policy dialogue, and analytical work. Bank operations were included if they supported any interventions that are part of the health financing functions (appendixes A, B, and C). Similarly, IFC operations are included if they support private or public health insurers or health maintenance organizations (HMOs) (appendix D). Bank support to health financing is managed by the Health, Nutrition, and Population, Social Protection (SP), and Poverty Reduction and Economic Management (PREM) Sector Boards. In addition, the Regions and the HNP anchor produce a large number of knowledge products. Health financing is a relatively new topic for the IFC, which offers advisory services and investments including loans and equity to private, for-profit insurance companies. Between FY03 and FY12 the IFC made six investments, including two investments in private health insurance, two in Nigeria’s largest integrated HMO-provider network, and two in health-specific private equity funds, which have invested in insurance companies and HMOs. IFC also provided nine advisory services and funded two output-based aid operations to health financing (appendix Table D.1). Advisory services aim at generating knowledge and advising governments as well as private and public insurers (appendix D). Most IFC projects in health financing aim to improve the financial protection of underserved populations, expand access to private insurance covering the mandatory package, and improve access to care among the poor. The business case for direct investment in stand-alone private health insurance does not exist to the extent envisaged in the 2002 IFC health strategy. Thus, the operational execution of IFC’s strategy has emphasized 7 CHAPTER 1 BANK GROUP SUPPORT TO HEALTH FINANCING increasing health care access through direct investments in health care networks, centers of excellence, and wholesaling (see appendix D). In the same period, the Bank provided 188 loans that included health financing interventions (appendixes A and C). These loans were implemented in 68 countries through development policy operations (56 percent) and investment loans1 (44 percent) (appendix B). The number of operations with health financing peaked in 2006 and then in 2010 when a large number of multisector development policy operations provided fast-disbursing financial support to ensure funding for social sectors during the economic crisis (Figure 1.2). About 40 percent of the Bank’s portfolio includes health financing. The share of health financing operations managed by HNP and the number of newly approved projects, have decreased as more health financing operations are implemented through development policy operations managed by PREM. Most health operations with health financing activities fund a variety of interventions, including infrastructure costs, but the actual lending amount for health financing activities is unknown. Figure 1.2. Number of Bank Operations with Health Financing Activities by Sector Board 30 25 Number of operations 20 15 PREM SP 10 HNP 5 0 FY03 FY06 FY07 FY08 FY09 FY10 FY11 FY12 FY04 FY05 Financial year of project approval Note: HNP = Health, Nutrition, and Population; PREM = Poverty Reduction and Economic Management; SP = Social Protection. Most health financing projects support revenue collection from public sources (Figure 1.3). However, there has been a marked decline in this type of Bank support, whereas support to purchasing has increased substantially. Almost half of the projects support compulsory risk pooling, but few Bank operations focus on revenues from private sources, including user payments. The distribution of project interventions and objectives by Region and sector are presented in appendix B. 8 CHAPTER 1 BANK GROUP SUPPORT TO HEALTH FINANCING Figure 1.3. Health Financing Interventions in Bank Projects, FY03–12 63 Revenue public sources 79 16 Revenue private sources 14 FY08-12 FY03-07 41 Compulsory risk pooling 35 32 Purchasing 17 0 20 40 60 80 100 Percentage of projects with interventions Note: The total for FY03–07 is 96 projects; for FY08–12, 92 projects. Most Bank projects in this evaluation aim to contribute to one of the four strategic objectives of the HNP strategy, namely, improving the health status of a population (Figure 1.4). Only a few health financing projects have a financial protection objective. Less than 20 percent of projects aim to improve equity in access, with access often defined as increased utilization or coverage of care. About one-fourth explicitly target the poor in their objectives (appendix B).2 Figure 1.4. Objectives in HNP Health Financing Operations, FY03–12 Intermediate Outcomes Outcomes Financial protection Final Health outcome Equity Sustainability Efficiency Quality Access/utilization/coverage Target Poor 0 20 40 60 80 Percentage of HNP projects with objectives Note: The IEG project portfolio review is based on 78 HNP operations with health financing interventions. 9 CHAPTER 1 BANK GROUP SUPPORT TO HEALTH FINANCING Bank lending operations are accompanied by a large body of analytical and advisory work, knowledge products, technical assistance, and training programs including the flagship course which is organized by the World Bank Institute (WBI). From 1997 to 2008 the WBI and its collaborating partners delivered 314 short-term training events on health sector reform and sustainable financing to 19,400 participants from 51 countries (Shaw and Samaha 2009). In FY03–12, the World Bank undertook analysis and promoted knowledge sharing on health financing reforms through 98 public expenditure reviews, at least 10 public expenditure tracking surveys,3 20 poverty assessments, about 70 economic and sector work activities, 8 fiscal space studies, and a small but growing number of impact evaluations (appendixes A and B). A large number of health financing workshops have been organized in the Regions,4 some of them in collaboration with the WBI and with initiatives such as the South–South Network and the Joint Learning Network. In addition, the HNP anchor supports health financing, including through the Results-Based Financing (RBF) initiative5 and the Universal Health Coverage initiative, which has conducted 25 country case studies.6 The Bank’s Development Research Group launched the ADePT health module software in 2011 which allows users to produce standard tables for health equity analysis.7 Objective of the Evaluation The evaluation’s objective is to examine the effectiveness of World Bank Group support to health financing in improving health system performance as measured by improved equity in service use, financial protection, and efficiency. The evaluation applies the health financing framework illustrated in Figure 1.1. The methodology is described in appendix A. This is the first time that IEG has evaluated the effectiveness of Bank and IFC support to health financing. IEG’s 2009 HNP evaluation analyzed IFC and Bank portfolio performance in achieving health outcomes for the poor, conducted analysis of communicable diseases, and examined health in transport and water and sanitation operations. This evaluation will not examine lending to finance health care delivery including human resources, equipment, pharmaceuticals, and construction of facilities, nor the procurement of these products. Some of these aspects of health systems improvements were evaluated previously (IEG 2009). Also, as procurement in the health sector is not part of health financing, it will not be addressed (IEG 2014). Social safety nets through conditional cash transfers were evaluated previously (IEG 2011b). 10 CHAPTER 1 BANK GROUP SUPPORT TO HEALTH FINANCING The evaluation recognizes that reforms in health financing only are not enough to improve quality of care, ensure utilization according to need, or remove barriers in the use of care, and that additional investments are needed to assure health care. However, financing reforms are necessary to influence the provision of health care. Other factors also influence the performance of health systems and outcomes, including economic growth, demographic and epidemiological changes, new medical technologies, and the environment. However, examining these factors is beyond the scope of this health-financing-focused evaluation. The evaluation addresses four questions, each of which is the main topic of the next four chapters:  What is the evidence that Bank Group support to revenue collection for health leads to improved equity in health financing and service use, financial protection, and efficiency?  What is the evidence that Bank and IFC support to pooling of funds and health risks leads to improved equity in health financing and service use, financial protection, and efficiency?  What is the evidence that Bank Group support to purchasing leads to improved equity in health financing and service use, financial protection, and efficiency?  What are the factors in successful Bank Group support to health financing reforms? The evaluation offers lessons to inform future lending and knowledge activities. This evaluation covers FY03–12 and draws on several sources (appendix A). They include a review of 188 closed and ongoing Bank operations (appendixes B and C), a review of Bank impact evaluations, 43 poverty assessments, 8 IEG project performance reports, semi-structured key informant interviews with 25 international health financing experts, and an electronic survey of Bank staff working in HNP. All IFC health-related advisory services and investment operations were reviewed (appendix D). The evaluation team also carried out 16 new country case studies, summarized in appendix E. Country case studies review reimbursable advisory services where relevant. Two caveats stand out. Evidence on the achievements of the Bank and IFC project portfolio has been difficult to obtain, mainly because project M&E frameworks do not collect the relevant indicators (appendix A). Further, limitations in project data severely constrain the ability to assess the contribution of Bank and IFC support to health financing (chapter 5). 11 CHAPTER 1 BANK GROUP SUPPORT TO HEALTH FINANCING References Gottret, Pablo, and George Schieber. 2006. Health Financing Revisited: A Practitioner’s Guide. Washington, DC: World Bank. Gottret, Pablo, George Schieber, and Hugh R. Waters. 2008. Good Practices in Health Financing: Lessons from Reforms in Low- and Middle- Income Countries. Washington, DC: World Bank. Hoffman, Steven J., John-Arne Rottingen, Sara Bennett, John N. Lavis, Jennifer S. Edge, and Julio Frenk. 2012. “Conceptual Issues Related to Health Systems Research to Inform a WHO Global Strategy on Health Systems Research.” Background Working Paper, World Bank, Washington, DC. Hsiao, William C. 2007. “Why Is a Systematic View of Health Financing Necessary?” Health Affairs 26 (4): 950–961. IEG (Independent Evaluation Group). 2009. Improving Effectiveness and Outcomes for the Poor in Health, Nutrition, and Population. Washington, DC: World Bank. ______. 2011a. The Global Fund to Fight AIDS, Tuberculosis, and Malaria, and the World Bank’s Engagement with the Global Fund. 2 vols. Washington, DC: World Bank. ______. 2011b. Social Safety Nets: An Evaluation of World Bank Support, 2000 –2010. Washington, DC: World Bank. ______. 2014. The World Bank and Public Procurement—An Independent Evaluation. Volume I: Building Procurement Capacity and Systems. Washington, DC: World Bank. IFC (International Finance Corporation). 2002. Investing in Private Health Care: Strategic Directions for IFC. Washington, DC: IFC. ______. 2007. Health Strategy for Sub-Saharan Africa. Washington, DC: IFC. IMF (International Monetary Fund). 2011. “Revenue Mobilization in Developing Countries.” Fiscal Affairs Department, Washington, DC. https://www.imf.org/external/np/pp/eng/2011/030811.pdf Kutzin, Joseph. 2013. “Health Financing for Universal Coverage and Health System Performance: Concepts and Implications for Policy.” Bull World Health Organ. 91: 602–611. OECD (Organisation for Economic Co-operation and Development)/Eurostat/WHO (World Health Organization). 2011. A System of Health Accounts: 2011 Edition. Paris: OECD Publishing. Savedoff, William D., David de Ferranti, Amy L. Smith, and Victoria Fan. 2012. “Political and Economic Aspects of the Transition to Universal Health Coverage. ” Lancet 380 (9845): 924–32. Shaw, R. Paul, and Hadean Samaha. 2009. Building Capacity for Health System Strengthening: A Strategy that Works: The World Bank Institute’s Flagship Program on Health Sector Reform and Sustainable Financing, 1997–2008. Washington, DC: World Bank Institute. WHO (World Health Organization). 2000. The World Health Report 2000. Health Systems: Improving Performance. Geneva: WHO. World Bank. 1993. World Development Report 1993: Investing in Health. Washington, DC: World Bank. ______. 2004. World Development Report 2004: Making Services Work for Poor People. Washington, DC: World Bank. ______. 2007. Healthy Development: The World Bank Strategy for Health, Nutrition, and Population Results. Washington, DC: World Bank. 12 CHAPTER 1 BANK GROUP SUPPORT TO HEALTH FINANCING 1 Investment lending to the public and private sector finances project costs such as goods, infrastructure, and consultancies. Development policy operations are nonearmarked loans, credits, or grants that support the country’s economic and sector policies and institutions; they finance transition costs, institutional strengthening, and consensus building on reforms. Using its RBF experience, the Bank introduced a new lending instrument, Program-for-Results financing, in January 2012, which supports government programs and links the disbursement of funds directly to the delivery of defined results, with a focus on strengthening institutions. The Bank has approved one health project under Program-for-Results financing, namely the Ethiopia Health Millennium Goals Program for Results (P123531), approved in February 2013. It is not included in this evaluation. Public Financial Management for Results Program in Mozambique (P124615) includes public financial management, health and education and is scheduled for approval in June 2014. 2Health, Nutrition, and Population (HNP) operations tend to target the vulnerable in their objectives. But as these projects often include disease-specific components, “vulnerability” could be interpreted as vulnerable to higher infection risk and not necessarily vulnerable to weaker socioeconomic status. 3 The Bank’s business warehouse database does not have a special code to identify public expenditure tracking surveys. 4For example: South Asia Regional High Level Forum on Health Financing in June 2010, http://web.worldbank.org/WBSITE/EXTERNAL/TOPICS/EXTHEALTHNUTRITIONANDPOPUL ATION/0,,contentMDK:22669883~menuPK:282516~pagePK:148956~piPK:216618~theSitePK:282511, 00.html. 5 For more information, visit http://www.rbfhealth.org/. 6For more information, visit http://www.worldbank.org/en/topic/health/publication/universal- health-coverage-study-series. 7 ADePT is a software platform that uses micro-level data from various types of surveys, such as household budget, demographic and health, and labor force, to automate economic analysis. The ADePT health module allows users to produce most tables that have become standard in applied health equity analysis with a very low margin of error, and covers inequalities and inequities in health and health care utilization, benefit incidence, financial protection, and equity in health financing. 13 2. Revenue Collection for Health Highlights  The challenges governments face in revenue collection include how to raise revenues efficiently and equitably to finance health care. In low-income countries, direct payments made by patients to providers are the main source of revenue, raising concerns about access to care for the poor.  The Bank did not take an ideological stance in its work in revenue collection for health; rather it worked within the different country contexts. Bank advice focused on increasing the health budget in low-income countries. In middle-income countries, the Bank recommended managing the level of public spending and subsidizing insurance enrollment. Some timely advice on a greater role for alcohol and tobacco taxes has been given but this is very limited. The Bank gave limited attention to user payments through lending operations. In few countries did it help institutionalize monitoring and evaluation to examine the effect of health financing.  There have been some notable successes. Bank support has helped raise domestic revenues for health and subsidize contributions to risk pools for low-income groups. Support to reduce user payments lacked the necessary fiscal and equity analysis, and evidence is missing that it has improved service use and financial protection.  Bank support was more successful with strong government commitment at both the economy-wide and sector levels and when Bank staff drew on a variety of skills across sectors to engage government. Revenues for health are collected from public and private sources and allocated to health care providers. Governments face challenges in raising revenues efficiently and equitably. Chapter 2 introduces these challenges, describes how the Bank supported countries in addressing them, and evaluates the effect of this support. Challenges While some countries set targets for public revenues for health, such as the Abuja target of allocating at least 15 percent of the annual government budget to health,1 there is no consensus on how much revenue governments should allocate to health. The reasons for this diversity reflect different economic circumstances and the range of social contracts that governments have with their citizens for ideological or historical reasons. The economic rationale for devoting public revenues to health are (i) to correct for market failures (e.g., private markets do not work well when consumers and providers have different levels of information regarding the appropriate type and amount of care to purchase); (ii) to ensure that public goods are correctly funded (e.g., immunization may be undervalued if the benefits flow to 14 CHAPTER 2 REVENUE COLLECTION FOR HEALTH society at large); and (iii) to ensure that the poor and other disadvantaged groups are not excluded (to meet equity objectives). The concern in many developing countries is that the very low amount that many governments now devote to health is too low to fund these necessary functions. A government’s revenue-raising capacity is affected by factors such as the country’s economic development, institutional constraints, level of formalization of the labor market, and tax administration capacity. Where these are weak, countries rely more on revenues from private and external sources for health. Private revenue—mainly user payments on fees charged by providers—amounts to 62 percent of total health funds in low-income countries (Figure 2.1). User fees have raised concerns about the financial consequences for poor households and the negative effect on health service use (Table 2.1). As countries grow economically, public revenue for health comes to predominate. Figure 2.1. Share of Total Health Revenues in Low- and Middle-Income Countries in 2011 a. Low-income countries b. Middle-income countries External 1% External 26% Private 48% Public Private Public 12% 51% 62% Source: World Development Indicators. Note: Private = user payments. Voluntary insurance is negligible and is not shown. Recent syntheses of impact evaluations find that increasing public spending and lowering payments for patients positively affects health outcomes. Using a large panel dataset at the country level, with annual data for 14 years (1995–2008), Moreno-Serra and Smith (2011) applied a two-step instrumental variables approach that directly estimates the reverse causal effects of mortality on coverage indicators. They found that higher public spending on health leads to better population outcomes, measured either by under-five or adult mortality rates.2 A synthesis report of 16 impact evaluations found that introducing user fees decreases utilization of care, whereas removing them sharply increases utilization of curative services (Lagarde and Palmer 2011). A systematic review of 20 impact evaluations of user fees for maternal health services found that the removal of such fees contributes 15 CHAPTER 2 REVENUE COLLECTION FOR HEALTH to increased facility delivery but has no clear impact on health outcomes (Dzakpasu et al. 2013). Table 2.1. Incentives and Challenges in Revenue Collection Revenues Types Incentives Challenges General taxes Individuals underreport income to pay lower Low tax ratios; inadequate taxes; governments allocate funds to other levels of public revenues sectors for political reasons allocated to health in low- income countries Labor taxes and Individuals reduce work in formal sectors; Increased informality; increased contributions to employers underreport number of employees revenue collection costs in tax health insurance and their salaries to tax authorities authority Voluntary premium Individuals hide true health status to pay Few people can pay high paid by individuala lower premiums premium; financial sustainability User payments by Poor seek care with lowest-price provider Poor report lower utilization of patients (e.g., pharmacies) care and worse health External sources Governments allocate funds for political Rigidity because of fragmented from donors reasons and to priority diseases and earmarked funding a. Private insurers charge premiums that reflect the risk of illness for an individual or a group of individuals. However, many developing countries struggle to mobilize adequate and stable resources because they report low tax ratios, with tax revenues often below 15 percent of gross domestic product (GDP) (IMF 2011). Thus these governments have little room to increase spending on health through domestic revenues. Still, the financing of increasing demand for costlier treatment for noncommunicable diseases (such as diabetes) and the treatment of infectious diseases put a heavy strain on their budgets. In response, governments try to manage public spending on health by setting caps on sector spending, prioritizing spending within the sector, and using central oversight (IMF 2011). Governments have introduced taxes on wages and alcohol and tobacco to raise additional revenues for health, which can have efficiency and welfare implications.3 If governments impose taxes on wages to finance insurance enrollment, this may affect efficiency if it leads to a reduction in the quantity of hours worked and increases informality (Table 2.1). Indirect taxes levied on goods with externalities, such as alcohol and tobacco, can increase efficiency because they aim to influence individual behavior, reduce negative externalities on others, and subsequently curb the incidence of costly diseases caused by alcohol and tobacco consumption (Begg et al. 2000). Although excise taxes on alcohol and tobacco are regressive, they have a welfare effect if the poor benefit disproportionately more than the rich in health. Nor do excise taxes have adverse effects on labor and capital (IMF 2011). 