Malawi Public Expenditure Review 2020: Strengthening Expenditure for Human Capital December 2020 Acknowledgements This public expenditure review was led by a team comprising Patrick Hettinger (Senior Economist, TTL), with sector chapter leads Salman Asim and Nobuyuki Tanaka (Education), Collins Chansa (Health), Colin Andrews (Social Protection), and Musharraf Cyan (Fiscal Transfers). The core team also included Yalenga Nyirenda (Consultant), Tamara Mughogho (Consultant), and Laston Manja (Consultant). The Education review was written by a core team led by Salman Asim (Senior Economist) and Nobuyuki Tanaka (Economist), and included Kebede Feda (Senior Economist), Innocent Mulindwa (Senior Education Specialist), Ravinder Casley Gera (Consultant), Mofioluwasademi Ayobami Odunowo (Consultant), and Tanya June Savrimootoo (Economist). Technical inputs were also provided by Lizzie Chiwaula (Consultant), Polycarp Otieno (Consultant), Karishma Silva (Consultant), Laston Manja (Consultant), and Christopher Burningham (Consultant). The peer reviewer was Quentin Wodon (Lead Economist). The health review was prepared by a core team led by Collins Chansa (Senior Economist) and Katelyn Jison Yoo (WBG Analyst), with technical support and guidance from Dominic Nkhoma (Consultant), Mariam Ally (Senior Economist), Moritz Piatti (Senior Economist), Collins Owen Francisco Zamawe (Consultant), Toni Lee Kuguru (Health Specialist), Pia Schneider (Lead Economist), and Angela Zeleza (Consultant). From the Ministry of Health and Population, we would like to thank Gerald Manthalu, Pakwanja Desiree Tewa, and Kate Langwe. The Peer Reviewers were Owen K Smith (Senior Economist) and Ellen Van De Poel (Senior Economist). The health chapter was financed by the Global Financing Facility for Women, Children and Adolescents (GFF). The GFF is a multi-stakeholder partnership hosted in the World Bank that supports country-led efforts to improve the health of women, children and adolescents. The GFF supports 36 low and lower-middle income countries with catalytic financing and technical assistance to develop and implement prioritized national health plans and to maximize the use of domestic financing and external support to scale up access to affordable, quality care and achieve better, more sustainable health results. Since the GFF was founded in 2015, partner countries have made significant progress to improve maternal and child health. The Social Protection review was written by a core team led by Colin Andrews (Senior Economist), and included Ivan Drabek (Sr. Operations Officer), Chipo Msowoya (Social Protection Specialist), Boban Paul (Economist), Claudia Rodriguez Alas (Economist) and Edward Archibald (Consultant)). Technical inputs were also provided by Niall Kelleher (Consultant). The task was carried out under the overall guidance of Dena Ringold (former Practice Manager) and Robert S. Chase (Practice Manager). The team is grateful for data collection and research assistance from Laston Manja. The peer reviewers were Ramya Sundaram (Senior Economist) and Silas Udahemuka (Human Development Specialist). The team gratefully acknowledges development partners who provided data including UNICEF, the World Food Programme (WFP), the European Union (EU), and Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ). The Intergovernmental Fiscal Transfer System and Service Delivery review was prepared by a core team including Musharraf Cyan (Consultant and Field Work Lead), Michael Roscitt (Public Sector Specialist), Debbie Isser (Lead Governance Specialist), Douglas Porter (Governance Consultant) Atamandike Chingwanda (Health Consultant), Lizzie Chiwaula (Education Consultant), and Bauyrzhan Yedgenov (Data Analysis Consultant). The team is grateful for the research collaboration the National Local Government Finance Committee (NLGFC) provided. Abebe Adugna (Practice Manager, Macroeconomics, Trade and Investment); Vivek Suri (Practice Manager, Macroeconomics, Trade and Investment); Greg Toulmin (Country Manager, Malawi); and Mara Warwick (Country Director, Malawi) provided overall guidance. The team wishes to thank William Battaile (Lead Economist), Inaam Ul Haq (Program Leader), and Ernest Massiah (Practice Manager, Health Nutrition and Population) for their constructive inputs. This report benefited from fruitful discussions, comments and information provided by representatives of the Ministry of Finance; Ministry of Economic Planning and Development and Public Sector Reforms; the Reserve Bank of Malawi; the National Statistical Office; the Malawi Revenue Authority; and a number of other Government Ministries, Departments and Agencies. Henry Chimbali (Communications Officer), Miriam Kalembo (Team Assistant), and Tinyade Kumsinda (Team Assistant) provided assistance with external communications, design and additional production support. Sharon Chetty (Consultant) provided editorial support. The findings, interpretations, and conclusions expressed in this publication do not necessarily reflect the views of the World Bank’s Executive Directors or the countries they represent. The report is based on information current as of November 2020. Acronyms ADC Area Development Committees AIP Affordable Inputs Program AGYW Adolescent Girls and Young Women’s ASPIRE Atlas of Social Protection: Indicators of Resilience and Equity BIA Benefit Incidence Analysis CBO Community Based Organization CDF Constituency Development Fund CDSS Community Day Secondary School CFOA Common Fiduciary Oversight Arrangement CHAM Christian Health Association of Malawi CHE Current Health Expenditure COMSIP Community Savings and Investment Promotion CSPS Civil Service Pension Scheme DAHSP Decent and Affordable Housing Subsidy Program DDF District Development Fund DFID Department for International Development DHMT District Health Management Teams DHS Demographic and Health Survey DP Development Partners ECD Early Child Development EHP Essential Health Package EPWP Enhanced Public Works Programme EQUALS Equity with Quality and Learning at Secondary ESIP Education Sector Implementation Plan ESJF Education Services Joint Fund ESPIG Education Sector Implementation Grant EU European Union FAO Food and Agriculture Organization FFA Food for Assets program FISP Farm Input Subsidy Program GDP Gross Domestic Product GER Gross Enrollment Ratio GGE General Government Expenditure GHE Government Health Expenditure GHSA Global Health Security Assessment GIZ Deutsche Gesellschaft für Internationale Zusammenarbeit GPE Global Partnership for Education GPI Gender Parity Index GRF General Resources Fund HAQ Health Access Quality HCI Human Capital Index HFS Health Financing Strategy HMIS Health Management Information System HSSP Health Sector Strategic Plan IFMIS Integrated Financial Management Information System IGFTS Intergovernmental Fiscal Transfer System IHS Integrated Household Survey ILO International Labor Organization KfW Kreditanstalt für Wiederaufbau JFA Joint Financing Arrangement LAYS Learning-Adjusted Years of Schooling LGA Local Government Authority MASAF Malawi Social Action Fund MDGS Malawi Growth Development Strategy MIERA More Income and Employment in Rural Areas of Malawi MLSS Malawi Longitudinal Schools Survey MNSSP Malawi National Social Support Program MoEST Ministry of Education, Science and Technology MoEPDPSR Ministry of Economic Planning and Development and Public Sector Reforms MoF Ministry of Finance MoGCDSW Ministry of Gender, Children, Disability and Social Welfare MoH Ministry of Health MOU Memorandum of Understanding MSCE Malawi School Certificate of Education M&E Monitoring and Evaluation MVAC Malawi Vulnerability Assessment Committee NAPE Nutrition and Access to Primary Education NCD Non-Communicable Diseases NDP National Decentralization Policy NER Net enrollment ratios NESIP National Education Sector Investment Plan NGO Non-Governmental Organizations NHA National Health Accounts NLGFC National Local Government Finance Committee NNR Net National Revenues NPS National Pension System NSSP National Social Support Policy ORT Other Recurrent Transactions OSS Open Secondary Schools PBF Performance-Based Financing PCR Pupil-Classroom Ratio PER Public Expenditure Review PFM Public Finance Management PqTR Pupil-Qualified Teacher Ratio PSLCE Primary School Leaving Certificate Examination PMT Proxy Means Test PTA Parent-Teacher Association PTR Pupil-Teacher Ratio PWP Public Works Program RBM Reserve Bank of Malawi SCTP Social Cash Transfer Programme SDG Sustainable Development Goal SIG School Improvement Grant SLA Service Level Agreements SMP School Meals Program SSA Sub-Saharan Africa SSN Social Safety Nets SSRLP Social Support for Resilient Livelihoods Project STEP Skills and Technical Education Programme STI sexually transmitted infections SWAp Sector-Wide Approach TEVET Technical and Vocational Education and Training TNP Targeted Nutrition Program UBR Unified Beneficiary Registry UHC Universal Health Coverage UN United Nations UNDP United Nations Development Programme UNICEF United Nations Children’s Fund USAID United States Agency for International Development VSL Village Savings and Loans WFP World Food Programme WHO World Health Organization’s Table of Contents Acknowledgements..................................................................................................................................................ii Acronyms ................................................................................................................................................................. iv Overview................................................................................................................................................................... 