16 CHAPTER 2 REVENUE COLLECTION FOR HEALTH Only a small share of total health revenue comes from voluntary premiums paid to private insurance. Outside the United States, revenues from voluntary health insurance contribute less than 15 percent of total health revenues in Organisation for Economic Co-operation and Development (OECD) countries (OECD 2013). In developing countries, voluntary private health insurance raises a negligible share of total health revenues (Gottret and Schieber 2006). Few countries have community- based health insurance (CBHI), which is financed by individual contributions and in some countries subsidized by government and donor funds. Despite the low share of government spending in low-income countries, governments still have much influence as most external resources are routed through them to finance the public health sector. External funds can, however, contribute to fragmentation in financing and service delivery, especially if they are earmarked for specific diseases, and draw away health workers and other resources from general care (Table 2.1). External funding can also be driven by external priorities, introducing both rigidity and instability into a country’s health sector funding. Bank Group Support to Revenue Collection for Health The World Bank Group has tried to help countries address the above challenges. Two-thirds of the Bank’s health financing portfolio includes interventions related to public revenue collection for health; however, this type of Bank support has been decreasing over time. Development policy operations are almost twice as likely as investment lending projects to advise governments on public revenues (appendix Table B.4). The Bank’s approach has been to help countries raise revenues to address market failure, public goods, and equity objectives. The Bank assisted countries in raising adequate levels of revenues to finance the government health budget and health insurance. It advised governments on revenues raised in the form of labor taxes and other contributions to social health insurance paid by employees, employers, and the self-employed, and on user payments made by patients to providers. ANALYTICAL WORK Multisector Bank teams produced analytical work, including Public Expenditure Reviews (PERs), tracking surveys, and fiscal space analysis, that informed governments and other donors about the level of public revenues for health and the allocation of funds within the sector (appendix Table A.9). Bank teams conducted 98 and at least 10 Public Expenditure Tracking Surveys since 2006 (appendix Table 17 CHAPTER 2 REVENUE COLLECTION FOR HEALTH A.10). However, the number of PERs with a health chapter has fluctuated and decreased over time to less than 10 reviews per year. Medium-term expenditure frameworks were supported by the World Bank in Madagascar, Nepal, and Rwanda, and helped inform governments and donors about health expenditure planning. In 2010 the Bank developed a conceptual framework for assessing fiscal space for health (Tandon and Cashin 2010). Since 2009 the Bank has conducted about eight fiscal space analyses to advice governments on how to feasibly increase revenues for health in a way consistent with the country’s macroeconomic fundamentals. More recently in 2013, the Bank produced a series of macro-fiscal context and health financing fact sheets for all Regions (Pande et al. 2013). DOMESTIC REVENUES FOR HEALTH FROM GENERAL TAXES The Bank tailors its advice to the country context. In low-income settings the Bank advised governments to increase their budgets for health, often with the support of Poverty Reduction Support Credits. In some countries tobacco taxation is earmarked for health and other social spending. Bank analytical work advised on using tobacco taxes to create fiscal space for health, mainly in middle-income countries, including China (2003), Estonia (2004), Morocco (2004), Indonesia (2005), Brazil (2007), The Gambia (2012), and the Philippines (2012) as well as the Southeast Asia Region (2004). COMPULSORY CONTRIBUTIONS AND VOLUNTARY PREMIUMS TO INSURANCE In a few European countries the Bank advised on labor taxes and on domestic revenue financing for social health insurance mainly through development policy operations. Where labor tax rates were already high, the Bank warned about adverse effects for the labor market and for informal workers. The World Bank and the International Finance Corporation (IFC) did not advise on the level of premiums paid to voluntary private health insurance. USER PAYMENTS Only 14 percent of Bank health financing projects advised governments, (mainly in the Africa and Europe and Central Asia Regions), on the level of user payments (appendix Table B.5). The Bank through development policy operations recommended introducing copayments with exemptions for lower-income groups in Romania and for preventive services in Burkina Faso. Analytical work by the Bank on under-the-table payments made by patients to providers (Cherecheş et al. 2013) has not been followed up in projects, even though the measurement of progress toward the objective of improved governance, accountability, and transparency is an indicator in the Health, Nutrition, and Population (HNP) strategy. 18 CHAPTER 2 REVENUE COLLECTION FOR HEALTH Effectiveness of World Bank Group Support to Revenue Collection This section focuses on how Bank advice has affected institution building and the level of revenues for government health budgets and social insurance. It presents evidence on Bank support to nonpooled funding in the form of user payments and how they have affected service use. Increased domestic revenues and subsidized contribution payments to social insurance mean that more pooled public funds are available for health. The effects associated with pooled financing are presented in chapter 3. STRENGTHEN REVENUE RAISING INSTITUTIONS In its analytical work, the Bank emphasized that strong institutions are crucial in ensuring that higher public spending positively affects the provision of care. Several Bank studies find a correlation between public health spending and utilization of care when institutions are strong, and Bank teams found that public sector spending improves health indicators in low-income and transition countries, mainly those with good governance systems (Gupta et al. 2002; Baldacci et al. 2008). In 2009, during the financial crisis, the Bank’s Europe and Central Asia Region reiterated the importance of good governance in revenue management in its Knowledge Briefs for client countries and staff. The Independent Evaluation Group (IEG) found that the Bank through lending and policy dialogue helped governments build institutional and technical capacity, and in some countries, Bank teams worked well with government staff. Institution building took place in Argentina where the Bank helped re-establish the ministerial and provincial health committee to coordinate health financing decision in the country (IEG 2011). The Bank supported technical capacity building through the introduction of National Health Accounts in governments to track the flow of funds, mainly in middle-income countries in the Europe and Central Asia Region (including Albania, Armenia, Kosovo, Moldova, Serbia, Tajikistan, and Uzbekistan) as well as in Mauritania and Vietnam. Information produced by government National Health Accounts was used by Bank public expenditure review teams, other donors and the government in health expenditure planning. While health accounts proved to be useful and informative, this support did not always succeed in institutionalizing the health account function within Ministries of Health (IEG 2014). Because many institutions are involved, coordination is important. IEG found that in Tanzania, the Ministry of Health and the Bank produced their own individual PER in 2011, drawing from different datasets and thus producing different results. The Bank’s PER was distributed but never published officially. As the discrepancies between the two PERs were not reconciled, the Ministry of Health uses its own 19 CHAPTER 2 REVENUE COLLECTION FOR HEALTH report. Collaboration between the Bank and government teams could have helped ensure that health financing analysis is coordinated and institutionalized in Ministries of Health. INCREASED AND PROTECTED HEALTH BUDGETS With the support of the Bank’s development policy operations and policy dialogue, health budgets increased during the loan period in several lower-income countries (for example, Afghanistan, Albania, Bolivia, Burkina Faso, Cape Verde, El Salvador, Lao People’s Democratic Republic, Mali, and Niger). However, these budget increases were not always sustained. IEG found in Tanzania, the Bank worked closely with other donors to ensure the government would maintain the share of public funding for health. Donors and the Bank decided to disburse earmarked funds to local government health budgets (in a sectorwide approach) and not move to general budget support when concerns were raised that this change would lead to a decrease in overall public spending for health. Despite these efforts, government spending on health decreased from 16 percent in 2007 to 11 percent of total government expenditures in 2011. The reasons for this decrease included a shift in government priorities from social sectors to infrastructure as outlined in the 2010 National Strategy for Growth and Reduction of Poverty, commonly known as MKUKUTA. Similarly, in the Kyrgyz Republic, Bank policy lending in 2002 supported the government in increasing its budgetary share for health to finance health insurance coverage for pensioners and unemployed persons and implement a categorical grant formula for health financed from the central government with the goal of decreasing the share of user payments among the poor. Until 2006, government spending on health increased steadily but then declined again to similar levels as in early 2000 because of increased government priorities for other sectors, including education (IEG 2008). The decline in budget financing for health was addressed under a Bank-supported follow-up operation (sectorwide approach), and the government implemented a set of rules governing the allocation and execution of public funds to the health sector. Subsequently, spending on health increased from 10.3 percent of total government spending in 2005 to 13 percent by 2012. Thus, competing government priorities play a role in raising revenue for health. Bank advice through lending and technical assistance helped raise additional taxes. One Bank policy operation (Romania) advised an increase in tobacco taxes in 2009; however, as Romania follows European Union rules,4 this increase would have happened without the Bank’s input. In the Philippines, the Bank responded quickly with a multisector team to government requests to help it get the “sin tax on tobacco” through a reluctant Congress. The government of the Philippines reports substantial revenue increases from the tax, which will translate into higher funding for health 20 CHAPTER 2 REVENUE COLLECTION FOR HEALTH programs (AER 2014). While some countries have increased tobacco taxation, recent studies from Brazil (Euromonitor International 2013) and Indonesia (Nasrudin et al. 2013) suggest that the tax rate is not high enough. There is scope for the Bank to address tobacco taxation in low-income countries. At the same time, impoverishing effects caused by regressive taxation need to be addressed. Bank advice on managing or protecting public revenues for health was informed by analytical work and implemented in close collaboration between the Health and Public Sector teams. IEG’s review of project completion reports found that in Bosnia and Herzegovina, Colombia, and Serbia the Bank’s Public Sector and Health teams (mainly through development policy operations) supported improvements in the tax collection from employers and employees which increased revenue transfers from the tax authority to social health insurance. In Eastern Europe, including in Croatia and Turkey, the Bank recommended budgetary caps on spending to manage spending growth. Bank policy lending protected the level of budgetary spending from cuts in Latvia and Tajikistan during the financial crisis, and thereby helped the government implement recommendations from recent PERs. During the economic crisis in 2010, Bank policy lending advised the Latvian government to subsidize health payments for low-income households and raise the number of nurses in health facilities to accommodate increased patient demand. In light of fiscal austerity in Argentina, Bank lending and policy advice ensured that basic and cost-effective health programs were protected and financed by the government, including the availability of reproductive health care services for low-income groups in public facilities (IEG 2011). In these countries, the Bank’s Health and Public Sector teams leveraged support through a program of policy and investment lending that was informed by analytical work such as PERs and fiscal analysis. SUBSIDIZED CONTRIBUTION PAYMENTS The Bank also helped increase revenues for health by subsidizing contributions to various insurance institutions for low-income groups. This type of Bank support was implemented through lending and policy dialogue in countries such as Benin, Bolivia, Cambodia, the Dominican Republic, Ghana, Mexico, Rwanda, Turkey, and Vietnam. The Bank supported the explicit targeting of subsidies to finance contributions for low-income groups through means testing in Georgia, Rwanda, and Turkey and through geographic location in Cambodia and Egypt. In other countries, similar support served to subsidize access to health insurance for low-income groups. Vietnam’s public insurance fund is financed from payroll taxes and general tax revenues. For households not active in the formal sector, the government makes contributions, defined as a proportion of the minimum wage, from the state budget. In Vietnam’s Mekong Region the Bank health project 21 CHAPTER 2 REVENUE COLLECTION FOR HEALTH cofinanced enrollment for near-poor households in the Health Care Fund for the Poor. In Georgia, Bank lending supported the publicly funded Medical Insurance Program for the poor, which provides an extensive benefit package with zero copayments. Low-income beneficiaries receive a publicly funded voucher to enroll with a private insurance company (Bauhoff et al. 2011). IEG found that in Rwanda the Bank provided technical assistance on the law for CBHI.5 Under this law, the government and donors subsidize CBHI enrollment for the three lower-income quintiles through means-tested targeting while the remaining households pay full contributions. By 2010 about 44 percent of CBHI revenue was from the government budget, 31 percent from households, and 22 percent from donors. As a result of this Bank support, insurance enrollment has increased in these countries. Whether these subsidized contribution payments have also improved service use among pool members is discussed in chapter 3. SERVICE USE AND FINANCIAL PROTECTION As shown in Figure 2.1, user payments are the most important revenue source for the health sector in lower-income countries. Bank advice on reducing user fees and copayments has fiscal and equity implications. However, this type of support often lacked the necessary analytical underpinnings, and—contrary to findings from other researchers (Lagarde and Palmer 2011; Dzakpasu et al. 2013)—evidence is missing that reducing copayment levels improved service use and financial protection. While the Bank had recommended introducing copayments with exemptions for lower-income groups in Romania, a recent study found that, compared with those in neighboring countries, households in Romania are far more likely to forgo care because they cannot afford the fees, and young people are more likely to borrow or sell assets to pay for care (Tambor et al. 2013). In El Salvador the Bank supported the elimination of copayments in hospitals but did not prepare providers enough for the resulting demand increase. A 2011 evaluation finds a 40 percent increase in service use after user fees were abolished, and raises concerns about transparency and corruption in finances at unaudited hospitals. It recommends increasing staffing in hospitals to reduce waiting lists caused by the higher number of patients (AfGH 2011). A case study prepared by IEG for this evaluation found similar concerns in Kenya where the new government had just eliminated user fees for primary care in 2013 and was to allocate higher public funding for primary care to compensate for forgone revenues from user fees. The Bank estimated that an additional $8.1 million is needed to compensate providers. It also identified challenges on the flow of health funds to health facilities in a decentralized system. The Bank could analyze the fiscal and equity implications of changes in user fees, as emphasized in the HNP strategy, and inform governments on the amount of public funds needed to compensate providers for reduced or eliminated fees. 22 CHAPTER 2 REVENUE COLLECTION FOR HEALTH In sum, most Bank support in health financing went to public revenue collection for health. As a result of the Bank’s help, government health budgets were increased; health spending was protected against budget cuts during an economic crisis; advice on fiscal space for health was considered; and governments were assisted in subsidizing compulsory contributions to various health insurance for low-income groups. Some timely advice on a greater role for alcohol and tobacco taxes has been given, but this is very limited. Public Sector and Health teams emphasized strong institutions and monitoring and evaluation to ensure revenues positively affect the provision of care. While this type of support has been decreasing over time, there have been some notable successes. Evidence indicates that these have occurred with strong government commitment at both the economy-wide and sector levels that the Bank has supported and when Bank staff drew on a variety of skills across the Bank to engage government. Bank advice and operations have also supported governments which have tried to lower user payments as a source of revenue. However, this type of support often lacked the necessary fiscal and equity analysis, and evidence is missing that Bank support to reduce copayments has improved equity in service use and financial protection. References AER (Action for Economic Reforms). 2014. “DOH Budget Increase for 2014 ‘Biggest Ever’ due to Sin Tax Law.” Press release, January 15. http://aer.ph/doh-budget-increase-for-2014-biggest- ever-due-to-sin-tax-law/ AfGH (Action for Global Health). 2011. “Health Spending in El Salvador: The Impact of Current Aid Structures and Aid Effectiveness.” Country Briefing 3, EU Health ODA and Aid Effectiveness, German Foundation for World Population, Hannover, Germany. Baldacci, E., B. Clements, S. Gupta, and Q. Cui. 2008. “Social Spending, Human Capital, and Growth in Developing Countries.” World Development 36 (8): 1317–1341. Bauhoff, Sebastian, David R. Hotchkiss, and Owen Smith. 2011. “The Impact of Medical Insurance for the Poor in Georgia: A Regression Discontinuity Approach.” Health Economics 20: 1362–1378. Begg, David, Stanley Fischer, and Rudiger Dornbusch. 2000. Economics. Sixth revised edition. New York City: McGraw–Hill Publishing Company. Cherecheş, R.M., M.I. Ungureanu, P. Sandu, and I.A. Rus. 2013. “Defining Informal Payments in Healthcare: A Systematic Review.” Health Policy 110: 105–114. Dzakpasu, Susie, Timothy Powell-Jackson, and Oona M.R. Campbell. 2013. “Impact of User Fees on Maternal Health Service Utilization and Related Health Outcomes: A Systematic Review. ” Health Policy and Planning 2013: 1–14. Euromonitor International. 2013. Country Report: Tobacco in Brazil. http://www.euromonitor.com/tobacco-in-brazil/report. Gottret, Pablo, and George Schieber. 2006. Health Financing Revisited: A Practitioner’s Guide. Washington, DC: World Bank. 23 CHAPTER 2 REVENUE COLLECTION FOR HEALTH Gupta, S., M. Verhoeven, and E. R. Tiongson. 2002. “The Effectiveness of Government Spending on Education and Health Care in Developing and Transition Economies.” European Journal of Political Economy 18 (4): 717–737. IEG (Independent Evaluation Group). 2007. “Ghana: Second Health and Population Project (Credit No. 2193) Health Sector Support Project (Credit No. 2994). ” Project Performance Assessment Report 39118, World Bank, Washington, DC. ______. 2011. “Results-Based Health Programs in Argentina and Brazil: Project Performance Assessment Report.” Report No. 62571, World Bank, Washington, DC. ______. 2014. “Republic of Albania Health System Modernization Project (P082814) and Social Sector Reform Development Policy Loan (P116937).” Project Performance Assessment Report No. 88074, World Bank, Washington, DC. Lagarde, M., and N. Palmer 2011. “The Impact of User Fees on Access to Health Services in Low- and Middle-Income Countries.” Cochrane Database of Systematic Reviews 2011, Issue 4. Art. No.: CD009094. DOI: 10.1002/14651858.CD009094. Moreno-Serra, Rodrigo, and Peter Smith. 2011. “The Effects of Health Coverage on Population Outcomes: A Country-Level Panel Data Analysis.” Working Paper, Business School and Center for Health Policy, Imperial College London, United Kingdom. Nasrudin, Rus’an, Ledi Trialdi, Djoni Hartono, and Abdillah Ahsan. 2013. “Tobacco Economic of Indonesia: Poor Households’ Spending Pattern, Tax Regressivity and Economic Wide Impact of Tax Simplification.” Working Paper in Economics and Business Volume III, No.2/2013, University of Indonesia, Jakarta. http://econ.fe.ui.ac.id/uploads/201302.pdf OECD (Organisation for Economic Co-operation and Development). 2013. OECD Health Statistics 2013. Current Health Expenditure by Healthcare Financing, ICHA-HF, 2011. Paris: OECD Publishing. http://www.oecd.org/els/health-systems/health-expenditure.htm. Pande, A., A. Leive, M. Smitz, P. Eozenou, and E. Ozcelik. 2013. “Macro Fiscal Context and Health Financing Factsheet,” World Bank, Washington, DC. Savedoff, William D., David de Ferranti, Amy L. Smith, and Victoria Fan. 2012. “Political and Economic Aspects of the Transition to Universal Health Coverage.” Lancet 380 (9845): 924– 932. Tambor, Marzena, Milena Pavlova, Bernd Rechel, Stanislawa Golinowska, Christoph Sowada, and Wim Groot. 2013. “The Inability to Pay for Health Service in Central and Eastern Europe: Evidence from Six Countries.” European Journal of Public Health 1–7. doi: 10.1093/eurpub/ckt118. Tandon, Ajay, and Cheryl Cashin. 2010. “Assessing Public Expenditure on Health from a Fiscal Space Perspective.” HNP Discussion Paper No. 56053, World Bank, Washington, DC. World Bank. 2009a. “Health Insurance and Competition.” Europe and Central Asia Region, Report No. 44316-ECA, World Bank, Washington, DC. 1In April 2001, heads of state of African Union countries met and pledged to set a target of allocating at least 15 percent of their annual budget to their health sector. http://www.who.int/healthsystems/publications/Abuja10.pdf 2An increase of $100 in government spending per capita results in a reduction of 13.2 per 1,000 in under-five mortality as well as a decrease of 2.6 and 2.2 per 1,000 in adult female and male mortality rates, respectively. 24 CHAPTER 2 REVENUE COLLECTION FOR HEALTH 3 Governments raise direct taxes from earnings, indirect taxes on consumption, and wealth taxes. A tax system is considered fair if it generates higher taxes on the rich to finance public goods and services, such as health services, predominantly used by the poor. Indirect taxes on consumption (e.g., value-added taxes) are regressive if the poor spend a higher proportion of their income on goods subject to these taxes than the rich. While such indirect taxes reduce the redistribution effect from the rich to the poor, they may still have a pro-poor effect if they finance public services predominantly used by the poor (Begg et al. 2000; IMF 2011). 4According to a directive from the European Commission’s Taxation and Customs Union, member states must apply to cigarettes a specific excise duty per unit of the product and a proportional excise duty calculated on the basis of the weighted average retail selling price. For more information visit http://ec.europa.eu/taxation_customs/taxation/excise_duties/tobacco_products/legislation/index _en.htm. 5On December 30, 2007, Parliament passed Law No. 62/2007 on establishing and determining the organization, functioning, and management of the community-based health insurance. For more information visit http://lip.alfa-xp.com/lip/AmategekoDB.aspx?Mode=r&pid=7586&iid=2031. 25 3. Pooling Health Funds and Risks Highlights  Risk pooling is important to address equity and financial sustainability in health. Countries have multiple pooling arrangements, leading to unequal risk distributions across pools and to different pools for the various socioeconomic groups.  World Bank support contributed to increased risk pooling in middle-income countries. Similarly, the International Finance Corporation supported risk pooling through public and private insurance.  The World Bank Group built institutional, management, and technical capacity in government and insurance administration to manage funds and risks, and Bank analytical work informed policy decisions. However, Bank assistance to health insurance has been diminishing over time. Projects were less effective in countries with decentralized health systems.  Equity in pooling has improved where the Bank helped subsidize coverage of the poor. But coverage did not always lead to pro-poor spending, improved service use, or greater financial protection. Fragmented pooling remains an issue and can affect efficiency. Success factors included strength in institutions, management, technical capacity, and information. With the exception of user payments, all revenues for health are pooled in public and private health insurance and in central and local government budgets, and then transferred to providers. Pooled financing reallocates funds from healthy to sick individuals—that is, from individuals with a low risk of illness to those with a high risk who are more likely to occur higher health care costs. As countries grow economically, pooled health financing in national health systems and health insurance comes to dominate revenues from user payments. The objective of pooling is to reduce the out-of-pocket price the patient pays when using services and to ensure financial protection against catastrophic health payments and equity in service use. But managing health revenues in a way that ensures equitable and efficient pooling is a major challenge (Gottret and Schieber 2006). Also, increased pooling contributes to higher health spending by increasing the demand for health care. Increased pooling, in national health systems or through insurance, benefits consumers and providers. Individuals who are insured or covered in the public system will copay less when seeking care. They are thus expected to report greater service use and lower copayments than those who pay user fees. Increased pooling is also good for providers because user payments from patients are erratic revenues in low-income environments. Instead, contracted providers will receive a stable amount of revenues from the government and insurers to treat patients (Box 3.1). 26 CHAPTER 3 POOLING HEALTH FUNDS AND RISKS Box 3.1. Risk Pooling Arrangements Countries introduce different risk-pooling arrangements to protect individuals against the financial risk of illness. In health systems with automatic coverage, public revenues are pooled in the government’s health budget, and the public sector plays an insurance role, even if it is not formally constituted as an insurance plan (Kutzin 2007; Savedoff et al. 2012). The government transfers revenues from the central and local government budgets to providers to pay for health care services provided to the population. In countries with decentralized health systems (such as Argentina, Kazakhstan, and the Philippines), health revenues from the central and local governments are pooled at the local level (state or region) and transferred to providers to finance health care delivery to patients. In addition, public and private health insurers pool health funds, including from individual contributions, premium payments, and government subsidized contributions, to pay for the financial risk of illness among their members. Chapter 3 summarizes the challenges related to automatic coverage and to mandatory and voluntary pooling. It then describes support from the International Finance Corporation (IFC) and World Bank to countries in meeting them. It offers evidence on how this support to pooling affects equity in health financing and service use, financial protection, and efficiency. Challenges In most countries, multiple pooling arrangements coexist, leading to a risk of fragmentation. Generally, formal sector workers are covered under mandatory social insurance; higher-income groups can afford paying higher premium to enroll with private voluntary insurance (to access specialist care and private providers); and the government provides automatic coverage in public health facilities for those who are excluded from these insurance arrangements, mainly the lower-income and informal sector groups. As a result, different socioeconomic groups pool their health risk among themselves in different institutions with different revenue raising capacity and access to different health benefits. The resulting fragmentation raises concerns about equity in service use across different groups. It also raises concerns about the financial sustainability of small risk pools (Box 3.2). One of the main challenges in government health systems (automatic coverage) in low- and middle-income countries is that government allocations are often not pro- poor. Instead, a higher share of funds is allocated to hospitals in urban areas, which are mainly used by the wealthier (Table 3.1). To improve pro-poor spending, some countries have earmarked transfers to providers mainly used by low-income groups. In Mexico, for instance, the Seguro Popular is an intergovernmental revenue transfer within the national health system from the center to the states. The transfer is 27 CHAPTER 3 POOLING HEALTH FUNDS AND RISKS defined based on the number of low-income individuals affiliated with Seguro Popular and is cofinanced by the states. Table 3.1. Incentives and Challenges in Different Risk Pooling Arrangements Risk Pooling Pooling Institutions Incentives Challenge Automatic Government budget Governments allocate Pro-rich spending; coverage (central and local) funds for political underfunded services in low- reasons (e.g., urban income areas hospital) Mandatory Public health insurance Moral hazard among Exclusion of informal groups; pooling Private health insurancea insured individuals inefficient service use; cost increase Voluntary Private health insurance Moral hazard; risk Financial sustainability pooling Community-based health selection insurance a. In some countries including Georgia, India, and Slovakia, private insurers offer mandatory insurance coverage. Mandatory participation by law in public or private health insurance limits insurance coverage to all or a defined population group (e.g., formal sector employees). The excluded are mainly lower-income groups who work in the informal sector—the majority of the population in developing countries (Table 3.1). They receive automatic coverage through the public system; however, this is often less comprehensive. The small membership size of social insurance in countries with nascent formal sectors can endanger an insurer’s financial viability. A fiscal and equity problem can arise if the government has to finance the deficit caused by the medical service use of wealthier insurance members (Box 3.2). Box 3.2. Financial Incentives May Endanger the Sustainability of Risk Pools Risk pooling involves trade-offs between equity gains caused by reduced uncertainty about the financial consequences of ill health and efficiency losses created by financial incentives (Arrow 1963; Cutler and Zeckhauser 2000; Zeckhauser 1970). These incentives, which include adverse selection, moral hazard, and supply side–induced demand, can lead to higher costs for the risk pool and endanger its financial sustainability. Adverse selection arises when those enrolling in voluntary risk pools are mainly high-risk individuals, resulting in high-cost pools that may not be financially viable. Moral hazard occurs when pool members overuse medical services because they copay at a reduced price for care, which can increase costs. Finally, providers who are reimbursed by the pool based on fee-for-service payments have a financial incentive to oversupply care, which also leads to higher costs. Some governments introduced separate institutions to pool public funds and health risk for informal sector groups. Thailand, for example, has established the government-funded Universal Coverage Scheme, which uses tax revenues to 28 CHAPTER 3 POOLING HEALTH FUNDS AND RISKS provide coverage for individuals not covered by formal sector social insurance. Benin and Cambodia created health equity funds to strengthen financial risk protection among the poor and informal sector groups. While they differ in design, these funds were set up to manage health care subsidies for eligible population groups. The volume of voluntary pooling in private health insurance is inconsequential. Few people can afford voluntary private insurance, and there are generally not that many private providers to contract with in developing countries (appendix D). Private health insurance pays mainly for services not covered by social insurance or by the government (e.g., specialist care in the private sector) and is used to avoid waiting lists for elective treatment. To ensure their financial viability, private insurers have an incentive to enroll people at low risk of being ill. Community-based health insurance (CBHI) is more prominent in Africa for informal sector groups. While members of such schemes often report better access than those who pay user fees, they tend to have access to a smaller benefit package than those with social health insurance (SHI). In addition, voluntary pooling in private insurance and CBHI may lead to adverse selection resulting in small, high-risk pools with predominantly sicker individuals, and may undermine pool finances (Box 3.2). Bank Group Support to Risk Pooling The World Bank Group has tried to help countries address challenges in pooling arrangements. About 40 percent of the Bank’s health financing operations support automatic coverage through national health systems or enrollment into mandatory health insurance in 36 countries. The Bank Group does not take an ideological stance on risk pooling arrangements. It did not advocate for SHI or automatic coverage through national health systems, or promote private health insurance. Rather, it works within different country and risk pooling contexts. The Bank assisted Ministries of Health and local governments managing and implementing their health budgets. Management and information capacity was also built with Bank and IFC support in public and private health insurance. ANALYTICAL WORK The Bank produced several analytical reports, mainly in middle-income countries, to inform governments about the challenges of different risk pooling arrangements in different contexts. Of the 70 economic and sector work activities, a relevant word search suggests that 15 have tackled pooling. In addition, the Bank conducted eight 29 CHAPTER 3 POOLING HEALTH FUNDS AND RISKS impact evaluations on risk pooling in China and Vietnam (appendix Table A.8). It also carried out limited analysis on the welfare effect of increased pooling of domestic revenues and whether the poor benefit. Only 10 Bank poverty assessments examined the distributional aspects of public spending and conducted benefit incidence analysis (see chapter 5). AUTOMATIC COVERAGE THROUGH NATIONAL HEALTH SYSTEMS The Bank provided technical assistance to improve resource management in national health systems. Decentralization of funding to lower levels of government was supported in several countries including Cambodia, Indonesia, Kosovo, Pakistan, Rwanda, and Serbia. It helped build technical capacity to manage fund pooling and health resources in Afghanistan, Cambodia, and Vietnam. The Bank has helped countries target the poor through automatic coverage in national health systems. In Argentina it supported the introduction of the Plan Nacer program, which targets supply-side subsidies to health facilities used by the poor (IEG 2011). Depending on the country context, Bank support to automatic health coverage was accompanied by public sector measures to strengthen the public management of health funds. In Afghanistan the Bank supported expenditure management through program-based budgeting which links health sector spending to the national strategy and to prioritize allocation. In Kenya and Tanzania it helped with public expenditure tracking to identify inefficiencies in spending. In Kenya the Poverty Reduction and Economic Management and Health, Nutrition, and Population (HNP) teams worked with the government to improve budget transparency by adding more detailed line items to track health expenditure. PUBLIC AND PRIVATE HEALTH INSURANCE The Bank has advised several low- and middle-income countries on the level of contributions paid to social insurers and on subsidized enrollment. The Bank also supported the strengthening of health insurance governance and management. The IFC has made only two direct investments in private health insurance (both originating from the Financial Markets Group) and two investments in a health maintenance organization (HMO) provider network (appendix Table D.3). The IFC has supported the expansion of Nigeria’s largest integrated HMO provider network with two investments and one advisory services project (2007 and 2009), and provided funding through its Performance-Based Grants initiative to support a project under the Global Program on Output-Based Aid to support the HMO’s community-based health plan targeting the informal sector. And it has supported the expansion of private insurance providers in the Europe and Central Asia and the Middle East and North Africa Regions with two equity investments in 2011 and 30 CHAPTER 3 POOLING HEALTH FUNDS AND RISKS 2012, respectively. IFC has recently approved a micro health insurance advisory project in India (appendix D). Two equity funds have been established as part of the Health in Africa Initiative (HiA)—the Africa Health Fund and the Investment Fund for Health in Africa. Both funds include insurance as target investments. They have invested in insurance companies and HMOs in Kenya, Nigeria, and Tanzania. The initiative has also supported governments in Kenya and Nigeria in strengthening their public health insurer through IFC’s advisory services. IFC support to the government of Uganda in reforming legal and regulatory frameworks aims to increase private sector participation in publicly funded health programs (appendix D). Effectiveness of World Bank Group Support to Pooling In practice, risk pooling may not work as expected for several reasons. This section examines how effectively Bank Group support to risk pooling has helped countries develop management and technical capacity, and improve equity in pooling, service use, financial protection, efficiency and financial sustainability of risk pools. MANAGEMENT AND TECHNICAL CAPACITY Management and technical capacity is important to ensure that health budgets in central and local governments are effectively implemented. Although the Bank has supported governments in increasing their health budget, budget implementation is limited in some countries. Low budget-execution rates were reported in Afghanistan and the Democratic Republic of Congo because of constrained technical capacity, lack of financial authority, and complicated financial and procurement procedures. In Afghanistan, Bank analytical work also found no clear targeting of public funds to areas with worse health outcomes, and the funds spent by the government and donors on health were not coordinated across provinces (Belay 2010). More recently, the IEG case study found, the Bank played an effective coordinating role among donors and emphasized the use of M&E which supported the government in evidence-based fund management. In Brazil, Bank support to building the institutional foundations at municipality level—including budget management, accounting, monitoring, financial management, and managerial capacity— contributed to the timely execution of the health budget in a decentralized health system (IEG 2011). Through its multisector analytic work, the Bank provides important information about public sector reforms (such as decentralization) that affect health financing. The Independent Evaluation Group (IEG) found that in Kenya, the Bank’s Health and Public 31 CHAPTER 3 POOLING HEALTH FUNDS AND RISKS Sector teams have spearheaded several analytical products and policy discussions in health financing. The Public Sector team was supporting a fiscal space analysis to ascertain the efficiency of the health sector and determine the value for money aspect. Bank work through the 2012 Public Expenditure Review and a Public Expenditure Tracking Survey with sub-national analysis and frontier analysis helped identify inefficiencies and informed the government about challenges that need to be addressed in health sector devolution (appendix E). This type of Bank analysis helped inform policy makers in their management and technical decisions. Health financing projects in decentralized health systems with automatic coverage performed worse than the average health financing projects in ratings of the IEG.1 Examples include projects in Brazil, the Dominican Republic, Ghana, and Mauritania, where the Bank overestimated the political commitment and technical knowledge in the government to decentralize, and failed to calibrate project design to local capacity. In Indonesia, three Bank health projects assisted the government in shifting responsibilities for planning and management of resources to district authorities. While district health budgets tripled, the additional resources were insufficient and ineffectually allocated. Allocation formulas were not adequately poverty calibrated, and the limited own-revenue raising capacity of poorer districts negatively affected horizontal equity across districts. Few districts reached the target of allocating at least 15 percent of local government spending to health, and health service use among the poor and near-poor increased very modestly in some districts and decreased in others. IEG found several factors that limited the success of Bank- supported health sector decentralization in Indonesia, including insufficient attention to define roles and responsibilities at different government levels, inadequate information systems, and considerable overlap and duplication of tasks across government (including in management of human resources) that generated inefficiencies in the organization and delivery of services. The Bank did help introduce better planning and budgeting methods that over time have helped improve information systems for regional monitoring. IEG concluded that Bank support to health decentralization needs to be grounded in a realistic understanding of how institutions work, and how they can be expected to change in the political context in which projects operate (IEG 2013). Bank technical assistance and lending helped governments develop new laws and administration to strengthen social insurance management, mainly in middle-income countries. Following Bank advice governments in Albania, the Dominican Republic, Serbia, and Vietnam introduced changes to health insurance laws and regulations. In Turkey it gave advice on the Social Security Administration law and hospital budgets. 