1 Education Sector Review ...................................................................................................................................... 20 Introduction ........................................................................................................................................................... 20 Education Sector Performance ........................................................................................................................... 21 Enrollment, completion, and gender parity .................................................................................................. 21 Repetition and Dropout ................................................................................................................................... 22 Out-of-school children ..................................................................................................................................... 22 Overage enrollment ......................................................................................................................................... 23 Literacy and learning outcomes ..................................................................................................................... 25 External efficiency ............................................................................................................................................. 26 Expenditure Trends and Funding Sources ........................................................................................................ 27 Budget allocation and execution .................................................................................................................... 35 Internal inefficiency related to repetition and dropout............................................................................... 37 Equity of public spending..................................................................................................................................... 38 Role of government in protecting equity ....................................................................................................... 38 Unit cost analysis .............................................................................................................................................. 40 Affordability of schools and the role of households .................................................................................... 41 School-level inequities in financing ................................................................................................................ 42 Institutional arrangements and effectiveness of service delivery.................................................................. 44 Transparency of policy, budget, and human resources and accountability framework ........................ 44 Assessment of policies on recruitment, deployment, qualification, and retirement .............................. 44 Effectiveness of service delivery ..................................................................................................................... 45 Equity of service delivery ................................................................................................................................. 47 Conclusions and recommendations ................................................................................................................... 49 Health Sector Review ............................................................................................................................................ 52 Introduction ........................................................................................................................................................... 52 Health Sector Performance ................................................................................................................................. 52 Health Expenditure Trends and Funding Sources ............................................................................................ 56 Level of Total Health Expenditure .................................................................................................................. 56 Composition and Growth of Health Expenditures ....................................................................................... 57 Public Expenditure on Health ......................................................................................................................... 60 Efficiency and Value for Money ........................................................................................................................... 67 Public Financial Management ......................................................................................................................... 68 Technical Efficiency ........................................................................................................................................... 69 Value for Money ................................................................................................................................................ 71 Equity Analysis ....................................................................................................................................................... 72 Spending on Health Services at Household Level ........................................................................................ 73 Catastrophic Health Expenditure ................................................................................................................... 74 Equity of Access to Health Services ................................................................................................................ 75 Conclusions and recommendations ................................................................................................................... 77 Social Protection Sector Review .......................................................................................................................... 82 Introduction ........................................................................................................................................................... 82 Objective, Methodology and Scope .................................................................................................................... 82 Key Risks and Vulnerabilities, and the Role of Social Protection .................................................................... 84 Expenditure Trends and Sustainability .............................................................................................................. 86 Overview of total social protection expenditure .......................................................................................... 86 Social Safety Nets .................................................................................................................................................. 87 Expenditure ....................................................................................................................................................... 87 Coverage ............................................................................................................................................................ 90 Financing and Sustainability ............................................................................................................................ 92 Social Insurance .................................................................................................................................................... 93 Expenditure ....................................................................................................................................................... 93 Coverage ............................................................................................................................................................ 94 Financing and Sustainability ............................................................................................................................ 94 Labor Market Programs ....................................................................................................................................... 95 Expenditure ....................................................................................................................................................... 95 Coverage ............................................................................................................................................................ 95 Financing and Sustainability ............................................................................................................................ 