32 CHAPTER 3 POOLING HEALTH FUNDS AND RISKS Governments in European countries introduced measures to stem the deficit in health insurance as recommended in PERs and analytical work supported under policy lending. Bank loans financed beneficiary identification and medical claims management systems in Georgia, Montenegro, and Serbia, and improved efficiency in insurance management. In these countries, Bank Public Sector and Health teams worked closely on these reforms, which contributed to successful implementation. Although the HNP strategy highlights the Bank’s comparative advantage in health insurance analytical capacity, Bank teams did not always maintain support to health insurance and address shortcomings in management and institutional capacity. In Tanzania in the 1990s and early 2000s, the Bank was instrumental in setting up the National Health Insurance Fund, the National Social Security Fund and the Community Health Fund. However, by the mid-2000s, the Bank reduced its engagement in these funds, partly because of changes in Bank resources as well as reduced demand for Bank assistance by the government as other donors increased their health financing technical assistance. By 2013, only about 8.6 percent of the population is insured in Tanzania, enrollment is highly unequal, and the National Health Insurance Fund has high financial reserves because investments are tied up, and because of a lack of understanding among health facility managers about how to submit a medical claim to the fund to get paid for care provided to insured patients. Little additional reforms were introduced to address these shortcomings. IEG finds that given the strong Bank involvement in the past, the Bank team could have provided further analysis and technical assistance to the funds to strengthen institutions and management capacity, especially since the government is now considering scaling-up insurance and raising labor taxes to finance enrollment. In Bolivia, IEG found that the Bank was instrumental in helping the government establish health insurance for mothers, children, and the elderly; however, this support was not maintained over time. Similar concerns were identified in Ghana and Rwanda, where Bank support to health insurance management has been reduced over time partly due to a shift in government priorities and Bank resources (appendix E). In the Kyrgyz Republic the Mandatory Health Insurance Fund is nearly all funded from general government revenues and manages a state-guaranteed benefit package for the whole population. The fund is also the sole purchasing agency for health services within the health system (Kutzin 2013). The Bank has supported these reforms in collaboration with other donors, through analytical work and capacity building and as a convener of donor efforts. Health reforms have benefited from strong political commitment and technical capacity, which are important success factors. They include the 10-year (1996–2005) government health reform strategy, several champions in the government in support of the reforms, and effective use of 33 CHAPTER 3 POOLING HEALTH FUNDS AND RISKS monitoring and evaluation (M&E) so that early successes could be publicized and used to generate support for subsequent reforms. In addition, Bank interventions and policy dialogue at key moments of political opposition sustained momentum for insurance reforms (IEG 2008). Under the HiA, IFC has conducted a strategic review of Kenya’s National Hospital Insurance Fund (IFC 2011a) and a market assessment of prepaid plans (IFC 2011b). Its recommendations were accepted by the government and are now being implemented. Through a follow-on advisory project, IFC is assisting the government on integrating private hospitals into the national health system (e.g., regulatory framework for accreditation and contracting with private providers). The evaluation’s case study found the Bank is also working on recommendations made by IFC through a project on health insurance subsidies for low-income groups (appendixes D and E). EQUITY IN POOLING Ensuring that pooling arrangements are equitable requires an effective way to cross- subsidize across pool members and to inject the pools with a sufficient amount of public funds that are sustainable. In health systems with automatic coverage, the Bank conducted a few incidence analyses to alert governments to issues in pro-poor allocation of funds. In Indonesia the Bank found that overall allocation of public spending on health is low and spending needs to be increased strategically to reach the poor effectively and to include demand-side measures. In Ghana the Bank identified increased pro-poor spending over time. However, in Nicaragua it found that public spending on social services is not pro-poor—it benefits all socioeconomic groups about equally. The share of poor included in risk pools increased where the Bank helped governments subsidize their enrollment (Box 3.3). In Turkey insurance coverage for the poorest increased more than fourfold between 2003 and 2011, generating a coverage rate of 85 percent for the poorest (Atun et al. 2013). The public health insurance has recently incorporated the Green Card Program, which subsidizes health care for the poorest income group and is funded by general government revenues (Atun et al. 2013). Similarly, in Colombia the Bank’s development policy operation helped increase the enrollment of low-income groups in government- subsidized insurance from 10.7 million in 2002 to 18.2 million in 2007. By March 2014, about 43 million individuals or 90 percent of the population was insured in Colombia (www.sispro.gov.co). In Rwanda, IEG found that Bank support to the CBHI law makes insurance enrollment mandatory and increased CBHI enrollment to about 85 percent of the population by 2012. A Bank project in the Philippines 34 CHAPTER 3 POOLING HEALTH FUNDS AND RISKS reached the poor in the National Health Insurance Indigent Program, but the percentage enrolled is unknown. Box 3.3. Bank Analysis Informed Risk Pooling in China and Vietnam Researchers have found that individuals enrolled with the voluntary Vietnamese Health Insurance (VHI) program were more likely to use outpatient care, and the poorest insured are 10 times more likely to seek care than the uninsured. But there were concerns that adverse selection and the use of unnecessary care would threaten the financial sustainability of VHI (Jowett et al. 2004). The Bank, using data from the late 1990s, confirmed improved health outcomes among VHI members (Wagstaff and Pradhan 2005). Informed by these studies, the Bank and other donors through the Second Poverty Reduction Support Credit in 2003 helped the government establish the Health Care Fund for the Poor (HCFP), which provides the same benefits as the VHI. In a follow-up study the Bank found that 60 percent of eligible households were covered by 2006, and the HCFP was well targeted to the poor; however, there was adverse selection (Wagstaff 2010). To address selection problems, Bank lending helped increase HCFP enrollment to 96 percent among the poor and 42 percent among the near-poor by 2011. In China, based on data from the late 1990s up to 2004, the Bank found that the Rural Cooperative Medical Scheme diminished the risk of high user payments for households (Wagstaff and Lindelow 2005), and ill health can have a large impact on household income, labor supply, and medical expenditures, even for the insured, raising concerns about the effectiveness of the Scheme (Lindelow and Wagstaff 2005). This was followed up by an impact evaluation of the government-subsidized New Cooperative Medical Scheme (NCMS) established in 2003, which found lower enrollment among the poor and higher enrollment among the chronically sick, pointing to adverse selection. Service use increased, but the NCMS did not reduce user spending for the poor (Wagstaff et al. 2007). Informed by these studies, the Bank has supported the NCMS since 2009. Enrollment increased to 99 percent in 2012. In some countries, Bank support was less successful in targeting the poor for inclusion in risk pools. In Tunisia only 9 percent of the eligible poor are covered under the government-funded Free Medical Assistance Program. The poor are not reached because of institutional constraints including nontransparent eligibility criteria that are subject to manipulation,2 and the Bank could in fact have addressed weak targeting of the poor in the policy dialogue (IEG 2014). In Georgia, despite means testing supported by the Bank, a significant proportion of eligible households were excluded from the Medical Insurance Program, mainly because of insufficient information (Bauhoff et al. 2011). In Ghana the National Health Insurance System covers 40 percent of the population, which are predominantly the nonpoor. The Bank was instrumental in convincing the government to extend coverage to children and youths under the age of 18 and pregnant women to achieve the relevant Millennium Development Goals. It also 35 CHAPTER 3 POOLING HEALTH FUNDS AND RISKS discussed a more generous definition of the poor, which according to the National Health Insurance System is only 1.7 percent of the population and far below the national poverty rate of 30 percent (appendix E). But reaching the poor requires commitment by government. IEG found that insufficient political and financial commitment by the government and limited implementation capacity were constraining factors to reforms as was weak M&E systems to track equity in access for insured and uninsured individuals (IEG 2007). The IFC supported private and public health insurers that provide both mandatory and voluntary coverage mainly for formal sector employees. In IFC’s managed-care investments in Nigeria, HMO enrollees are primarily federal employees and employees of large corporations and members of the National Health Insurance Scheme. There is no evidence from the IFC’s Development Outcome Tracking System to suggest the 1.2 million HMO enrollees in Nigeria and 613,000 patients served as of FY13 were poor. In Kenya, IFC advisory support to the public insurer contributed to expanding coverage to civil servants. IFC support also resulted in the government’s decision to expand health insurance subsidies to the indigent population (poorest 9 million Kenyans). IFC assisted the government of Meghalaya (India) with the contracting of a private insurer to offer health insurance to low- income individuals. In Tanzania, the Investment Fund for Health in Africa invested in the largest private insurance company; its clientele is primarily corporate employees who are mainly higher- and middle-income individuals (appendix D). Dalberg (2012) finds that the equity investment through the Africa Health Fund in Kenya is reaching the poor but not the very poor. SERVICE USE RELATIVE TO NEED AND FINANCIAL PROTECTION Few Bank projects report how increased pooling of domestic revenues affects service use, particularly in automatic coverage systems. In Uzbekistan primary health spending to facilities mainly used by low-income groups rose from 41 percent in 2004 to 45.2 percent of public health expenditure in 2011, and the number of visits per person per year has increased from 3.8 in 2005 to 4.4 in 2010. In Tajikistan the reforms in public revenues had no effect on care seeking; as patients did not seek care, households also spent less on health. In Argentina service use of protected programs remained at a high level and increased for the treatment of tuberculosis and HIV vertical-transmission prevention (IEG 2011). Risk pooling does not necessarily translate into improved service use and financial protection. In China no recent information is available on the impact of the use of care and how effectively the New Cooperative Medical Scheme protects households against the financial consequences of ill health. The Turkey insurance reform supported by the Bank contributed to improved equity in health financing across 36 CHAPTER 3 POOLING HEALTH FUNDS AND RISKS income groups and substantially reduced catastrophic expenditures for the poor while increasing their service use (Atun et al. 2013). The Colombia health insurance for the poor lowers mean inpatient spending for patients and is associated with the use of preventive series and health gains for children. However, insurance does not affect spending for outpatient care nor does it increase utilization of curative care (Miller et al. 2013). In Georgia, insurance did not affect utilization of care (Bauhoff et al. 2011). Among the reasons why were low quality of care and the exclusion of pharmaceutical products from coverage (World Bank 2012). Bank analytical work should identify and address the reasons why pooling does not lead to the expected outcomes, as done in Georgia and Vietnam (Somanathan et al 2014), for example. Insufficient information about benefits is a limiting factor. In Vietnam a Bank impact evaluation finds that while the Health Care Fund for the Poor (HCFP) has reduced user payments for members, it did not affect their service use (Wagstaff 2010). Among the reasons were that HCFP members were not well-enough informed about benefits. Thus the Bank helped improve knowledge about HCFP benefits for 98 percent of members, and by 2011, 46 percent of the poor HCFP members used hospital and outpatient care. Similarly, service use among members of the Medical Insurance Program in Georgia was low because the program provided too little information on the benefit package; beneficiaries failed to receive vouchers for enrollment; and providers continued requesting under-the-table payments from patients (Bauhoff et al. 2011). Some countries report substantial improvements for the poor insured, but this information is limited. Other researchers report improved utilization and reduced out-of-pocket spending for the insured in some Bank-supported risk pools. Based on 2006 household survey data, CBHI in Rwanda is associated with significantly increased utilization of health services when they are needed and with lower user payments. The incidence of catastrophic health expenditure was almost four times as high for noninsured households as for the insured (Saksena et al. 2011). In Ghana the insured poor have greater access to health care, lower copayments, and better health outcomes than the noninsured poor. Insurance has also reduced catastrophic spending on health and protected households against impoverishment (Nguyen et al. 2011). In Cambodia the health equity funds led to sharp gains in utilization of key services and reduced spending by the poor, and they significantly lowered copayments, catastrophic expenditures, and debt by the poor (Flores et al. 2013). Little evidence of the impact of IFC’s support to health financing on improved service use or financial protection is due to the newness of the projects or scarcity of data (appendix D). As IFC investments in private health insurers do not monitor utilization of care and copayments by the insured, no information is on hand about 37 CHAPTER 3 POOLING HEALTH FUNDS AND RISKS their effectiveness. Partly this is due to the transaction-like nature of some IFC’s advisory services; however, recent projects have recommended post-completion reporting on access to improved services. EFFICIENCY AND FINANCIAL SUSTAINABILITY Efficiency concerns arise where pooling is fragmented. In East Asian countries the Bank warned about high administrative costs, duplications of benefits, and loss of negotiating power with providers (Langenbrunner and Somanathan 2011). In European countries the Bank highlighted risk-equalization issues to address fragmentation (World Bank 2009a). Bank advice to the government of Turkey in 2003 cautioned about fragmented pooling (World Bank 2003). Since then, the government has consolidated the five insurance schemes into a unified general health insurance program with harmonized benefits (Atun et al. 2013). The government has also improved revenue allocation to primary care which reduced referral rates to more costly specialist care. In Hungary and Poland the Bank successfully advised against breaking up the social health insurer into multiple pools, which would have increased fragmentation. The Bank also warned about adverse selection in countries with multiple insurance funds, including China, the Slovak Republic (World Bank 2009a), and Vietnam. Achieving development results in IFC's private insurance investments have been difficult. For example, an investment aimed at reaching underserved populations in a multiple pooling environment proved difficult. Addressing fragmentation needs political commitment. In Bosnia and Herzegovina, Bank advice to consolidate various health insurers faced political resistance, and the government did not follow it. In Mexico a Bank study found that government- subsidized risk pooling among the poor through Seguro Popular incentivizes informality. While Seguro Popular improves access to care, it was associated with a 3.1 percentage point fall in the flow of workers into formality. The Bank also found that Seguro Popular has income effects. Members can avoid having to contribute to the formal SHI program by moving to the informal sector and receiving services under Seguro Popular (Aterido et al. 2011). Yet the government has chosen not to consolidate Seguro Popular with the formal SHI program to reduce fragmentation. The findings of this evaluation show that almost half of the Bank’s health financing portfolio supported the strengthening of pooling through automatic coverage or mandatory public health insurance. The Bank helped strengthen regulatory frameworks, resource management, and institutional foundations for budget execution, and invested in M&E. Bank and IFC analytical work and technical assistance helped inform governments about fund management and the expansion of coverage to the uninsured. However, in some countries the Bank could have 38 CHAPTER 3 POOLING HEALTH FUNDS AND RISKS provided further analysis and technical advice to help governments strengthen insurance institutions and address weaknesses, including in targeting the poor. Equity in pooling improved where the Bank helped subsidize enrollment of the poor. But coverage did not always lead to pro-poor spending, improved service use, or financial protection. The IFC supported private and public health insurers that provide both mandatory and voluntary coverage mainly for formal sector employees, but evidence that this has improved service use is missing. Fragmented pooling remains an issue in several countries and can reduce efficiency. References Arrow, Kenneth J. 1963. “Uncertainty and the Welfare Economics of Medical Care. ” American Economic Review 80 (5): 941–973. Aterido, Reyes, Mary Hallward-Driemeier, and Carmen Pagés. 2011. “Does Expanding Health Insurance Beyond Formal-Sector Workers Encourage Informality? Measuring the Impact of Mexico’s Seguro Popular.” Policy Research Working Paper 5785, World Bank and Inter- American Development Bank, Washington, DC. Atun, Rifat, Sabahattin Aydın, Sarbani Chakraborty, Safir Sümer, Meltem Aran, Ipek Gürol, Serpil Nazlıoğlu, Şenay Özgülcü, Ülger Aydoğan, Banu Ayar, Uğur Dilmen, and Recep Akdağ. 2013. “Universal Health Coverage in Turkey: Enhancement of Equity. ” The Lancet 382: 65–99. Bauhoff, Sebastian, David R. Hotchkiss, and Owen Smith. 2011. “The Impact of Medical Insurance for the Poor in Georgia: A Regression Discontinuity Approach.” Health Economics 20: 1362–1378. Belay, Tekabe A. 2010. Building on Early Gains in Afghanistan’s Health, Nutrition, and Population Sector: Challenges and Options. Washington, DC: World Bank. https://openknowledge.worldbank.org/handle/10986/2459 License: CC BY 3.0 IGO. Cutler, David M., and R.J. Zeckhauser. 2000. “The Anatomy of Health Insurance.” In Handbook of Health Economics, edited by A. Culyer and J.P. Newhouse, 563 –642. Amsterdam: Elsevier Science. Dalberg. 2012. “Base of Pyramid Baseline Assessment Report: Aureos Africa Healthcare Managers Limited.” October 8, 2012, The Avenue Group, Nairobi, Kenya. (Unpublished) Figueras, Josep, Ray Robinson, and Elke Jakubowski. 2005. Purchasing to Improve Health Systems Performance. New York City: McGraw–Hill International. Flores, Gabriela, Por Ir, Chean R. Men, Owen O’Donnell, and Eddy van Doorslaer. 2013. “Financial Protection of Patients through Compensation of Providers: The Impact of Health Equity Funds in Cambodia.” Health Economics 32 (6): 1180–1193. Gottret, Pablo, and George Schieber. 2006. Health Financing Revisited: A Practitioner’s Guide. Washington, DC: World Bank. IEG (Independent Evaluation Group). 2007. “Ghana: Second Health and Population Project (Credit No. 2193) Health Sector Support Project (Credit No. 2994). ” Project Performance Assessment Report No. 39118, World Bank, Washington, DC. ______. 2008. “Kyrgyz Republic Health Sector Reform Project (Credit 2860-KG) and Second Health Sector Reform Project (Credit 3506-KG).” Project Performance Assessment Report No. 44513, World Bank, Washington, DC. 39 CHAPTER 3 POOLING HEALTH FUNDS AND RISKS ______. 2011. “Results-Based Health Programs in Argentina and Brazil: Project Performance Assessment Report.” Report No. 62571, World Bank, Washington, DC. ______. 2013. “Indonesia—First and Second Phase of the Provincial Health Project and Health Work Force and Services Project/Provincial Health III.” Project Performance Assessment Report No. 77878, World Bank, Washington, DC. http://documents.worldbank.org/curated/en/2013/11/18537800/indonesia-first-second- third-phase-provincial-health-project ______. 2014. Tunisia Country Program Evaluation, FY05–13. Washington, DC: World Bank. IFC (International Finance Corporation). 2011a. Strategic Review of the National Hospital Insurance Fund—Kenya. Washington, DC: World Bank. ______. 2011b. Market Assessment of Private Prepaid Schemes in Kenya. Washington, DC: World Bank. Jowett, Matthew, Anil Deolalikar, and Peter Martinsson. 2004. “Health Insurance and Treatment Seeking Behavior: Evidence from a Low-Income Country.” Health Economics 13: 845–857. Kutzin, Joseph. 2007. “Myths, Instruments, and Objectives in Health Financing and Insurance.” In Extending Social Protection in Health: Developing Countries’ Experiences, Lessons Learnt, and Recommendations, edited by J. Holst and A. Brandrup-Lukanow. Eschborn, Germany: Verlagfuer Akademische Schriften. ______. 2013. “Health Financing for Universal Coverage and Health System Performance: Concepts and Implications for Policy.” Bull World Health Organ. 91: 602–611. Langenbrunner, Jack, and Aparnaa Somanathan. 