95 General Subsidies ................................................................................................................................................. 96 Expenditure ....................................................................................................................................................... 96 Coverage ............................................................................................................................................................ 96 Financing and sustainability ............................................................................................................................ 96 Sector Coverage .................................................................................................................................................... 97 Total sector coverage ....................................................................................................................................... 97 Equity of Spending and Targeting Effectiveness............................................................................................... 99 Efficiency and Effectiveness of Spending......................................................................................................... 101 Conclusions and Recommendations ................................................................................................................ 107 Intergovernmental Fiscal Transfers and Local Service Delivery Review ...................................................... 111 Introduction ......................................................................................................................................................... 111 Decentralization in Malawi: Promise vs. Practice ........................................................................................... 111 Malawi’s Intergovernmental Fiscal Transfer System and Service Delivery ................................................. 112 IGFTS Performance – Parameters and Coping Mechanisms ........................................................................ 115 Adequacy ......................................................................................................................................................... 115 Predictability .................................................................................................................................................... 117 Timeliness ........................................................................................................................................................ 120 Transparency ................................................................................................................................................... 121 Equalization ..................................................................................................................................................... 124 Conclusions and Recommendations ................................................................................................................ 125 Annexes ................................................................................................................................................................ 127 Education Annexes ......................................................................................................................................... 127 Social Protection Annexes ............................................................................................................................. 136 Intergovernmental Fiscal Transfers Annexes ............................................................................................. 141 References ........................................................................................................................................................... 145 List of Figures Figure 1: Malawi’s performance on the Human Capital Index Figure 2: Share of budgeted expenditure by function Figure 3: Trend in GERs, NERs, completion rates and gender parity by level of education Figure 4: Trend in GERs by level of education, area and wealth status Figure 5: Trend in repetition and dropout rates by educational level and gender Figure 6: School-age cohort children breakdown by never attended and dropout Figure 7: Enrollment by group: underage, on time, overage by 1-2 years and 3+ years Figure 8: School attendance tree for age 15-24 Figure 9: Literacy rate trends among working age population, by gender, area, wealth status Figure 10: Average years of schooling Figure 11: Returns to additional year of education Figure 12: Sources of education finance Figure 13: Breakdown by level of education Figure 14: Trend in share of education budget Figure 15: Total spending of household by school type Figure 16: share of household spending in private schools Figure 17: Share of spending by public and household Figure 18: Enrollment distribution by level Figure 19: Annual expenditure, primary education, 2020-25 Figure 20: Budget allocation Figure 21: Execution rate Figure 22: Benefit Incidence Analysis of public expenditure on education Figure 23: Average household size in each quintile Figure 24: School participation by quintile, public schools Figure 25: School participation by quintile, private schools Figure 26: Lorenz Curve for public spending on education by level Figure 27: Total unit cost comparison by level of education and type of schools attended Figure 28: Education spending as share of household per capita spending and of total household consumption, by household wealth quintile Figure 29: Household education expenditures per child by household wealth quintile Figure 30: Overall public allocations per secondary education student Figure 31: Deployment of secondary education teachers Figure 32: Human Capital Index Figure 33: Trends in Total CHE, 2009/10-2017/18 Figure 34: Composition of Total CHE, 2006/07-2017/18 Figure 35: Nominal Growth in Health Expenditures: 2007/08-2017/18 Figure 36: Spending on the Top Causes of DALYs by Financing Sources in Malawi Figure 37: GHE by Low Income Countries and other Countries in SSA, 2017 Figure 38: Level of Public Spending on Health and Budget Execution Rates: 2014/15-2018/19 Figure 39: Level of Public Spending on Health in Per Capita Nominal and Real Terms Figure 40: Public Spending on Health; and Drugs and medical supplies Figure 41: Public Expenditure on Health by Level of Healthcare Figure 42: Trends in Public Spending on Health by Districts Figure 43: Per capita expenditure by district vs health outcomes Figure 44: Public Expenditure by Key Health Systems Inputs Figure 45: Expenditure by Key Health Systems Inputs at District Level Figure 46: Total Current Health Expenditure by Disease: 2017/18 Figure 47: Pathways for potential inefficiencies in the Health Sector Figure 48: Number of Doctors per 1,000 Population, 2018/19 Figure 49: Productivity of Health Workers by Districts - 2018/19 Figure 50: UHC Index for Low Income Countries in SSA, 2017 Figure 51: Translation of Outputs in Health Outcomes Figure 52: Out-of-Pocket Expenditure on Health as a Share of total CHE Figure 53: Level and Composition of Households Spending on Health by Income Quintile Figure 54: Percentage of households spending above 40 percent of total non-food expenditure on health Figure 55: Use of health services by income quintiles Figure 56: Utilization Incidence of Outpatient and Inpatient Services: 2010/11 vs 2016/17 Figure 57: Expenditure on Social Protection Programme by Type (% of GDP), 2011 – 2019 Figure 58: Spending (% of GDP) Trends by Social Safety Net Program (2011 – 2019) Figure 59: The Malawi social safety net system through the life cycle lens Figure 60: Financing Sources for Social Safety Nets across Africa Figure 61: Expenditure on Social Insurance as % of GDP (2011 – 2019)* Figure 62: Pensions expenditures in civil service pension scheme, % GDP Figure 63: Share of population covered by various social protection programs (2011-2019) Figure 64: Percent of Total Population Covered by Social Protection, Humanitarian Assistance and Subsidies (2010-2017) Figure 65: Percentage of Total Population and Poorest Quintile Covered by Social Protection, Humanitarian Assistance and Subsidies (2016/17) Figure 66: Percentage of Beneficiaries that belong to the Poorest Quintile - Social Protection, Humanitarian Assistance and Subsidies (2010-2017) Figure 67. Distribution of Social Protection Beneficiaries - by National Poverty Lines, 2016/17 Figure 68. Social Protection Benefits as a Percentage of Pre-Transfer Household Consumption (2010-2017) Figure 69. Social Protection Benefits as a Percentage of pre-transfer Household Consumption - by National Poverty Lines (2010-2017) Figure 70: Average Administrative Expenditure as a % of Total Expenditure, 2015 - 2016 Figure 71: Decentralization Roles and Functions Figure 72: Composition of Transfers from FY 13-14 to FY 18-19 (by transfer) Figure 73: Composition of Development Transfers from FY 15-16 to FY18-19 (by transfer) Figure 74: Development transfers as a share of NNR, FY 2014-2018 (in MWK) Figure 75: Real, Per Capita District Transfers Changes Over Time (2014 – 2018) Figure 76: DDF Budgeted