2011. Financing Health Care in East Asia and Pacific. Washington, DC: World Bank. Miller, Grant, Diana Pinto, and Marcos Vera-Hernández. 2013. “Risk Protection, Service Use, and Health Outcomes under Colombia’s Health Insurance Program for the Poor.” American Economic Journal: Applied Economics 5 (4): 61–91. Nguyen Ha, Yogesh Rajkotia, and Hong Wang. 2011. “The Financial Protection Effect of Ghana National Health Insurance Scheme: Evidence from a Study in Two Rural Districts. ” International Journal for Equity in Health 10:4. Saksena, P., A.F. Antunes, K. Xu, L. Musango, and G. Carrin. 2011. “Mutual Health Insurance in Rwanda: Evidence on Access to Care and Financial Risk Protection.” Health Policy 99 (3): 203– 209. Somanathan, Aparnaa, Ajay Tandon, Huong Lan Dao, Kari Hurt, and Hernan L. Fuenzalida-Puelma. 2014. Moving toward Universal Coverage of Social Health Insurance in Vietnam: Assessment and Options. Washington, DC: World Bank. Wagstaff, Adam. 2010. “Estimating Health Insurance Impacts under Unobserved Heterogeneity: The Case of Vietnam’s Health Care Fund for the Poor.” Health Economics 19: 189–208. Wagstaff, Adam, and Magnus Lindelow. 2005. “Can Insurance Increase Financial Risk? The Curious Case of Health Insurance in China.” Policy Research Working Paper 3741, World Bank, Washington, DC. Wagstaff, Adam, Magnus Lindelow, Gao Jun, Xu Ling, and Qian Juncheng. 2007. “Extending Health Insurance to the Rural Population: An Impact Evaluation of China ’s New Cooperative Medical Scheme.” Policy Research Working Paper 4150, World Bank, Washington, DC. Wagstaff, Adam, and Menno Pradhan. 2005. “Health Insurance Impacts on Health and Nonmedical Consumption in a Developing Country.” Policy Research Working Paper 3563, World Bank, Washington, DC. 40 CHAPTER 3 POOLING HEALTH FUNDS AND RISKS World Bank. 2003. Turkey: Reforming the Health Sector for Improved Access and Efficiency . Washington DC: World Bank. ______. 2012. “Georgia: Managing Expenditure Pressure for Sustainability and Growth.” Public Expenditure Review Report, World Bank, Washington, DC. Zeckhauser, R. 1970. “Medical Insurance: A Case Study of the Trade-Off between Risk Spreading and Appropriate Incentives.” Journal of Economic Theory 2: 10–26. 1 Of the 188 projects included in this evaluation, 42 supported decentralization. Their rating in the Independent Evaluation Group (IEG) review of the project implementation completion report is lower for monitoring and evaluation and for efficacy. 2Eligibility in 2005 was determined based on visits by local social workers and included criteria such as family size, disability, age, capacity to work, and income. 41 4. Purchasing Highlights  Purchasing is important for incentivizing quality and efficiency. Countries are introducing complex payment reforms that require improved data collection and analysis and management to address adverse effects, including those on nonhealth sectors.  A growing share of Bank operations are supporting countries in purchasing reforms and most of this support is to performance- and results-based payments in low-income settings. In addition, the Bank helped in building institutional and administrative capacity and investment in information to assess provider performance.  Service use has increased when countries move from line-item budgets to paying providers for activities or performance. A change in provider payment method primarily benefits individuals seeking care. Demand-side barriers, such as user fees, and high administrative costs remain concerns for efficiency and financial sustainability of Bank support.  Purchasing reforms are likely not sustained unless they are embedded in overall health financing, and the broader public finance context and future financing are assured. The objective of purchasing is for providers to deliver quality care efficiently to individuals who need it. Purchasing is challenged by the financial incentives of various provider payment methods that transfer funds from the purchaser (e.g., government units and insurers) to providers and by the paucity of information on providers’ reactions to these methods. Whether these incentives lead to the desired outcome heavily depends on the institutional context for providers and how they react to them. How did the World Bank Group support countries in purchasing, and what evidence is there for improved equity in financing and service use, financial protection, and efficiency? This chapter addresses these questions. It also includes findings from the country case studies presented in appendix E. Challenges Financial incentives set by the payment method may encourage providers to change the number of services, manage costs, and improve quality of care, all of which can affect efficiency (Ellis and McGuire 1996) (Table 4.1). Line-item budgets and hospital per diem are still common in middle- and low-income countries. But these two payment methods do not set incentives for providers to become more efficient or offer better care. To increase the number of health services, governments and 42 CHAPTER 4 PURCHASING insurers increasingly pay providers based on their activities including through fee- for-service and Diagnosis-Related Groups (DRGs) which reimburse hospitals a fixed amount per patient depending on the diagnosis. In the U.S. Medicare system the average length of hospital stay fell by 15 percent in the first three years after the shift to DRGs (Cashin et al. 2005). But activity-based payment has cost implications for the payer. In the Czech Republic, the move from line-item budgets to fee for service led to an increase in activities and 46 percent growth in hospital expenditures from 1992 to 1995 (Langenbrunner et al. 2005). Table 4.1. Provider Payment Methods and Related Incentives and Challenges Payment Methods Financial Incentives to Providers Challenges Line-item budget Increase number of input factors (e.g., bed, staff) Low productivity and use full budget Fee for service Increase number of incentivized activities (e.g., Inefficient service use; cost Activity-based Per diem for hospital services per patient, hospital days, admissions, increase day cases treated) Case based (e.g., DRG) Pay for performance Increase number of services leading to improved Transparency of information quality or efficiency on performance “Code creep” (distortion of treatment toward Inefficient service use; cost those with higher payment) increase Capitation Treat patients within budget Substandard quality Global budget Exclude high-risk patients Source: Langenbrunner et al. (2009). Prospective payment—including capitation based on the number of individuals registered with the provider and global budgets to hospitals to provide a set of services—shifts the financial risk to the providers, setting an incentive to increase efficiency. In Ireland the move to capitation led to a decline of 20 percent in the number of outpatient visits (Langenbrunner et al. 2005). In the worst case, providers reduce their costs by skimping care or discourage individuals with costlier health problems from registering. Mixed payment methods are frequently used, such as capitation adjusted by some activity indicator (e.g., number or coverage of preventive services or quality). To improve quality and efficiency, some countries including the United Kingdom and the United States have introduced performance payment to compensate providers for meeting preset quality and efficiency measures.1 The evidence base linking performance-based payments to better quality of care is thin in Organisation for Economic Co-operation and Development (OECD) countries (Box 4.1). Similarly, no systematic review shows the impact of performance-based payments in low- and 43 CHAPTER 4 PURCHASING middle-income countries. A review of nine impact evaluations finds that their effect on quality of care, antenatal care, institutional deliveries, preventive care for children, and outpatient visits is unclear. While performance-based payments increase facility revenues, their impact on efficiency is yet to be ascertained. The review also found little evidence that such payments have triggered for managerial autonomy in health facilities (Witter et al. 2012). Box 4.1. What Is the Effect of Performance-Based Payments in OECD Countries? The evidence base linking performance-based payments to better quality of care is thin. Most studies from the United States and United Kingdom show inconsistent efficacy or have revealed unintended effects, such as improved documentation without much change in quality of care (Epstein 2007). Maynard (2012) also finds that most studies were conducted without control groups and had methodological flaws. He finds that studies with control groups show modestly improved quality scores in participating health facilities, with the lowest improvements in the already highest performing hospitals, and that larger financial incentives produced greater effects than smaller incentives. Still, the financial incentive seems to diminish over time. U.S. hospitals reported gains for the first three years, but afterwards showed no difference in the performance of the two study groups. While the design and implementation costs of performance-based payment are considerable, none of the studies has conducted a relative cost-effectiveness analysis of these programs. Maynard (2012) concludes that the scale up of performance-based payments in OECD countries was made based on poorly designed, executed, and evaluated pay-for-performance programs, which may even raise costs and worsen efficiency. Although performance-based payment is often introduced alongside public reporting of performance results, few studies have identified whether improved performance stems from the performance-based payment itself or from the information on provider performance, which affects the reputation of the health care provider. One study suggests that the incremental effect of performance-based payment over public reporting of performance is small, around 3 percent performance improvement over two years, and varies according to baseline performance with the largest improvements observed among the poorest performing hospitals (Lindenauer et al. 2007). This finding has important cost implications given the high implementation costs of performance payment. In addition to these constraints in provider payments, governments are challenged by insufficient information, unclearly defined benefit packages, and institutional limitations when designing the purchasing function. Asymmetric information is a major constraint for activity- and performance-based payments, particularly in low- income settings. The difficulty is choosing appropriate measures and benchmarks, and collecting reliable and valid information on provider performance on the basis of which payments are made. But the purchaser—especially in countries with weak data systems—often has little information on how health care was delivered. Another problem is that in many countries the health care benefit package is nominally comprehensive, but in practice it is narrowly defined owing to provider and financial constraints. This means that patients continue to pay user fees for 44 CHAPTER 4 PURCHASING goods and services that are meant to be in the package and financed by the government or insurance. Institutional reforms support the effect of provider payments in public health facilities. While private providers can adjust their resources, managers in public facilities seldom have the autonomy to respond to the financial incentives set by the payment method and improve efficiency by adjusting the input mix, such as staff and medical supplies. Thus changing provider incentives needs to be accompanied by public sector reforms. Concerns arise if incentives cause adverse behavior among providers, as it may lead to cost shifting across different payers or to spillover effects in other sectors (e.g., payments may lead to wage increases in the health sector and put pressure on the government to increase wages in other sectors). Given the complexity of purchasing, purchasing needs to be fully integrated into the overall health financing and public finance context (Box 4.2). Unintended consequences such as spillover effects on other sectors need to be examined and addressed. Box 4.2. Integrated Approach to Purchasing An integrated approach consists of strategic purchasing, defining which interventions should be purchased in response to population needs, how they should be purchased, for whom, and from which providers (Figueras et al. 2005). IEG also finds that an integrated approach to purchasing considers the broader public sector context, including relevant institutional reforms needed to implement purchasing reforms, while attempting to foresee and forestall any potentially adverse effects. Bank Group Support to Purchasing An increasing share of Bank health financing projects supports governments and insurers in purchasing. Bank projects generally support health care providers and the purchaser who pays providers. In line with the Bank’s 2007 Health, Nutrition, and Population (HNP) strategy, the Bank has introduced a focus on results and better performance in health facilities. The Bank assisted governments and insurers with changing their provider payment methods. It also helped build institutions through information, monitoring and evaluation (M&E), and regulations to define benefits. In some countries, these operations also support the abolition or reduction of user fees paid by patients, including in Argentina, Benin, Burundi, Djibouti, Nigeria, Senegal, and Zimbabwe. 45 CHAPTER 4 PURCHASING ANALYTICAL WORK The Bank’s health teams have prepared limited analytical work on purchasing (Moreno-Serra and Wagstaff 2010); however, this body of work is growing as shown by an increasing number of impact evaluations on provider payment reforms supported by results-based financing projects (appendix Figure B.4). So far, the Bank’s impact evaluations have not analyzed the distributional effect of payment reforms and whether the poor benefit. M&E frameworks in Bank projects are presented in chapter 5. PROVIDER PAYMENT METHODS Some 60 percent of provider payment methods supported by Bank projects include a performance- or results-based component; project documents use these terms interchangeably (Figure 4.1). Most are introduced in health systems with automatic coverage in low-income settings and not by using the health insurer as the purchaser. A few Bank operations, mainly in Europe, supported DRG payment from health insurers to hospitals or some other case-based payment. Similarly, capitation payment adjusted by some activity and case-mix indicators are introduced mainly in Europe and Central Asia, Ghana, Latin American countries, and Vietnam. In Armenia, RBF is managed by the country’s single payer state health agency and is implemented nationally. In some countries such as Rwanda the Bank helped scale up performance-based payment developed by other donors.2 Figure 4.1. Provider Payment Methods Supported by Bank Projects, FY03–12 Capitation payment 27% Case-based Performance payment based 2% 51% DRG Results 7% Output based based 9% 4% Note: The total number of HNP projects with purchasing is 45. A more-detailed review of the Bank’s project documents categorizes Bank support in performance- or results-based payment along geographic lines. In reality, most of these payment methods are mixed payments that combine some aspects of activity- based payments (e.g., fee-for-service and case-based payment) with some quality 46 CHAPTER 4 PURCHASING indicators. More recently, this type of Bank support is implemented with the support of results-based financing (RBF) operations.  Afghanistan. The Bank supported the Ministry of Public Health to establish performance-based payment for contracted nongovernmental providers. Nongovernmental organizations (NGOs) are paid a capitation amount adjusted by scorecard indicators such as patient satisfaction and staff availability. The payment is a 1 percent bonus if the quality score improves. Bank support has recently transitioned to helping government paying providers a fee-for-service amount to increase the number of maternal and child health services.  Africa, including Benin, Burundi, Central African Republic, and Nigeria. In these countries the Bank supports governments in the introduction of mixed payment methods to providers consisting of fee-for-service and case-based payment for specific treatments or diagnoses, such as number of hypertension cases. This payment sets an incentive to increase the number of services for which a fee is paid and to diagnose more patients with diagnoses that have higher reimbursement. It can also have adverse effects if providers “up-code” patients and diagnose them with higher-priced diseases, leading to higher costs. The payment includes a quality component as measured in scorecards or during health supervision. These projects often include a demand-side component that abolishes or reduces user fees or provides vouchers for care to poor patients.3  Argentina and Brazil. Bank loans supported provider payment reforms and co-financed central government transfer payments to the health budgets of local governments (in addition to line-item health budgets) based on institutional performance in Brazil, and capitation adjusted by the achievement of 10 preventive care indicators in Argentina (Box 4.3). Payments encourage local governments in Brazil to invest in local administration, institutions, and fiduciary management in the health sector and in Argentina to ensure that providers have the resources to achieve the targets for preventive care indicators. Providers in both countries continue to receive line-item funding. In Argentina local governments also pay providers fee for service for preventive care to incentivize increased service provision (IEG 2011). An increasing number of Bank projects support a shift to performance- or results- based payments in national health systems. The majority of them are run with the support of the Bank’s results-based financing (RBF) program implemented through the Bank’s policy and investment lending programs. In some countries (e.g., Benin and Burundi), RBF funds are pooled with funds from the government and other 47 CHAPTER 4 PURCHASING donors and then transferred to providers. RBF is a cash payment or non-monetary transfer made to a national or sub-national government, manager, provider, payer, or consumer of health services after predefined results have been attained and verified. Payment is conditional on measurable actions being undertaken (Musgrove 2011). RBF operations thus directly influence the provider payment method. The Health Results Innovation Trust Fund4 cofinances Bank operations that finance health interventions in countries in the International Development Association. It also funds technical dialogue and learning related to RBF operations and program evaluations. The Trust Fund has a dedicated work program to monitor countries’ progress using performance data and to support country teams in their data work.5 Box 4.3. Financing Based on Performance or Results in Argentina and Brazil In Argentina the Bank loan disbursed an earmarked capitation amount to the government’s Plan Nacer based on the number of individuals registered in the provinces. Plan Nacer is a supply-side subsidy program that provides reproductive, maternal, and child health care in contracted public and private health facilities to uninsured children and women. The National Ministry of Health signed a performance contract with the provincial governments to transfer funds from the Bank loan and the central government budget to the provincial Plan Nacer based on the number of individuals registered (capitation) and on the results achieved on 10 health indicators. The capitation part of the transfer sets an incentive to provinces to increase the number of plan members; the second part an incentive to achieve treatment targets. The Brazilian central government, supported by the Bank, paid a bonus payment and a performance prize to municipalities for achieving explicit governance and fiduciary targets to improve management in primary care. The bonus payment was distributed as a lump sum to the 35 (of 188) municipalities that met the three criteria, and the performance prize was shared by 12 municipalities. Source: IEG (2011). INSTITUTIONS, BENEFIT PACKAGE, AND INFORMATION The Bank helps countries in purchasing their essential benefit package mainly from public sector providers and NGOs. In Afghanistan the Bank supported contracting of the basic package of cost-effective services from NGOs. In middle-income countries, Bank technical assistance advised insurers to purchase care strategically. In the former Yugoslav Republic of Macedonia primary care providers were privatized in 2007. There, the national health insurance fund, with Bank support, contracted the essential package from private providers and paid capitation adjusted by age, gender, region, and preventive care indicators (IEG 2013). Bank loans financed information systems in insurance and governments to manage beneficiary information. In low-income countries where the Bank supports performance- or results-based payments, substantial investments in information were made. Data 48 CHAPTER 4 PURCHASING allow for real-time monitoring and evaluation of provider performance and corrective actions.6 Effectiveness of Bank Group Support to Purchasing To understand the effectiveness of Bank and International Finance Corporation (IFC) support to purchasing, one must first ask whether that support was appropriate to country conditions. In particular, was the support accompanied by an assessment of financing needs, and where appropriate, financial assistance? Also, given the demanding managerial requirements of purchasing systems, did the Bank Group tailor its support to local capacity? INTEGRATING PURCHASING WITH HEALTH FINANCING AND PUBLIC FINANCE In several countries the Bank has taken a more single-track approach to purchasing without integrating payment reforms with other health financing functions. This approach is contrary to the multisector approach described in the HNP strategy as the Bank’s comparative advantage. IEG’s case studies found that in the Republic of Yemen the results-based payment approach supported by the IFC and the Bank for a narrowly defined disease area is not linked to any broader health financing efforts