vs Actual End-of-Year Receipts (by year) Figure 77: DDF Budgeted vs Actual End-of-Year Receipts (by year) Figure 78: Average Transfer Delay (Planned vs. Actual Transfer, 2018 – 2019) Figure 79: Composition of Education Transfer Formula Figure 80: Composition of Sector Transfer Formulas Figure 81: Equalization Effect of Transfers List of Tables Table 1: Projected Financing Gap - Achievement of Secondary Education for All in Malawi Table 2: Costs of internal efficiency and dropout rates (MWK thousands) Table 3: Key Health Outcome Indicators Table 4: GHE as a Share of Total GGE and GDP, 2014/15-2018/19 Table 5: Public Health Sector Wage Bill as a share of the total Public Expenditure on Health Table 6: Budget Formulation, Execution and Evaluation by Facility Ownership Table 7: Total transfers as a share of NNR, FY13-14 – FY17-18 (in MWK billion) Table 8: Actual End-of-Year Receipts as a % of Budgeted Amounts (by transfer) List of Boxes Box 1. Malawi’s Education System Box 2. Impacts of COVID-19 Box 3. Malawi’s early marriage challenge Box 4: High Levels of Fertility and Stunting are Impeding Progress on Human Capital Box 5: Responsiveness of Social Protection to COVID 19 in Malawi Box 6: Managing at the Front Lines of Health Service Delivery Box 7: Navigating the School Improvement Grant Overview This is the second and last module of a programmatic Public Expenditure Review (PER). The programmatic PER is intended to support the Government of Malawi to analyze: 1) fiscal risks and sustainability; and 2) the efficiency and equity of public expenditure, particularly in human development sectors. The first module, “Putting Fiscal Policy on a Sustainable Path”, analyzed recent macro-fiscal trends, fiscal risks and sustainability, and the policies required to manage risks and maintain sustainability. This second module, “Strengthening Expenditure for Human Capital,” includes both cross- sectoral and sector-specific analysis aimed to improve the effectiveness of using public resources for human capital development. The cross-sectoral work looks at aggregate spending levels for human capital and relevant aspects of the intergovernmental fiscal transfer system. Sector-specific reviews are presented for the education, health and social protection sectors. Human Capital Development in Malawi Despite low per capita growth, Malawi has made strong progress in many areas of human capital development in the past two decades. It not only more than halved the ratios of maternal, infant, neonatal and under-5 mortality ratios since 2000, but also fares better than the averages for regional and peer countries. As of 2018, maternal mortality had dropped from 749 to 349 deaths per 100,000 live births while under-5 mortality had declined from 173 to 56 per 1,000 live births. Life expectancy increased from 45 to 64 years between 2000 and 2018. Malawi has expanded access to primary education and literacy rates have increased, from 67 to 76 percent between 2010 and 2017. Although ‘moderate poverty’ has stagnated, ‘ultra-poverty’ has decreased from 24.5 to 20.1 percent between 2010 and 2016, which is partly attributed to expanding social protection systems.1 However, the country still faces considerable gaps in human capital, which will impede its ability to reduce poverty in the medium term. Malawi lags behind in health outcomes including HIV and malaria prevalence. High levels of stunting with approximately 39 percent of Malawi’s children below the age of five stunted, has potential to reduce the productivity of labor in future. Due to its reliance on rain- fed subsistence agriculture, largely focused on maize production, the country’s high vulnerability to weather shocks can increase the vulnerability of households to poverty either temporarily or permanently. These factors are exacerbated by high population growth of about 3 percent. Malawi has one of the highest rates of adolescent fertility in the world, with 132 births per 1,000 women aged 15-19. Strengthening human capital in Malawi will be critical to reduce poverty, increase inclusion in society, and create jobs. Low levels of human capital are an impediment to the country sustaining economic growth and structural transformation. This calls for a workforce that is prepared for more highly-skilled jobs, so that it can compete effectively in the global economy. Strengthening human capital requires investing in people through nutrition, health care, quality education, jobs and skills. As part of the World Bank Group’s Human Capital Project, it launched a new Human Capital Index (HCI) in October 2018. The HCI measures the amount of human capital that a child born today can expect to attain by age 18. It conveys the productivity of the next generation of workers compared to a 1 The Government of Malawi has two poverty measures – “moderate poverty” and “ultra-poverty.” Ultra-poverty is calculated with reference to a national poverty line measuring the consumption level needed to satisfy daily calorie requirements. Its value is approximatively US$0.82 in 2011 PPP. The moderate poor are those whose household expenditure per capita is below the total poverty line–a sum of the food and non-food poverty lines—of US$ 1.32 (2011 PPP). benchmark of complete education and full health. Globally, 56 percent of all children born today will grow up to be, at best, half as productive as they could be; and 92 percent will grow up to be, at best, 75 percent as productive as they could be. Malawi’s Human Capital Index of 0.41 indicates that a child born in Malawi today will be 41 percent as productive when she grows up as she could be if she enjoyed complete education and full health. It ranks 125 out of 157 countries in the HCI, which is relatively higher than what would be predicted for its income level. Malawi is in the bottom quartile of countries across all six indicators in the HCI. It scores particularly poorly on adult survival rate and healthy growth, as well as for harmonized test scores (see Figure 1): Probability of Survival to Age 5: Only 95 out of 100 children born in Malawi survive to age 5. Adult Survival Rate: Across Malawi, 74 percent of 15-year-old children will survive until age 60. This statistic is a proxy for the range of fatal and non-fatal health outcomes that a child born today would experience as an adult under current conditions. Healthy Growth (Not Stunted Rate): Only 61 out of 100 children are not stunted. As such, 39 out of 100 children are stunted and at risk of cognitive and physical limitations that can last a lifetime. Expected Years of School and Learning-adjusted Years of School: In Malawi, a child who starts school at age 4 can expect to complete 9.6 years of school by her 18th birthday. However, when adjusting for quality of learning, this is only equivalent to 5.5 years: a learning gap of 4 years. Harmonized Test Scores: Students in Malawi score 359 on a scale where 625 represents advanced attainment and 300 represents minimum attainment. Figure 1: Malawi’s performance on the Human Capital Index Source: World Bank 2020 Note: Large circles represent Malawi in 2020, diamonds represent Malawi in 2010, small circles represent other countries. Thick, vertical lines and color of circles reflect quartiles of the distribution. Malawi’s high adolescent fertility rate is one of the major contributors to lower human capital outcomes. It contributes to high population growth and stunting, and lower health and education outcomes and earning potential. High population growth further increases the need for higher education, health, and social protection expenditure. Nearly one in every three adolescent girls aged between 15 and 19 years in Malawi has already begun childbearing. Adolescent pregnancies contribute about 29 percent of all deliveries in Malawi, and given that most of them are undernourished and their bodies are not fully developed, they are predisposed to delivering low birth weight babies. This increases the number of neonatal deaths, which accounts for 47 percent of the total number of under-five deaths in Malawi. Further, about 15 percent of all the maternal deaths in Malawi are from adolescent girls. Ultimately, the high adolescent fertility rate contributes to poor maternal, child health and nutrition outcomes in Malawi. High levels of childhood stunting—estimated at 39 percent of the children below the age of five—reduce Malawi’s HCI and are a serious impediment to human capital development, estimated to contribute to losses of US$597 million or US$39 per capita per year (see Box 4). The main underlying cause for the high adolescent fertility rate is the high rate of child marriage. Although the rate of child marriage has decreased in recent years, it is still about 10 percentage points over the regional average (see Box 3). Child marriage is the likely cause of two-thirds of cases of girls having children before the age of 18. High child marriages contribute to high adolescent fertility which in turn leads to high population growth and stunting, and lower health and education outcomes and earning potential. Addressing cross cutting policy issues to reduce fertility and stunting levels, and reducing early marriage and childbirth, would improve human capital development, economic