in revenue collection or risk pooling. In Kenya, although the Bank and the IFC are providing health financing support to strengthen the National Insurance Fund, a parallel RBF program is being piloted in one county that does not appear to be connected to the health insurance reform. In Rwanda discussions about RBF support did not involve any analysis on the use of the existing health insurers as the fund holder for RBF, on the effect of RBF in health facilities that are paid capitation or fee for service by health insurance, or on what the overall impact would be on health insurance finances and future cost trends. In Benin there is no coordination with public sector budget reform. RBF is almost exclusively linked to using revenues from household payments and drug sales, and is not integrated with fiscal transfers to facilities. In Tanzania an RBF pilot has been introduced separate from overall ongoing public sector and health financing reforms which raises concerns about its sustainability. In Ghana the design of the Bank’s RBF operation is not coordinated with the payment reforms introduced by the health insurer and fails to clearly demonstrate how it will address performance problems, given already sizable increases in health worker salaries. Where purchasing is embedded in broader public sector reforms it is more effective in establishing the relevant institutions that are needed to make it sustainable. In Rwanda the government changed the public finance law to allow for incentive payments for public sector workers; greater autonomy to facilities in regard to 49 CHAPTER 4 PURCHASING recruitment, deployment, and dismissal; and direct accountability for the performance of mayors. IEG found that the Bank provided substantial support through policy dialogue and general budget support, including legal changes and management reforms in health facilities. In other countries, however, purchasing support has had less profound effects on strengthening regulatory and management functions. The above examples suggest that the Bank could give more attention to integrate purchasing with broader health financing functions and public finance to build the necessary institutions that make payment reforms sustainable, as emphasized in the Bank’s 2007 HNP strategy. MANAGEMENT AND INFORMATION SYSTEMS Bank support has strengthened management and information systems. In Poland the case-based payment (supported by Bank policy lending) contributes to transparency and improved data availability in the social health insurance fund (Czach et al. 2011). Similarly, in Afghanistan, Argentina, and Egypt, among others, supervision of health facilities improved as did information and reporting systems and the validity of routine data. In Serbia the Bank conducted a baseline analysis when the government considered shifting from line-item budgets to capitation. The Bank recommended additional measures to prevent possible adverse effects under capitation, including consolidated pooling of funds from the insurer and other public sources to prevent fragmented incentives; a comprehensive capitation rate to cover salaries along with public sector reforms for public employees and provider autonomy; and M&E (World Bank 2009). But in Brazil a recent Bank project failed to introduce performance-based payment because it omitted to collect the relevant information to link disbursement to results or to provide for independent technical audits of data to compute disbursement indicators. Institutional strengthening contributed to transparency. Bank loans financed information-technology management systems in governments to monitor provider performance and in insurance companies to help process medical claims submitted by providers and monitor and assess their performance. In Argentina the Bank helped provinces invest in detailed data collection and analysis of performance indicators. Results were audited by an independent firm hired under the Bank loan, and provinces and providers were fined for incorrect data reporting (IEG 2011). In Benin, the Bank supported health strategic planning and information systems, and in Bolivia, performance agreements between the central government and the regional departments with clearly defined objectives and results. However, the contracting of international firms to carry out the verification of performance to pay providers has proved very costly in these countries. 50 CHAPTER 4 PURCHASING In some cases, the Bank did support public reporting of provider performance results with the aim of informing consumers and financers about better providers (as is done in some OECD countries; see Box 4.1). RBF programs have introduced online real time performance and financial reporting at the facility level including in Benin, Cameroon, Chad, Zambia, and elsewhere.7 The Bank did also help a few communities that were getting involved in monitoring and verification of the use of public health resources in Afghanistan, Benin, and Rwanda. IEG case studies in Nicaragua found that the Bank supported the government in bringing the community into the health sector by strengthening the use of locally collected data, including technical and social audits. The Bank in Kenya channels resources directly to health facilities and has contributed to improved reporting and accounting of revenues collected by facilities. Findings from Bank-supported research in Uganda suggest that community participation in monitoring providers can improve quality and quantity of health services as evidenced by large increases in utilization, significant weight gains for infants, and markedly lower death rates among children (Bjoerkman and Swensson 2007). AVAILABILITY OF CARE AND SERVICE USE Availability of care and service use has increased where countries moved from line- item budgets to activity- or performance-based payments, including where this was supported by the Bank. In Afghanistan the bonus payment has been associated with an increase in availability of skilled health workers and administrative personnel. In Egypt performance-based payment positively affected provider behaviors toward patients, reduced staff turnover, and helped lift medical encounters from three to 16 per day. In Argentina a 1 percentage point provincial health budget increase, combined with a results focus for local authorities, contributed to higher utilization rates among low-income groups who seek care in public health facilities (IEG 2011). Few studies use control groups to compare the effect of performance payment against other payment changes such as higher salaries independent of performance, or against intrinsic factors. In an impact evaluation in Rwanda, the Bank compared higher budget funding for health facilities in control districts with performance- based payments for selected services in pilot districts. It found that the payments mainly increased utilization of services that had higher unit payments and that providers could more easily control for. There was no impact on other rewarded services such as childhood immunization, malaria prophylaxis, or curative care visits for children (Basinga et al. 2011). Distributional effects across the insured and uninsured and between socioeconomic groups were not identified. Bank research also highlighted the importance of altruistic concerns that drive the behavior of health care providers. In Uganda a Bank team found that more funding to religious 51 CHAPTER 4 PURCHASING health facilities in 1999 was passed on to patients in the form of more diagnostic services and lower user fees, which reduced financial barriers; however, public facilities performed less well (Reinikka and Svensson 2003). While the Bank did not analyze the payment effect over the insurance effect on service use, other researchers find that insured individuals in Rwanda report large and significant improvements in several rewarded and unrewarded services and a decrease in child anemia prevalence (Sherry et al. 2012). Similarly, Skiles et al. (2012) report increased service use across all wealth quintiles and insurance as a positive predictor for service use. These findings suggest that it is mainly the care-seeking insured who benefit from payment reforms. It also shows that much still needs to be learned about how to combine payment reforms with risk pooling to address barriers in access to care for patients, such as user payments. FINANCIAL PROTECTION As a change in provider payment method primarily benefits individuals seeking care (e.g., the insured and wealthier), provider payment reforms without measures to reduce user payments and improve risk pooling are unlikely to improve equity in service use and financial protection. So far, the Bank’s impact evaluations have not analyzed the distributional effect of payment reforms and whether the poor benefit or the uninsured. HEALTH OUTCOMES The evidence linking payment reforms to better health outcomes is thin and mixed. Sherry et al. (2012) used 2005 and 2007 demographic and health survey data in Rwanda and found mixed results and no significant impact of performance-based payment (which was supported by Bank policy lending) on maternal and child health outcome indicators. However, one Bank study found (based on household survey data comparing households in 10 districts with the payment reform and 9 districts with traditional input-based financing) that performance payment led to large and significant improvements in child health between 2006 when the payment was introduced and 2008, and the payment is more effective among higher-skilled providers (Gertler and Vermeersch 2012). Distributional effects in health outcomes across the insured and uninsured and between socioeconomic groups were not identified. EFFICIENCY Insurers and government entities can purchase care strategically to ensure the efficient use of health funds. There is some indication that Bank support contributed to improved efficiency where pooling of public funding increased to purchase a 52 CHAPTER 4 PURCHASING benefit package with cost-effective interventions. In Afghanistan, efficiency is considered to have been improved by the performance-based NGO contracting model because access to the contracted services increased, and those services in the package are the most cost-effective interventions available for improving overall health outcomes. In Armenia spending shifted from hospitals, which had to reduce bed overcapacity, to more cost-effective primary health care. In Serbia the Bank helped streamline benefit packages where social health insurers had committed to cover overly generous benefits. The number of staff was rationalized in the health sector in Croatia and Serbia, and the Bank recommended increasing wages for primary care workers in Tajikistan. Strong government support was essential in introducing these efficiency enhancing measures in these countries and supporting the effectiveness of purchasing. However, adverse effects of activity- and performance-based payment reforms on sector efficiency were insufficiently examined in Bank analysis. Only in a few countries (e.g., Serbia) did the Bank conduct cost and productivity analysis to identify the payment amount. Nor did Bank analysis sufficiently examine adverse reactions by providers to the payment. A case study prepared by IEG for this evaluation found that in Bolivia, fee-for-service payment for mothers and children created an incentive for beneficiaries to seek costlier tertiary care, contributing to inefficiency. In Ghana fee-for-service payment leads providers to refer patients to hospitals or clinics for more expensive treatment. It has also resulted in lengthy processing times and prolonged the reimbursement time to providers to five months. In Vietnam moving from fee-for-service to capitation payments for hospitals contributed to efficiency gains by reducing recurrent expenditures and did not have a negative effect on health outcomes. However, a recent Bank-supported study warns of adverse spillover effects because hospitals are shifting costs across different payers from insurance paying capitation to the uninsured who pay user fees, which may have implications for access and financial protection (Nguyen et al. 2013). A ruse in Rwanda was used by hospital pharmacies to prevent drugs from running out of stock—a performance indicator—by refusing to dispense the last box of pharmaceuticals (Kalk et al. 2010). The Bank could analyze adverse effects of payment reforms systematically and help countries addressing them. FINANCIAL SUSTAINABILITY Purchasing reforms are not sustainable unless they are embedded in overall health financing and the broader public finance context and future financing is assured. Mainly in middle-income countries, the Bank supported purchasing reforms in social health insurance and cautioned governments about the financial consequences. A small body of Bank analytical work looked at the financial 53 CHAPTER 4 PURCHASING sustainability of insurance and advised governments on streamlining generously defined benefits (Chawla 2007; La Forgia and Nagpal 2012; Langenbrunner and Somanathan 2011; Preker et al. 2013; Smith and Nguyen 2013). In the Europe and Central Asia Region, the Bank warned that social health insurers are simple disbursement agents who create a high financial risk for the government budget and recommended strategic purchasing (Chawla 2007). The Bank also recommended selective contracting with providers and shifting to capitation and case-based payment with broader health sector reforms (Langenbrunner et al. 2009; World Bank 2010). Where performance and activity payment was introduced through the government budget, the Bank did not give sufficient attention to the financial sustainability of this support. Many RBF projects are pilots that aim to build evidence on the impact and cost effectiveness of the intervention. Still, administrative costs and the financial implications for the payer are major concerns when introducing activity- and performance-based payments, which the Bank did not address sufficiently. In Kenya verification costs for performance payment are estimated at 20 percent of the performance budget. Governments have not assumed financing responsibility in the recurrent budget for the cost of performance- or results-based payment programs, and so even programs considered effective have not been taken over by governments. In Egypt the government scaled up the Family Health Model but not the performance-payment component after donor funding ended, even though it was perceived as successful. In Ghana plans are going ahead with a new RBF program supported by the Bank in parallel to the provider payment reforms of the National Health Insurance System, but because of financial sustainability concerns, the program will need to be financed by the Bank or donors. In Argentina cofinancing of the Plan Nacer by provincial governments encountered long delays (IEG 2011). In low-income countries, RBF is a separate budget line and mainly financed by donors from sector budget support. In Rwanda challenges of sustainability came to the fore as donor funding was scaled back (Ministry of Finance and Planning 2011). The Bank did not examine how a change to the provider payment in the public sector affects the governments wage bill, yet this may have substantial financial implications for the government. A case study prepared by IEG for this evaluation found that the Rwandan government allocated about 10 percent of the domestic health budget to performance payments in 2010, which were used by health facilities as they saw fit, including topping up salaries and improving the facilities (Ministry of Finance and Planning 2011). In the first two years, performance-based pay increased facility budgets by 22 percent on average, most of which (77 percent) was paid as a salary top-up, resulting in an average 38 percent salary increase for staff 54 CHAPTER 4 PURCHASING (Basinga et al. 2011). In absolute terms, results-based financing helped lift health workers’ salaries by $75–750 a month depending on their function and facility performance (Kalk et al. 2010). This is considerably higher take-home pay for health workers than for other public employees, such as teachers, and can pressure governments to increase other public sector wages. However, in none of the countries did Bank teams analyze the public sector wage impact of results-based financing. As in OECD countries, decisions to scale up payment reforms were not always based on lessons from pilots, which may affect sustainability. Most of the Bank’s performance- or results-payment reforms supported by RBF programs were piloted. However, decisions were made to scale up regardless of weak, inconclusive, or incomplete pilot results. In Benin several failed RBF pilots introduced in 2007 were redesigned and reintroduced with a grant from the Bank’s Health Results Innovation Trust Fund. The evaluation of the pilot supported by this fund is not yet complete. Nonetheless, plans have been made to scale up RBF to all 34 health zones. Similarly, in Tanzania, despite multiple failed RBF programs and before the evaluation of the most recent pilot becomes available, the Bank and other donors have voiced interest in supporting a scale up to a national RBF scheme. This is a concern as early findings from an evaluation by the Ifakara Health Institute suggests that the RBF programs are burdensome, incompletely understood, unevenly implemented, not particularly effective, and unlikely to be sustainable (Chimhutu et al. 2014). In Argentina the Bank’s impact evaluation (Gertler et al. 2011) was not ready to inform the scale up of Plan Nacer nationwide. Instead, the government decided to scale up based on routine administrative data and qualitative analysis (IEG 2011). The IEG found that in some countries, the Bank supported additional payments and transfers with questionable sustainability. In Rwanda, the Ministry of Finance raised concerns over the sustainability of the Bank-supported payment program to community health workers who are volunteer members of a cooperative (Ministry of Finance and Planning 2011). Under a former Bank project, pregnant women also received a baby kit if they delivered in health facilities or an umbrella if they have four antenatal care visits—but these in-kind transfers were stopped due to high procurement costs. Again, these examples emphasize the importance of integrating purchasing with broader health financing and institutional reforms to support sustainability. In sum, countries are introducing complex payment reforms to incentivize providers to improve quality and efficiency. A growing share of Bank operations are supporting countries in these purchasing reforms, and most of this support is 55 CHAPTER 4 PURCHASING through RBF in low-income settings. In addition, the Bank helped in building institutional and administrative capacity and investment in information to assess provider performance. IEG’s country case studies found that where Bank purchasing support was integrated with other health financing functions (risk pooling and revenue collection) and linked to public sector reforms rather than limited to narrowly defined provider payment methods, it has been relatively more effective because it addressed broader institutional reforms which in turn support sustainability. The availability of care and service use improved in countries where the Bank helped introduce activity- and performance payment often with RBF support. However, limited evidence from impact evaluations with control group points to broadly similar effects for performance-based payment and budget increases for health. Moreover, performance-based payment systems have high overhead costs for performance verification. As a change in provider payment method primarily benefits individuals seeking care (e.g., the insured and wealthier), Bank-supported provider payment reforms without measures to reduce out-of- pocket user payments and improve risk pooling are unlikely to improve equity in service use and financial protection. Payment reforms are likely not sustained unless they are embedded in overall health financing and the broader public finance context. References Basinga, P., P.J. Gertler, A. Binagwaho, A.L. Soucat, J. Sturdy, and C.M. Vermeersch. 2011. “Effect on Maternal and Child Health Services in Rwanda of Payment to Primary Health-Care Providers for Performance: An Impact Evaluation.” The Lancet 377 (9775): 1421–1428. Bjoerkman, Martina, and Jakob Swensson. 2007. “Power to the People: Evidence from a Randomized Filed Experiment of a Community-Based Monitoring Project in Uganda.” Policy Research Working Paper 4268, World Bank, Washington, DC. Cashin, Cheryl, S. O’Dougherty, Y. Samyshkin, A. Katsaga, A. Ibraimova, Y. Kutanov, K. Lyachshuk, and O. Zuys. 2005. “Case-Based Hospital Payment Systems: A Step-by-Step Guide for Design and Implementation in Low- and Middle-Income Countries.” Consultant report for the U.S. Agency for International Development, Abt Associates Inc., Bethesda, Md. Chawla, Mukesh. 2007. Health Care Spending in the New EU Member States: Controlling Costs and Improving Quality. Washington, DC: World Bank. https://openknowledge.worldbank.org/handle/10986/6771 License: CC BY 3.0 IGO. Chimhutu, Victor, Ida Lindkvist, and Siri Lange. 2014. “When Incentives Work Too Well: Locally Implemented Pay for Performance (P4P) and Adverse Sanctions Toward Home Birth in Tanzania—A Qualitative Study.” BMC Health Services Research 14: 23. Czach, Katarzyna, Katarzyna Klonowska, Maria Swiderek, and Katarzyna Wiktorzak. 2011. “Polish Experiences with DRGs, in Diagnosis-Related Groups in Europe.” In European Observatory on Health Systems and Policies, edited by ReinhardBusse, Alexander Geissler, Wilm Quentin, and Miriam Wiley. New York City: McGraw–Hill International. 56 CHAPTER 4 PURCHASING Ellis, Randall P., and Thomas G. McGuire. 1996. “Hospital Response to Prospective Payment: Moral Hazard, Selection, and Practice-Style Effects.” Journal of Health Economics 15: 257–277. Epstein, A.M. 2007. “Pay for Performance at the Tipping Point.” New England Journal of Medicine 356: 515-517. Gertler, Paul, Martinez Sebastian, and Pablo Celhay. 2011. “Impact of Plan Nacer on the Use of Service and Health Outcomes: Intermediate Results Using Administrative Data from Misiones and Tucuman Provinces.” Unpublished manuscript, World Bank, Washington, DC. Gertler, P., and C. Vermeersch. 