growth and poverty reduction in Malawi. This would call for investments in the key determinants of fertility and stunting such as agriculture, nutrition, health, education, water and sanitation, social protection, and good governance. Measures to ensure that adolescent girls remain in school are one of the best ways to end child marriage and early childbearing. Incentives and programs to keep girls in school may also lead to “tipping points” in communities whereby more and more girls remain in school and are able to delay marriage. The Government is making efforts to strengthen human capital, although this will face even greater challenges with the COVID-19 pandemic. Malawi is an early adopter of the Human Capital Project, signaling its commitment to prioritize investments in human capital, and it has prepared a Human Capital plan of action using a whole-of-government approach. Yet the pandemic will put recent progress on human capital at risk. The closure of schools during the crisis (from April to mid-October) has slowed learning and calls for developing new means of educating students. Progress in maternal and child health outcomes could be reversed, as health service delivery has largely been focused on COVID- 19 at the expense of routine essential health services. For example, disruptions in the supply of maternal and child health services could negatively impact access, utilization, and health outcomes especially for the poor and vulnerable populations. Moreover, with the slowdown in economic activity, poverty is likely to increase, especially in urban areas, where cash transfers have historically had limited reach. Fiscal space for human capital spending has been reduced by high debt service costs and is being further constrained by the COVID-19 pandemic. Fiscal deficits have averaged over 6 percent of GDP since 2013. As a result, total public debt increased from 30.5 to 59.4 percent of GDP between 2012 and 2019. This buildup has largely been high cost domestic debt, which almost tripled between 2012 and 2019, from 10.4 to 29.7 percent of GDP. Repeatedly optimistic revenue assumptions, combined with an increase in government expenditure, expenditure overruns, and arrears have contributed to high deficits. In FY2019/20, revenue shortfalls and expenditure overruns, further impacted by the COVID-19 pandemic, led to a deficit of 9.4 percent of GDP. The impact of the pandemic on growth has led to lower revenues, combined with the need for higher expenditure, which will further shrink fiscal space. Tax revenues averaged 17.3 percent of GDP from FY2016/17 to FY2018/19, to decline to 16.3 percent in FY2019/20, and even further to a projected 15.8 percent of GDP in FY2020/21 (although part of the FY2020/21 reduction is due to payroll tax reductions introduced in the FY2020/21 budget). This reduction of more than one percent of GDP of tax revenues would correspond to over 20 percent of Government spending on education, or more than one-third of Government expenditure on health or social protection (including subsidies and pensions). Malawi spends a significant portion of the national budget on human capital sectors. Expenditure on education is the largest proportion of any single sector (see Figure 2). It has averaged 18.1 percent of total Government expenditure from 2014-2018. 2 This is followed by agriculture (excluding the Farm Input Subsidy Program (FISP)) at 12.3 percent, and health (9.5 percent). Government expenditure on social protection is similar to health, at around 9.7 percent, although almost all of this is for pensions and the FISP. Figure 2: Share of budgeted expenditure by function Share of revised budget, average of FY14/15 – 17/18 Education , 18.1 Other , 24.4 Agriculture (ex-FISP) , 12.3 Interest payments , 14.0 Health , 9.5 Roads & Social Protection transport , 5.4 Defense & (incl. FISP) , 9.7 security , 6.6 Source: Authors’ calculations based on MoFEPD data The need to strengthen human capital combined with limited fiscal space heightens the need for Malawi to improve the efficiency and effectiveness of human capital expenditure. Improving the equity of spending will be critical to support the most vulnerable and improve human capital outcomes. Strengthening institutions will also be needed to improve service delivery and to ensure that resources are effectively spent at the local level. Cross-sectoral Issues in Using Public Expenditures to Improve Human Capital Malawi’s Government prioritizes spending on human capital. At an aggregate level, spending on human capital is relatively higher than the average for Sub-Saharan Africa (SSA), although high population growth increases spending needs. Malawi’s education spending of 5.0 percent of GDP (approximately 2 These figures are based on revised budget estimates from FY2014/15 to FY2017/18. US$16 per capita) is higher than the SSA average of 4.2 percent of GDP. However, it has not consistently met the Global Partnership for Education guideline of allocating at least 20 percent of the national budget to education. Similarly, for health, Government health expenditure at 2.4 percent of GDP in Malawi is higher than the average for low-income countries (1.2 percent) and the average for SSA excluding high income countries (1.9 percent). However, expenditure on social safety nets is significantly lower – at 0.8 percent of GDP in Malawi, compared to 1.2 percent in SSA, and domestic financing comprised only 6 percent of this funding. Yet for broader social protection spending—including subsidies and pensions— Malawi’s Government expenditure is around 2.8 percent of GDP. Human capital spending depends heavily on external financing, particularly for health and the social protection sectors, raising concerns of financial sustainability. Donor funding comprised between 54 and 70 percent of total health expenditure, amongst the highest levels in SSA. This high level of dependency on donors to finance the health sector could pose a risk of making health financing unsustainable and disrupting service delivery. For social protection—including subsidies and pensions— funding from development partners is over 40 percent of total funding, although donor funding for social safety nets is over 90 percent. The education sector has slightly less funding from donors, at around 15 percent. However, some education sub-sectors are highly dependent on donor financing, such as Early Child Development (ECD) programs, for which development partners fund approximately 80 percent of nutrition programs. Donor funding could be coordinated better and make better use of Government systems. Over 70 percent of the funding for health and education is off-budget. The education sector has developed a common financing mechanism to strengthen donor coordination, but expenditures under this mechanism are off- budget, limiting its integration with Government systems and planning. For health, since the Cashgate scandal, most donors have funded the sector through vertical programs and projects, with aid agencies and Non-Governmental Organizations (NGOs) using their own financing and reporting systems. In doing so, they miss an opportunity to improve the Public Finance Management (PFM) systems in the country. Moreover, Government and household funding in health is more aligned with the order of priority of the disease burden than donor spending. Social protection expenditure, as well, is largely off budget, and poorly coordinated by donors. Significant improvements could be made by leveraging the Unified Beneficiary Registry (UBR) and harmonizing social safety nets provision with humanitarian assistance. The efficiency and equity of public spending can be improved in order to strengthen outcomes. Across sectors, funding can be better allocated, particularly for teachers and health care workers. Teachers are misallocated across and within schools. Pupil-teacher ratios range from below 39 in the best-staffed schools to over 100. In addition, within schools, teachers are poorly distributed between grades, with severe shortages in lower grades and overstaffing in upper grades. The condition of classrooms, teaching, and learning materials also varies widely, even within a single sub-district zone, affecting learning outcomes. In order to achieve equitable levels of learning, there is an urgent need to increase the targeting of inputs to schools to address these inequities. Moreover, public spending is skewed toward wealthier households due to highly inequitable allocations at upper levels of education. In the health sector, more than half of public funds are spent on personal emoluments, crowding out expenditure on drugs and medical supplies, and other recurrent transactions (ORT). This contributes to persistent shortages of drugs and medical supplies at hospitals and health centers. The distribution of health workers is also highly inequitable, with higher number of doctors in Blantyre and lower staffing levels in districts with more outpatients. In social protection, funding is heavily skewed toward less efficient programs such as agricultural input subsidies and humanitarian assistance, rather than more effective, equitable cash transfer programs which target the poor. This raises questions about the choice of instruments to achieve Government objectives, particularly in light of Malawi’s continuing vulnerability to shocks. The introduction of the innovative UBR should help further improve the efficiency of Social Safety Nets (SSN) and could offer an avenue to harmonize targeting across multiple social protection programs. Financing of service delivery at the sub-national level continues to fall well short of the responsibilities that are being increasingly decentralized to local governments. Malawi’s intergovernmental fiscal transfer system (IGFTS) has failed to match policy commitments and fiscal outlays. Transfers do not adequately meet expenditure needs and are unpredictable, particularly for development transfers. Furthermore, they are not disbursed on a timely basis, disrupting service delivery. In addition, they do not sufficiently increase equitable access to services. For the education sector, grants to primary schools have shown positive results, but the needs-based component should be increased to offset the persistence of large disparities in school conditions. In the health sector, the existing formula for allocating resources to the districts is not being used, leading to wide variations across the districts with regard to per capita public health spending and health outcomes. Considering that this formula only focuses on the allocation of financial resources for drugs and operational grants in the public sector at district level, the Ministry of Health (MoH) also needs to look closely at the funding and distribution of human resources, infrastructure, and equipment. Focusing on the resources allocation formula alone will not lead to the desired improvements in efficiency and equity. Data quality, particularly at the district level, is poor and limited. The quality and availability of data on Local Government Authority (LGA) level expenditure is weak. This is partially due to the poor integration of the district and central Government financial reporting systems. This severely limits the Government’s ability to monitor and report on budgets and expenditure and assess service delivery across sectors. Recommendations The Government can address cross-cutting issues to promote human capital development by: Government resources can be more efficiently and equitably targeted. The management of key human resources can be improved. Teachers and healthcare workers can be more optimally distributed (in terms of numbers and skills mix) across districts, and for teachers even within schools, which would increase access and lead to more equitable outcomes in more remote locations. The Government can also strengthen social protection expenditure by rebalancing expenditure toward more efficient social safety nets rather than input subsidies, in order to more effectively target the poor and increase their resilience to shocks. Substantial donor resources can be better coordinated and aligned with Government priorities in order to increase their efficiency and equity. With a considerable amount of resources coming from donor funds, they should be better integrated with Government resources and strengthen planning and targeting in order to reduce inefficiency. In the health sector, this would call for increasing the alignment of donor funding with the highest levels of the disease burden and gradually making more use of Government systems. In education, donors can increase the alignment of their finance with Government systems and priorities by moving from a project-based to a programmatic approach which strengthens existing expenditure lines. The social protection sector should consolidate existing funding arrangements in order to reduce the additional burden that separate funding channels place on an already strained implementation system. Reform of the intergovernmental fiscal transfer system can help strengthen service delivery at the local government level. Intergovernmental transfers can be leveraged to strengthen service delivery at the local levels. To do so, they would need to be more timely, predictable, and transparent. Moreover, equity of spending can be strengthened by revising transfer formulations. Data quality should be strengthened in order to enable stronger monitoring and analysis . Expenditure data is not readily available for monitoring and analysis. Expenditure reporting should be strengthened in the new Integrated Financial Management Information System (IFMIS) at the central and decentralized level. One major hurdle in promoting this should be to link the local level and central financial reporting systems. Education Sector Review The education chapter aims to explore the equity and efficiency of education finance in Malawi, with a special focus on primary and secondary education. Education expenditure in Malawi is divided between secondary and higher levels, which are financed by the Ministry of Education, Science and Technology (MoEST); and the primary level, where the majority of funds are managed and expended by Local Authorities.3 Key performance indicators reflect the need to strengthen efforts to increase access at post- primary levels of education and learning outcomes at all levels. The Gross Enrollment Ratio (GER) for Malawi in 2017 is 118 percent at the primary level but much lower at post-primary levels (42 percent in lower secondary and 3 percent in tertiary), reflecting problems of access at upper levels. There are large and persistent disparities in access to post-primary education between rural and urban areas, and across richer and poorer households. Repetition is a core feature of the Malawi education system: almost one-quarter of primary students repeat in a given year. Rates of dropout are very high at the primary level (28 percent) and secondary level (23 percent), and more than one in ten Malawi school-age students are out of school. Although literacy has increased, learning outcomes are poor. Some 76 percent of 15-64 year old children and adults could read or write in 2017, compared with 67 percent in 2010. However, Grade 4 students struggle to complete even tasks aligned with the Grade 1 and 2 curricula. Even by Grade 6, only 44 percent of students reach minimum proficiency level in reading. Girls, overage students, poorer students and those without literate parents perform significantly less well in learning assessments. High rates of repetition and dropout pose large costs to public finances and to households, particularly at the primary level. The public sector is estimated to lose the equivalent of about 51 percent of total recurrent spending on primary and secondary education as a result of repetition and dropout. Similarly, households lose about 27 percent of their total current spending on education at the primary and secondary levels. This lost income and foregone output, over the lifetime of the affected children, is equivalent to 2.3 percent of GDP. 3 Although data on national-level expenditures, and transfers to districts, is relatively complete and transparent, data on district- level expenditures was not available during the preparation of this report. This gap has been partially accounted for through additional data sources. Authors use sources of empirical data on local-level school conditions and financing, including from surveys such as the Malawi Longitudinal Schools Survey (MLSS), conducted by the World Bank in partnership with Government. In addition, the team collected qualitative data through targeted district and school level field studies as case studies in selected education divisions/districts (Zomba, Mzuzu, Mzimba and Mangochi), as well as visits conducted by World Bank staff as part of preparation of pipeline projects in Malawi. Total Government spending on education increased steadily between 2015 and 2019. Education expenditure as a share of GDP increased from 4 percent in 2015 to 5.1 percent in 2019. The 2019 spending, at 21 percent of public expenditure, is in line with the Global Partnership for Education guideline that 20 percent of the national budget should be allocated to the education sector. However, it has not consistently met this target, missing it for 3 of the last 6 years. Malawi’s program of ECD is highly dependent on support from development partners. Public spending allocation by level of education is biased in favor of post-primary and tertiary education, despite significantly lower access to education above the primary level. Some 39 percent of total education spending in Malawi is funded by private households’ out-of- pocket contributions, while the Government contributes 46 percent, and development partners the remaining 15 percent. The Government’s contribution is highest at primary level, while households directed their support to the secondary and tertiary levels of education. The free secondary education policy, announced in 2018, is likely to contribute to rapid rises in the Government share of secondary spending in the next few years. Supporting the continued expansion of enrollment, while raising input standards, is likely to entail rapid increases in development expenditure in the coming years. At both primary and secondary levels, the expansion of the system poses a severe fiscal challenge. At current public unit costs, achieving