2012. “Using Performance Incentives to Improve Medical Care Productivity and Health Outcomes.” Policy Research Working Paper 6100, Impact Evaluation Series No. 60, World Bank, Washington, DC. IEG (Independent Evaluation Group). 2011. “Results-Based Health Programs in Argentina and Brazil: Project Performance Assessment Report.” Report No. 62571, World Bank, Washington, DC. ______. 2013. “Former Yugoslav Republic of Management Project (P086670). ” Project Performance Assessment Report 72810, World Bank, Washington, DC Macedonia Health Sector. Kalk, A., F.A. Paul, and E. Grabosch. 2010. “Paying for Performance in Rwanda: Does It Pay Off?” Tropical Medicine & International Health 15 (2): 182–190. La Forgia, Gerard, and Somil Nagpal. 2012. Government-Sponsored Health Insurance in India. Are You Covered? Washington, DC: World Bank. Langenbrunner, Jack, and Aparnaa Somanathan. 2011. Financing Health Care in East Asia and Pacific. Washington, DC: World Bank. Langenbrunner, Jack, Cheryl Cashin, and Sheila O’Dougherty. 2009. Designing and Implementing Health Care Provider Payment Systems. How-To Manuals. Washington, DC: World Bank and USAID (U.S. Agency for International Development). Langenbrunner, Jack, Eva Orosz, Joseph Kutzin, and Miriam Wiley. 2005. “Purchasing and Paying Providers.” In Purchasing to Improve Health Systems Performance, edited by Josep Figueras, Ray Robinson, and Elke Jakubowski. Maidenhead, UK: Open University Press, McGraw-Hill Education. Lindenauer, P. K., D. Remus, S. Roman, M.B. Rothberg, E.M. Benjamin, A. Ma, and D. W. Bratzler. 2007. “Public Reporting and Pay for Performance in Hospital Quality Improvement. ” N Engl J Med. 356: 486–496. Maynard, Alan. 2012. “The Powers and Pitfalls of Payment for Performance.” Health Economics 21:3– 12. Moreno-Serra, Rodrigo, and Adam Wagstaff. 2010. “System-Wide Impacts of Hospital Payment Reforms: Evidence from Central and Eastern Europe and Central Asia. ” Health Economics 29 (4): 585–602. Musgrove, Philip. 2011. “Financial and Other Rewards for Good Performance or Results: A Guided Tour of Concepts and Terms and a Short Glossary,” World Bank, Washington, DC. http://www.rbfhealth.org/system/files/RBFglossarylongrevised.pdf Nguyen, Ha, Thi Hong, Sarah Bayles, Adam Wagstaff, and Huyen Do. 2013. “Getting Incentives Right: An Impact Evaluation of District Hospital Capitation Payment in Vietnam. ” Policy Research Working Paper 6709, World Bank, Washington, DC. Preker, Alexander S., Marianne E. Lindner, Dov Chernichovsky, and Onno P. Schellekens. 2013. Scaling Up Affordable Health Insurance: Staying the Course. Washington, DC: World Bank. 57 CHAPTER 4 PURCHASING Reinikka, Ritva, and Jakob Svensson. 2003. “Working for God? Evaluating Service Delivery of Religious Not-for-Profit Health Care Providers in Uganda.” Policy Research Working Paper, World Bank, Washington, DC. Sherry, T., S. Bauhoff, and M. Mohanan. 2012. “Paying for Performance When Health Care Production Is Multi-Dimensional: The Impact of Rwanda’s National Program on Rewarded Services, Multitasking and Health Outcomes.” Economic Research Initiatives at Duke (ERID) Working Paper 136, Duke University, Durham, NC. Skiles, M. P., S.L. Curtis, P. Basinga, and G. Angeles. 2012. “An Equity Analysis of Performance-Based Financing in Rwanda: Are Services Reaching The Poorest Women? ” Health Policy and Planning 28 (8): 825. doi:10.1093/heapol/czs122. Smith, Owen, and Son Nam Nguyen. 2013. Getting Better. Improving Health System Outcomes in Europe and Central Asia. Washington, DC: World Bank. Witter, S., A. Fretheim, F. L. Kessy, and A.K. Lindahl. 2012. “Paying for Performance to Improve the Delivery of Health Interventions in Low- and Middle–Income Countries.” Cochrane Database of Systematic Reviews, Issue 2. Art. No.: CD007899. doi: 10.1002/14651858.CD007899.pub2. World Bank. 2009. “Serbia: Baseline Survey on Cost and Efficiency in Primary Health Care Centers before Provider Payment Reforms.” Report No. 45111-YF, World Bank, Washington, DC. ______. 2010. “Health Provider Payment Reforms in China: What International Experience Tells Us.” China Health Policy Notes No. 58414 V2, World Bank, Washington, DC. 1The underlying rationale for performance payment is that quality varies across health facilities because providers deliver care differently. Reducing this variation by setting financial incentives was expected to increase quality and productivity of the health system. Performance-based payment methods differ substantially, however, and reflect local conditions including information technology, data availability, and providers’ willingness to participate (Maynard 2012). 2Hospitals are paid on their performance in 52 quality indicators assessed quarterly. Health centers receive payments based on 24 indicators on service delivery measured monthly. Hospitals and health centers also receive a provider payment, including fee-for-service and case-based payments and capitation, from the social health insurance fund and from community-based health insurance. 3In Burundi, results-based financing (RBF) is a national program, and the Bank ’s funds are pooled with other donors and government funding to finance RBF and user fee abolition. In Senegal and Zimbabwe, poor women receive vouchers to seek care. The Nigeria health project introduces exemption policies for the poor. Most of these projects are still in the design phase or early implementation. 4This is financed by the governments of Norway and the United Kingdom. The number of HNP projects with cofinancing from this fund increased from three in 2007 to nine in 2013 for a total committed amount of $260 million. 5The trust fund supporting the RBF operations has a requirement that each country program is paired with a rigorous impact evaluation and is accompanied by a well-funded impact evaluation grant. In parallel with the requirement, extensive technical support is provided to the country teams to assure the rigor of such evaluations. At present, the portfolio consists of 34 impact evaluations and eight program and process assessments. 6 All RBF programs collect and verify performance data which allows for real-time monitoring of performance and necessary corrective actions. Most programs have an extensive database which stores administrative data and allows for a close examination of performance down to the facility level. Data collected for the most part are integrated in the countries Health Management and 58 CHAPTER 4 PURCHASING Information Systems (HMIS) and beyond that also contain information on quality of care. Some countries, like Zambia and Zimbabwe, make use of HMIS data to monitor possible negative spill-over effects of RBF on nonincentivized services. The Bank team designed an RBF module for ADePT to make it easier and faster to analyze data and focus on results (http://www.worldbank.org/adept/) Discussions with the World Bank Institute’s Innovation Lab explore options to show cross country RBF data on the web. 7RBF programs introduced online performance dashboards to show real-time performance and amount spent. This information is public. Examples of the online dashboard can be found at http://www.beninfbr.org/ (Benin), http://www.fbpsanteburundi.bi/ (Burundi), http://www.fbrcameroun.org/ (Cameroon), http://www.fbrtchad.org/ (Chad), http://cd.thenewtechs.com/ (Democratic Republic of Congo), https://nphcda.thenewtechs.com/ (Nigeria), and http://www.rbfzambia.gov.zm/ (Zambia). 59 5. Factors in Successful Bank Group Support Highlights  Government commitment to health financing reforms is influenced by political and fiscal constraints. The World Bank Group can reinforce commitment by building technical capacity but needs to be flexible and able to adjust to the local political and technical context.  Bank analytical work has informed the international health financing dialogue and could be expanded to help institutionalize health financing reforms and build local capacity.  To fully use its capabilities in health financing, the Bank Group should draw on the expertise from health staff jointly with public finance and fiscal experts and work across the new Global Practices.  The Bank’s health financing portfolio is changing and focusing on one subintervention — performance-payment reforms, which is increasingly supported by results-based financing operations. Purchasing needs to be integrated with other health financing functions and public finance to be sustainable as outlined in the 2007 strategy for Health, Nutrition, and Population.  This evaluation may be missing some successful Bank engagement in health financing because of weak monitoring and evaluation (M&E) in health projects. Learning from the Bank’s rich country experience is constrained by weak M&E in health projects. The common success factors seen in the previous chapters that would make for good engagement in revenue collection, risk pooling, and purchasing include:  Government commitment and technical and information capacity;  Depth and relevance in analytical work;  Capabilities and collaboration;  Integration of all health financing functions; and  Sound monitoring and evaluation (M&E). Chapter 5 discusses how these factors cut across the three pillars of health financing and influence the effectiveness of Bank Group support and the implementation of the Bank’s Health, Nutrition, and Population (HNP) strategy. Chapter 2 showed that in governments that are committed to improving health outcomes and increasing support to the sector, efficient and equitable revenue instruments need to be used, taking into account the overall public finance context. Similarly, to address the many challenges in risk pooling described in chapter 3, strong institutions, management, and technical capacity are needed as well as information capacity to manage pooled funds. Bank analytical work helps inform governments in these health financing decisions. As seen in chapter 4, purchasing depends on integration with risk pooling to address financial barriers and thus revenue collection and with public finance. Purchasing depends heavily on information systems, M&E, technical capacity to analyze performance and define payments, and government commitment to 60 CHAPTER 5 FACTORS IN SUCCESSFUL BANK GROUP SUPPORT institutional reforms to allow providers to respond to financial incentives. The Bank’s 2007 HNP strategy—chapter 1—sees health financing as a comparative advantage for the Bank because of its analytical capacity and multisector nature. Chapter 5 also contains lessons learned from the country case studies conducted for the evaluation (appendix E). Government Commitment, Technical and Information Capacity, and Flexibility Whether Bank support to health financing reforms is sustained depends on government commitments to allocate revenues to health, execute health budgets, address inequity in health financing, and introduce institutional reforms, including passing legislation, changing provider autonomy to allow responses to financial incentives, addressing fragmentation, and linking health financing to broader public sector reforms. Governments are also required to invest in information to document the flow of funds in the sector, build technical capacity to analyze information collected on finances and performance, and address adverse effects. Mounting political commitment by governments has ensured important health financing reforms in countries such as Afghanistan—where nongovernmental organizations (NGOs) were contracted by the government—and Ghana, Rwanda, and Turkey, which introduced impressive health financing reforms. Yet insufficient financial commitment has limited reform sustainability. Ghana has too little fiscal space to subsidize insurance coverage for the poor, leading the health insurer to use a 1.7 percent poverty rate rather than the true rate of 30 percent. Weak technical and information capacities in Tunisia stymie coverage for the poor under the Free Medical Assistance Program (IEG 2014). Government commitment is also affected by political, not just financial and technical constraints. In Bolivia, Bank lending and analytical work guided much of the policy discussion in health financing reform. However, since the government changed in 2006, the Bank has not been considered a technical partner in health financing. Similarly, in Egypt the Bank was engaged through lending and analytical support in consolidating insurance reform and the family health model, which was widely considered coherent and innovative. But discord within the government over the reform led to implementation difficulties, and the Bank has since found it hard to reengage. In Rwanda, Bank support was conducted through policy dialogue and general budget support, which addressed a wide spectrum of health financing reforms. But in 2009, the government decided under its division of labor policy that the Bank should redirect funds to other sectors. The Bank has since been absent from the health financing policy dialogue there. Although the Bank provided extensive 61 CHAPTER 5 FACTORS IN SUCCESSFUL BANK GROUP SUPPORT technical assistance to drafting the health financing policy in Rwanda, after the Bank’s health financing team left, momentum drained and the draft policy has yet to be finalized (IHP+ 2012). In Kenya, by contrast, the relationship with the government has improved markedly since 2009. The Bank is now engaged in an array of health financing activities, including analytical work on the level and allocation of fund, and the government has provided additional $50 million in FY 13–14 to compensate providers for free maternity care. The International Finance Corporation (IFC) has convened the stakeholder dialogue and supports management reforms in the public health insurer. In addition, the Bank financial collaboration with health facilities has contributed to improved reporting and accounting of revenues in Kenya. Technical capacity facilitated understanding for health financing reforms. The Bank’s HNP regional departments and the World Bank Institute (WBI) are building local technical capacity that helps facilitate dialogue on key health financing issues with governments. Country and donor representatives interviewed by the Independent Evaluation Group (IEG) indicated that WBI courses have improved their understanding of the technical and policy aspects of health financing reform. For this evaluation, the IEG conducted an online survey among graduates of the WBI flagship course. Almost three-quarters of the 109 participants who completed the survey said they had a chance to implement what they had learned during the course. More than 80 percent used the literature from the course in their work, and more than half of them said the course had a positive impact on their collaboration with Bank staff. The WBI should consider tracer surveys among future graduates to ensure that its courses build technical capacity for health financing (appendix A). In some countries, including Brazil, Bank support to building the institutional foundations at municipality level contributed to the timely execution of the health budget. Bank support to consumer information and information technology has created transparency in insurance management in Vietnam and elsewhere. The Bank could also help more countries institutionalize standardized methods commonly used in Organisation for Economic Co-operation and Development countries, such as National Health Accounts to track the flow of funds in the sector. Standardized government M&E helped inform Public Expenditure Reviews (PERs) in several countries, built local technical capacity, and created a better understanding for health financing reforms. In some cases, Bank support needs to be more flexible and adjust to the local political and technical context. In a low-capacity environment like Afghanistan, the Bank demonstrated such flexibility. Given the limited ability of the Ministry of Health to provide or purchase services, the Bank supported contracting with NGOs 62 CHAPTER 5 FACTORS IN SUCCESSFUL BANK GROUP SUPPORT and provided large-scale assistance. Similarly, in Benin, Bank support to pooling and purchasing was calibrated to the local context. In Tanzania, however, other donors viewed Bank financing as inflexible as it proved hard to finance one of the nine option papers on fiscal space to feed into a larger health financing strategy paper. Beyond flexibility, in some countries including Egypt, the Bank would have been more useful had it possessed a better understanding of the political economy. Depth and Relevance of Analytical Work The Bank has a unique ability to connect operational work with research and evaluation to inform policy making through its knowledge products. It has provided analytical support to health financing reforms including those for fiscal space, PERs, insurance analysis, and impact evaluations on insurance and results-based financing. Through this work, the Bank maintained a policy dialogue with governments that contributed to informing health financing reforms in countries such as Afghanistan, Ghana, Mexico, Kenya, Poland, Rwanda, Turkey, and Vietnam—and elsewhere. The Bank has launched a growing number of impact evaluations to examine the effect of health financing reforms. The World Bank Development Impact Evaluation database in the Development Economics Group had 178 health-related impact evaluations in 2013. Of these, 14 are on health financing and complete and available (appendix Table A.8). Most of them examine health insurance in China and Vietnam. Results-based financing has become the most frequently researched topic among the ongoing evaluations (appendix Figure B.4). None of these impact evaluations, however, includes a cost-benefit analysis. Another drawback is that the Development Impact Evaluation database is not comprehensive and may have missed some impact evaluations of Bank-supported reforms (e.g., impact evaluations financed under a Bank project or conducted by non-Bank researchers). Through health financing workshops, the Bank promotes international dialogue. It also adds value by creating and maintaining the global health databases with health financing indicators. International health financing experts and stakeholders in client countries interviewed agree that the Bank adds value through its knowledge work on health financing, health financing analysis in PERs, poverty assessments, fiscal space analyses, and a growing body of impact evaluations. They consider these reports of high quality and very useful. However, the Bank does not have a central registry that would make these studies or workshops easily accessible.1 63 CHAPTER 5 FACTORS IN SUCCESSFUL BANK GROUP SUPPORT The Bank has conducted an array of analytical work on health financing, including in sector analysis, PERs, and fiscal space studies. However, these reports do not necessarily examine the poverty and equity effect of health financing. The Bank could therefore deepen analysis on health financing in its poverty assessments. In 2007–2012, few poverty assessments with a health chapter looked beyond epidemiological changes to examine the poverty impact of health financing. Twenty of the 43 poverty assessments assess the country’s health financing situation (appendix Table A.5), but their approach varies greatly (Table 5.1). Ten included benefit incidence analysis, and five presented out-of-pocket spending as a share of total health expenditures. Few analyzed utilization combined with out-of-pocket spending. Health insurance enrollment among the poor is reported in 10 reports, but they did not analyze whether insurance improves utilization of care or protects the insured against catastrophic spending or falling into poverty. Two assessments reported on impoverishment from health shocks (Azerbaijan and Georgia) and compared household income before and after health payments. Table 5.1. Poverty Assessments with Relevant Health Financing Analysis, 2007–2012 Indicators used in poverty assessments Country of poverty assessment Azerbaijan, Bangladesh, Chad, Ghana, Indonesia, Kenya, Benefit incidence analysis Nicaragua, Paraguay, Senegal, Zambia OOP expenditure in health as percentage of Azerbaijan; Georgia; Tajikistan; Uzbekistan; Venezuela, RB total health expenditure OOP expenditure in health as percentage of Azerbaijan, China, Georgia, Iraq, Nicaragua, Tajikistan, total household expenditure Uzbekistan Percentage of lowest income quintile Argentina; China; Georgia; Ghana; Indonesia; households participating in risk-pooling Macedonia, FYR; Nicaragua; Paraguay; Russian Federation; schemes Venezuela, RB Percentage of households with catastrophic Azerbaijan, Georgia health expenditures Percentage of population falling below the Azerbaijan, Georgia poverty line because of illness Severity of poverty because of OOP Azerbaijan, Georgia expenditures Note: For Kenya see www.worldbank.org/SDI; OOP = out-of-pocket user payments. As in PERs, the Bank’s poverty assessments could usefully follow a common methodology to analyze the impact of health financing for households. This methodology could be developed based on the experience from the Azerbaijan and Georgia reports, and be integrated with the new Systematic Country Diagnostic framework which will focus on achieving the twin goals of reduced poverty and shared prosperity. Bank analytical work should be made easily accessible on the Internet. 