the expansion of secondary access to 90 percent of secondary school-age children would increase the annual cost of secondary education by around US$114 million, which would raise spending on secondary to account for more than one-third of total education expenditure. Careful consideration of unit costs and infrastructure investments is needed as part of the preparation of the new National Education Sector Investment Plan in order to ensure a fiscally sustainable path for education expansion. Overall public spending on education in Malawi is skewed toward wealthier households as a result of highly inequitable allocations at upper levels of education. However, at the primary level, public spending appears to be more equitable in the sense that the poorest quintile receives a high share of public benefits (24 percent), 4 percent more than their population share, while the wealthiest receive 15 percent of the benefits (5 percent less than their population share). The transparency of policies and budgets for education at the national level is adequate, but budgetary and human resources frameworks at the district level are not transparent. Following decentralization reforms in 2016/17, teacher personnel emoluments and other recurring transactions are managed at the district level. However, the recording of district fund receipts and expenditure in the IFMIS is incomplete and not robust. Grants to primary schools have shown positive results, but the needs-based component should be increased. The School Improvement Grant (SIG), the main form of discretionary finance provided to schools by Government, includes a needs-based component provided to schools with severe shortages of infrastructure, teachers, or basic amenities. However, the persistence of large disparities in school conditions, even within a single zone, suggest that the needs-based component may need to be increased. Per-student allocations at secondary level are significantly inequitable across Education Management Divisions, the six sub-regions which act as the primary middle management layer for secondary education. Primary school staffing levels are subject to a wide degree of variation, which represents a serious misallocation of resources. Nationally, the best-staffed 10 percent of schools have a pupil-teacher ratio (PTR) below 39 – well below the national target – and 10 percent have a PTR above 100. These disparities reflect the weaknesses of Malawi’s system for allocating teachers to schools, in which poor quality data leads to overly broad guidance to local officials, and teachers exploit this discretion to ensure appointments in low-hardship settings. Moreover, within a school, teachers are very poorly distributed between grades, with severe shortages in lower grades and overstaffing in upper grades. The quality of school environments, and availability of teaching and learning materials, varies widely. Infrastructure provision in primary schools has struggled to keep pace with the rapid expansion of the system, and the typical Malawian primary school lacks adequate infrastructure and faces challenges of maintenance. The top 10 percent of schools have fewer than 37 pupils per classroom, while the bottom one-tenth more than 156 pupils per classroom. 4 The condition of classrooms, teaching, and learning materials also varies widely. These inequities in staffing and learning environment directly and negatively affect learning outcomes. Controlling for a wide range of school characteristics, schools in remote areas achieve significantly lower learning outcomes in Grade 4. Other factors that lower outcomes include those schools with lower than average numbers of desks, notebooks and uniforms; those relying heavily on junior teachers; or schools with a greater incidence of salary delays 5. In order to achieve equitable levels of learning, there is an urgent need to increase the targeting of inputs to schools to address these inequities. Recommendations The following priority actions are recommended to improve equity, efficiency, and transparency of education spending: Misallocation of teachers to primary schools remains the single largest source of inefficiency in the Malawian education system and requires further reform. Further strengthening of policies and procedures regarding teacher deployment and redeployment are required to ensure enforcement of allocation decisions and empower district-level officials to resist pressure to change deployments. Low-cost, safe and sustainable classrooms are needed to provide equitable access to quality education at the primary level on a fiscally sustainable basis. Evidence from pilot implementation of additional discretionary finance for schools suggests that low-cost learning shelters, procured and constructed with community management, can be obtained for around US$3,000 (in contrast with current unit costs of US$20,000), with substantial positive implications for fiscal sustainability. Primary school grant finance requires reform to respond better to school needs. The persistence of chronic inequities in school conditions, even within a single subdistrict area, suggests that there is an urgent need to increase the needs-based allocation. In addition, payment of SIG remains prone to severe delays, particularly for schools in remote areas, necessitating reform to the procedures of allocation and payment to ensure more timely delivery to schools. Improvements in teacher utilization and equitable supply of qualified teachers for all subjects are required to raise efficiency at secondary level and ensure their adequacy for the likely rapid expansion of the system. Secondary teaching loads are estimated at an average 13 hours of teaching per week or less, well below global norms. Increased double-shift instruction can reduce the need for new teachers and infrastructure. 4 Malawi Longitudinal Schools Survey (MLSS) baseline, 2016 5 ibid. Exploration of lower cost models for delivery of secondary education is needed. The ongoing revitalization of the open secondary schools (OSS) system has the potential to increase the overall efficiency of secondary expenditure. The lower cost nature of OSS, if conducted and managed as designed, would reduce parental costs to secondary education and thereby increase access. Optimal use of existing placements can also raise the efficiency of the secondary sector. Urgent action is required to raise the quality and availability of data on LGA-level education expenditures. Despite LGAs now controlling the majority of primary education expenditure, there is a lack of collated and transparent data on per-LGA spending. In order for the general public to properly assess education service delivery by LGAs there is a need for greater transparency in these areas of expenditure. Health Sector Review Malawi performs better than most of the low-income countries in SSA in service coverage, and in transforming available health services and outputs into better child health outcomes. However, it is not as effective in translating the available services and outputs into better maternal health outcomes. This could be attributed to low quality of maternal healthcare services. The COVID-19 pandemic could reverse the gains in maternal and child health outcomes that Malawi has registered since 2000, if not addressed. Since the advent of the COVID-19 pandemic in Malawi, there has been a reduction in the utilization of some of the key reproductive, maternal and child health services. This is mainly due to: (i) disruptions in the procurement and distribution of medicines and other essential medical commodities, (ii) greater emphasis on COVID-19 as compared to other essential services, and (iii) patients’ fears of contracting the disease if they enter health facilities. Financing and provision of health services in Malawi are likely to be affected by the COVID-19 pandemic whether the number of cases remains low or increases. Additional investments in the health system are required to: (i) prevent further spread of the disease, (ii) to treat the sick, and (iii) to maintain the provision of other essential services. This will require additional domestic and external funding but given the negative impact of COVID-19 on economic growth and resource mobilization worldwide, expenditure on health and other social sectors could shrink rather than increase. The Malawi Government’s commitment to funding the health sector is already high. Malawi’s total spending on health in per capita terms and as a share of GDP is higher than other low-income countries. However, total health spending per capita estimated at US$39 per year is insufficient to provide essential healthcare as outlined in the country’s health benefit package—the Essential Health Package (EHP)— which has contributed to gaps in service delivery. Further, in real terms, public expenditure on health has been decreasing despite the increasing population and high and increasing disease burden. Donor funding is still the largest source of funding to the health sector, but the growth in donor expenditure on health has been low since the Cashgate Scandal. High dependency on donors to finance the health sector in Malawi poses a risk of making health financing unsustainable, which