64 CHAPTER 5 FACTORS IN SUCCESSFUL BANK GROUP SUPPORT Capabilities and Collaboration in Health Financing Health financing requires a different skill set from that of the general health specialist. Since 2007, the number of Bank staff affiliated with the HNP sector has increased slightly. However, the share of economists among HNP staff remained at 19 percent, and the economist team became more junior as suggested by a decreasing number of lead economists working in HNP (from nine in 2007 to two in July 2013). The number of senior economists doubled from 14 to 29 in 2013. If the Bank is to be a major player in health financing, it has to staff accordingly. The Bank’s comparative advantage lies in HNP’s ability to collaborate with the Public Sector and Macro and Fiscal Management teams and facilitate dialogue on health financing at all government levels, including the Ministry of Finance as outlined in the HNP strategy. To fully use its capabilities, the Bank Group could use a multisector team that draws on the expertise from Health and other sector experts and works across the new Global Practices and the IFC. Most IEG country cases found there was collaboration between the HNP hub and the Regions but limited cross-support from the Development Research Group2 or the Human Development chief economist, which is also confirmed in the HNP staff survey. Collaboration with the IFC was limited to the Health in Africa and India initiatives. Within the jointly established Health in Africa Initiative, an external mid-term review found that the Bank Group did not leverage synergies within the group and “operated without complementarity” (Brad Herbert Associates 2012). In several countries, including Argentina and more recently in Kenya, this collaboration has worked well and has improved results. Collaboration is essential in a broad-based systems approach to link health financing reforms (in purchasing, pooling, taxation, and user fees) to public sector reforms and lead the health financing dialogue at all government levels. However, the evaluation’s country case studies and the international health financing experts noted that the Bank does not often exert this leadership role. The Bank’s Health and Public Sector teams could enhance collaboration to fully embed health financing in broader public sector reforms. In Vietnam, health support has not been a significant part of overall public sector reform in the past, and a need to enhance coordination was emphasized in the IEG case study by the HNP and Poverty Reduction and Economic Management teams, with a view on macro-level issues of health financing reform such as fiscal space, costing of coverage, and affordability and sustainability of reforms. Similarly, in Benin reforms were intra-health focused without drawing enough on expertise on how to create fiscal space or considering longer-term implications of fiscal sustainability. In Tanzania the health team was involved in government-wide public financial management reform to track resource allocations to decentralized levels of government. Still, an attempt was made (without 65 CHAPTER 5 FACTORS IN SUCCESSFUL BANK GROUP SUPPORT success) to harmonize three independent reforms in this area. In Uzbekistan treasury reform supported by the Bank’s Public Sector team reintroduced rigidities in spending that reduced intended increases in autonomy and flexibility at health institutions. This created a contradiction between public financial management and health financing reforms—one that has yet to be resolved. The international health financing experts raised concerns over the Bank’s dwindling capabilities in health financing. While the Bank has added value in the area of strong technical skills, the general impression is that it is not as deep as it used to be, following the departure of several more experienced health financing staff. Partner agencies reported they do not know who to contact on health financing at the Bank, and they raised concerns that recent senior retirees are not being replaced. There are also concerns that the Bank is losing its edge and a perception that it has become less serious about health financing. An example was that the Bank sent public health specialists to international costing meetings, where a health financing expert was expected. HNP staff interviewed by IEG echoed these anxieties and worried that this could affect the Bank’s future collaboration with other organizations. IEG’s country cases identified similar concerns. In several countries, including Ghana and Rwanda, the Bank did not maintain its health financing expertise even so governments embarked on substantive reforms. Without doubt, the Bank’s capabilities affect partnerships with other bodies. At the country level as well as globally, the Bank works with other donors on health financing reforms.3 It often leads the donor collaboration agenda but not necessarily in health financing. Country-level engagements in health financing vary and are influenced by the perspectives of Bank staff, their available resources, and individual capabilities. In Nepal and Tanzania the Bank coordinated well with a multidonor sectorwide approach and basket funding arrangement, and has supported the health financing agenda as an active member of the technical working group of finance. Yet in some countries collaboration with Bank staff was described as informal, sporadic, and challenging, where staff are focused on disbursements instead of technical issues in health financing. In short, there is room to leverage synergies in collaboration between organizations so as to raise the quality of the health financing dialogue at the country level. Integrating All Health Financing Functions The evaluation showed that the Bank’s health financing portfolio is changing, and there is a growing focus on purchasing. Within purchasing most attention is given to one subintervention—performance- or results-based payment reform. This trend is 66 CHAPTER 5 FACTORS IN SUCCESSFUL BANK GROUP SUPPORT continuing. Since FY13, the Bank has approved 11 new RBF projects that are cofinanced by the Health Results Innovation Trust Fund. This is an impressive shift in the portfolio given that in previous years, HNP reported about six new health financing operations annually. This shift in the Bank portfolio raises concerns that the Bank’s approach to health financing is driven by availability of trust funds; draws on an insufficient evidence base; and is not integrated with the other health financing functions and public finance. Payment reforms supported by the Bank’s RBF program tend to operate in parallel to health financing and public finance reforms (Ghana, Kenya, and Tanzania) and did not examine broader fiscal effects (Rwanda). Similarly, stakeholders interviewed by IEG in countries and among the international health financing experts indicated that the Bank focus in health financing seems to change with the viewpoints of the Bank’s leadership, with individual staff, and with the availability of donor funding to promote specific topics. For example, the Bank seemed only temporarily committed to support National Health Accounts as long as the external funding was provided. With the availability of funding from the Health Results Innovation Trust Fund, the Bank’s focus shifted to results-based payments in low-income countries. Because of these shifting areas of focus, the Bank is perceived as not properly linking health financing to poverty reduction. IEG found that the Bank has integrated health financing reforms with public sector reforms in several countries, often in collaboration with Public Sector teams. Much of this support was provided through development policy operations. In Cambodia, cross-sector collaboration has been strong and effective, and the health economist in some instances took the lead in the overall policy dialogue with the government on public servant payment reform, for example. In Turkey health reform benefited from a dialogue between the public sector and health teams about the implications of insurance expansion for health spending, which were analyzed jointly. In Argentina, Bank support to the government’s budget helped protect pro-poor health spending. This was coordinated with the results-focused design of the Plan Nacer Program (Gertler et al. 2011). In several countries (e.g., Ghana and Serbia), the Bank provided support to public health insurers including to move to capitation payment and manage cost. In Bolivia, Cambodia, Turkey, and Uzbekistan, the Bank’s purchasing support could be considered a key entry point for other health financing reforms. The Bank can add value by stressing this comparative advantage via linking health financing with public finance and working across teams, as suggested in the 2007 HNP strategy. These findings suggest that the Bank could revisit its approach to performance- and results-based financing and integrate purchasing, including that in RBF projects, 67 CHAPTER 5 FACTORS IN SUCCESSFUL BANK GROUP SUPPORT with other health financing functions and public finance. The Bank could link provider payment methods with public reforms to help institutionalize these changes. It could also help governments disseminate information on provider performance that affects the reputation of health care providers and informs consumers (e.g., infection rates, cleanliness in health facilities). As performance- and results-based payments mainly benefit patients who seek care, payment reforms need to be linked to pooling so as to reduce demand-side barriers in accessing care. In countries with social health insurance, such as Ghana, Kenya, and Tanzania, the Bank could explore working with the insurer to implement its results-based financing activities with ongoing insurance payment reforms, instead of implementing a parallel activity. This would help streamline financial incentives, reduce fragmentation, and build institutions that support sustainability. Monitoring and Evaluation in Health Financing Projects This evaluation may be missing some successful Bank and IFC engagement in health financing because of weak M&E in health projects (appendix Table B.8 and Table D.4). HNP is among the sectors with the lowest ratings for the quality of project M&E. Of the 34 closed HNP projects with an IEG project completion review included in this evaluation, 25 percent were rated substantial or high M&E performance in IEG project ratings, which is considerably below the Bank average of 32 percent. Reasons for weak M&E ratings in Bank projects include missing indicators, indicators that are too vaguely defined or not measurable, use of national data to evaluate a pilot program, and unreliable data. However, looking forward, HNP is substantially investing in impact evaluations, and most of them are evaluating RBF pilot programs (appendix figure B.8). It remains to be seen how the results from these evaluations will inform future program expansion and progress in health financing projects. The HNP strategy stipulates that the Bank monitor how health financing affects equity in service use, risk pooling, and financial protection, but this information is rarely collected in health financing operations (Table 5.2). The majority of Bank projects that aim to improve access to care monitor changes in the utilization of services, though rarely across socioeconomic groups. Of the 12 HNP projects with equity objectives, the China Rural Health Project is the only one monitoring utilization patterns across income groups. Four of the 11 HNP projects with objectives to expand health insurance report enrollment, but only two—in Turkey and Vietnam—report enrollment among the poorest as stipulated in the HNP strategy results framework. Only three of the nine projects with financial protection objectives report on changes in catastrophic spending or the impact of out-of-pocket 68 CHAPTER 5 FACTORS IN SUCCESSFUL BANK GROUP SUPPORT spending on household incomes; even so, these are indicators identified in the 2007 HNP strategy results framework. Only seven of 19 HNP projects with poverty objectives monitor changes in utilization of care for the poor. Table 5.2. Health Financing Indicators in HNP Strategy and Health Financing Projects Number of HNP Projects with Health Relevant Health Financing Indicators Bank Projects with Indicators Financing Objective Equity in service use Utilization of care across socioeconomic China (P084437) (12 of 78) groups; visit rate of top versus bottom quintile Expand risk pooling Percentage of lowest income quintile Turkey (P074053); (11 of 78) households in risk-pooling schemes (HNP Vietnam (P079663) strategy indicator) Financial protection Percentage of households experiencing Kyrgyz Republic (P084977); (9 of 78) catastrophic health expenditures China (P084437); Vietnam (P082672) Out-of-pocket expenses in health as a Kyrgyz Republic (P084977); percentage of total household expenditure China (P084437) (HNP strategy indicator) Source: World Bank (2007). References Brad Herbert Associates. 2012. Health in Africa Mid-Term Evaluation: Final Report. Wittman, Md.: Brad Herbert Associates. https://www.wbginvestmentclimate.org/advisory- services/health/health-in-africa/upload/HiA-Mid-Term-Evaluation-Final-Report-Sept- 2012.pdf. Gertler, Paul, Martinez Sebastian, and Pablo Celhay. 2011. “Impact of Plan Nacer on the Use of Service and Health Outcomes: Intermediate Results Using Administrative Data from Misiones and Tucuman Provinces.” Unpublished manuscript, World Bank, Washington, DC. IEG (Independent Evaluation Group). 2014. Tunisia Country Program Evaluation, FY05–13. Washington, DC: World Bank. IHP+ (International Health Partnership + Related Activities). 2012. Joint Assessment of Rwanda’s Third Health Sector Strategic Health Sector Strategic Plan. Geneva and Washington, DC: IHP+. 1The Bank’s Business warehouse reporting does not capture reports that were prepared within a project or under a technical assistance code. It also does not identify technical assistance or workshops. 2Most collaboration with the Bank’s Development Research Group was on impact evaluations (see Table A.8). 69 CHAPTER 5 FACTORS IN SUCCESSFUL BANK GROUP SUPPORT 3Collaboration is on strategic issues, operational and country-level support, analytical product development, and knowledge and capacity building through joint health financing workshops, conferences, and the World Bank Institute flagship course. Some organizations provide temporary funding through trust funds to advance initiatives managed by the Bank Group (e.g., National Health Accounts, Universal Health Coverage, Health in Africa, and RBF). The Bank’s collaboration with universities is limited, mainly involving universities in the United States. 70 6. Conclusions and Recommendations This evaluation examined World Bank and the International Finance Corporation (IFC) support to health financing through lending, investment, policy dialogue, and analytical work. Over FY03–12 the World Bank supported health financing reforms through 188 operations in 68 countries and provided an active analytical program. The IFC delivered a small program with six investments and nine advisory services. The Bank’s Health, Nutrition, and Population (HNP) 2007 strategy sees health financing as a comparative advantage for the Bank because of its analytical capacity and multisector nature. The Bank and IFC do not have a joint strategy or strategic approach about the mix of public and private health insurance. The Bank Group did not take an ideological stance in its work in revenue collection and different risk pooling arrangements; rather, it worked within the different country contexts. In line with the Bank’s health strategy, the Bank did promote a focus on improved results and performance in health facilities by helping governments and insurers change the way they pay providers. An increasing number of this work is implemented with the support of results-based financing (RBF) operations. The Bank’s 2007 health strategy remains valid to guide support to health financing reforms. However, the evaluation finds that key elements of the strategy have proved to be elusive in its implementation, including integration and monitoring and evaluation (M&E). The reasons mainly evolve around capabilities and cross- sector collaboration and are areas for further reflection for the global practices. The factors for successful Bank support include government commitment to address the many challenges in revenue collection for health and risk pooling, ensure pro- poor spending, introduce institutional reforms, and build technical and information capacity. Purchasing needs to be integrated with risk pooling to address financial barriers for poor individuals in accessing care and with public finance to manage adverse effects. Bank analytical work and collaboration across teams helps inform governments in these health financing decisions. These health financing functions heavily depend on information systems, M&E, technical capacity to analyze performance, and government commitment to sustain reforms. The evaluation recognized that reforms in health financing only are insufficient, and additional investments are needed to ensure the supply of health care. But health financing decisions are necessary to influence the provision and use of health care and ensure financial protection. They include decisions about how to mobilize and 71 CHAPTER 6 CONCLUSIONS AND RECOMMENDATIONS allocate funds for health, how to pool these funds, and how to purchase care from health providers. The evaluation found that evidence is scant on the effect of Bank and IFC operations and programs on final outcomes, and much remains to be learned about the health benefits, equity in service use and finance, and financial protection value of public spending, pooling, and purchasing supported by the Bank Group. There is also a critical need to strengthen evidence on implementation processes so as to identify the reasons that contribute to success. Sound analytical work about adverse effects and financial sustainability are particularly important for all countries. The four main conclusions of the evaluation are the following:  There have been some notable successes of Bank support to all three health financing functions, including revenue collection, risk pooling, and purchasing. Evidence suggests that these have occurred when the Bank Health and other sector teams drew on a variety of skills across sectors to engage government and where government commitment to reforms was strong. The collaboration between the IFC and the Bank has been limited so far, given the small health financing IFC portfolio.  Bank support has helped increase governments’ health budgets and protect health spending against budget cuts during economic crisis. Equity in pooling increased where the Bank assisted governments in subsidizing compulsory contributions to various health insurance for low-income groups. However, increased pooling did not always lead to pro-poor spending, improved equity in service use, or greater financial protection. Support to reduce user payments was limited, and evidence is missing that it improved equity in service use and financial protection. This type of support often lacked the necessary fiscal and equity analysis.  The Bank has been shifting its focus on health financing to performance- or results-based payments supported by RBF projects. There is a greater focus on financial incentives to increase the number of specific services and monitoring of service use. Little attention was given to the impact on costs, broader public sector institutional reforms to allow providers to react to financial incentives and to demand-side barriers including user fees, and how to tackle these in a fiscally sustainable manner. This shift has happened without the necessary evidence for financial protection and sustainability and potential adverse effects on the broader public sector, including on wages.  An integrated approach that links health financing including RBF with public sector reforms is likely to be more effective than single-issue interventions in establishing the relevant institutions that are needed to sustain reforms. This 72 CHAPTER 6 CONCLUSIONS AND RECOMMENDATIONS approach comprises efficient and equitable revenue instruments (tax and non-tax) for health, taking into account the overall public finance situation. It also includes moving toward compulsory pooling and reduced fragmentation in pooling, and a focus on strategic purchasing that examines potential adverse effects in a public sector context. Linking health financing reforms to public sector reforms requires strong collaboration between the IFC and the Bank’s Health and Public Sector and Finance teams to help facilitate the dialogue on health financing at all government levels. In a reorganized World Bank Group, health financing operations could benefit more from thinking and coordination across the Bank’s HNP, Governance, Macro and Fiscal Management, Poverty, and Social Protection teams as well as the IFC. This could include, for example, streamlining the methodology in the Bank’s diagnostic program to include analysis on both financial protection and adverse effects set by financial incentives, and integrating health financing analysis into the new Systematic Country Diagnostic framework which will focus on the critical challenges to achieving the twin goals of reduced poverty and shared prosperity. This evaluation makes five recommendations to guide the Bank Group’s future work on health financing: 1. Support government commitment and build technical and information capacity to be able to inform health priorities and spending by:  Supporting countries through capacity building in standardized monitoring of total health expenditures (e.g., National Health Accounts), with attention to serving the needs of the poor; and  Expanding training in client countries in collaboration with local Institutions to build knowledge and technical capacity through health financing learning platforms. 2. Address health financing as a cross-cutting issue at the country level by:  Ensuring analysis of equity in health service use and finance, financial protection, and financial sustainability consistent with the aim of promoting Universal Health Care coverage. 3. Have Global Practices focus on health financing as a core comparative advantage of the Bank by:  Building and expanding technical capacity among staff working on health financing in different Global Practices (including Health, Macro and Fiscal 73 CHAPTER 6 CONCLUSIONS AND RECOMMENDATIONS Management, Governance, Poverty, and Social Protection) to ensure that staff capacity is adequate to respond to country demand; and  Having a clearly identified focal point on health financing for the World Bank Group. 4. Integrate all health financing functions by:  Integrating results-based financing interventions with other health financing functions and the broader public finance context at the country level to address sustainability and prevent distortions; and  Developing a joint strategic approach between IFC and the Bank and complementary implementation on the ground toward health insurance, including mandatory and voluntary coverage. 5. Strengthen M&E in Bank and IFC health financing projects by:  Improving appropriate M&E frameworks in Bank and IFC projects to put in place mechanisms to collect and monitor relevant indicators; and  Monitoring distributional indicators, including on access and outcomes, consistent with benchmarking and tracking progress toward Universal Health Care coverage. 74