could cause disruptions in service delivery. Since the Cashgate scandal, most donors have opted to provide funding to the health sector through vertical programs and projects. By 2017/18, about 74 percent of the donor funding to the health sector was off-budget. Moreover, most aid agencies and NGOs use their own planning, financing, procurement, and monitoring systems to manage donor funds. Use of vertical programs and parallel systems negates the five principles on aid effectiveness 6 and is a missed opportunity to improve Malawi’s PFM system. While all the three main financiers (donors, Government, and households) focus their spending on the top 10 causes of disability-adjusted life years (DALYs), Government and households’ funding is more aligned to the order of priority of the disease burden. Furthermore, donor funds are often released late and at the same time the absorption is low due to a number of reporting requirements. Resolving inefficiencies in the allocation and use of donor funds is critical because donor funding is the largest source of funding to Malawi’s health sector. In line with the national health policy, the bulk of the financial resources in the public health sector are spent at district level. However, more than half of the total public funds in the sector are spent on personnel emoluments. As a result, expenditure on drugs and medical supplies, and ORT is low. The share of spending on drugs in Malawi, at 16 percent, is lower than in other African countries. The current level of funding only caters for about six months’ supply, which leads to persistent shortages at hospitals and health centers. Despite the relatively high expenditure on personnel emoluments in Malawi, there is still a critical shortage of clinical health workers, while some of the existing health workers are underutilized. Most of the doctors in the country work in Blantyre, which has 0.03 doctors per 1,000 people— significantly less than the World Health Organization’s (WHO’s) 1 doctor per 1,000 population benchmark. However, districts with lower staffing levels see more outpatients than those with higher staffing levels which suggests that some health workers are not being fully utilized. Predictability of funding is low. Actual spending is not aligned to the budget. This could be due to weaknesses in domestic resource mobilization at national level, and gaps in health services planning. Constant expenditures below and over the budget allocations raise concerns about the credibility of the budget as a planning and resource allocation tool in the health sector. Malawi has over the years developed four needs-based formulas for distributing financial resources from the center to the districts aimed at achieving efficiency and equity objectives . However, the current formula is not being used, leading to wide variations across the districts between per capita public health spending and health outcomes. Further, the existing formula only focuses on the allocation of Government funds, and financial resources for drugs and ORT, which constitute a very small portion of the overall resource envelope in the sector. Gaps in PFM arrangements have made it difficult to deliver health services at Government hospitals and health centers. The PFM system is characterized by inadequate compliance with guidelines, poor accountability, and limited communication and dialogue. Budget execution also emphasizes control over flexibility, which has led to district authorities sidestepping the system. Moreover, poor integration of the district and central Government financial reporting systems limits budget monitoring and reporting. Although out-of-pocket expenditure on health as a share of the total current health expenditure has been increasing consistently since 2012/13, the poorest quintile of households has not been affected. Further, the proportion of poor households incurring catastrophic health payments has In line with the Paris Declaration on Aid Effectiveness, the five principles that make aid more effective are: Ownership, Alignment, Harmonization, Managing for Results, and Mutual Accountability. Several donors that operate in Malawi are signatories to the Paris Declaration on Aid Effectiveness. For more information see https://www.oecd.org/dac/effectiveness/34428351.pdf decreased. However, geographically, the incidence of catastrophic health payments has increased in rural areas while there has been a reduction in urban areas. Equity of access to health services in Malawi has improved over the years but there is room for further improvement. Increased equity of access to health services could be attributed to increased utilization of free health services by poor households at Government and Christian Health Association of Malawi (CHAM) health facilities. However, poor households still consume more health services at Government health facilities (where quality of healthcare is low) as compared to the CHAM and private health facilities. Furthermore, overall utilization of health services (across facilities) is still higher among the rich than the poor. Recommendations The efficiency and equity of expenditure in the health sector can be improved by addressing these issues: There is need to sustain the gains made in transforming the available health services and outputs into better child health outcomes while scaling-up the delivery of quality maternal health services. This will require addressing underlying issues such as the high adolescent pregnancy, high teenage marriages, low education status for the mothers, poor maternal nutrition, and low access to social protection services. Additional financial and material resources will be required to prevent further spread of COVID-19, to treat the sick, and to maintain the provision of other essential services. The Government needs to ensure that existing financial resources in the health sector are not reduced. If an economic crisis forces the issue, the Government could consider cutting certain areas of the national and/or health budget to sustain the provision of health services with the least possible impact on health and other social outcomes—e.g., postponing capital spending or temporarily suspending some non-essential services. The Government should continuously monitor how the COVID-19 pandemic is affecting supply and demand for health services, consistently map how available resources are allocated and spent, and undertake timely procurement and distribution of medicines and other essential medical commodities. Scaling-up risk communication on the COVID-19 pandemic and maintaining the provision of essential health services during the COVID-19 outbreak is critical to reducing morbidity and mortality due to the COVID-19 itself and from other diseases. The MoH needs to improve efficiency in the allocation and use of available resources by: (i) re-prioritizing Government spending as highlighted above, and (ii) developing a financial sustainability plan that could extend the available Government and donor funding for an additional 3-5 years. This is because the likelihood of increasing domestic and external funding is very low given the already high Government and donor spending in the health sector; and reduced revenue generation capacity worldwide. The Government could consider developing and implementing a comprehensive health financing strategy (HFS) to guide resource mobilization, pooling, allocation, and purchasing of health services. The strategy needs to encompass the financial sustainability plan (alluded to above) and viable strategies for promoting financial resilience and efficiency. The HFS being proposed in Malawi could focus more on resilience and measures to enhance efficiencies. In order to improve the overall allocation of funds, governance and accountability in the health sector, donor funding needs to be aligned to Government systems at both district and national levels. Immediate actions include: (i) developing a system for routine mapping and tracking external funds at both central and district levels, (ii) aligning donor funding to the order of priority of the disease burden, and (iii) increasing predictability of donor funding through the use of joint budgeting, disbursement, financial management, procurement, and reporting mechanisms. The Government needs to enforce use of the existing PFM guidelines, and facilitate improvements in communication and dialogue on planning and budgeting at district level. There is also need for greater flexibility on budget execution at Government health facilities similar to the flexibility at CHAM health facilities. Health budgets at district level also need to be ring-fenced to avoid inter-governmental transfers when the funds are disbursed to the district councils. There is also a need to integrate accounting systems at the district and central Government levels to enable the generation of integrated financial reports for the sector. To achieve its efficiency and equity objectives, the Government needs to apply the revised (2019) district-level resources allocation formula to both Government and donor resources across all districts. Furthermore, considering that this formula only focuses on the allocation of financial resources for drugs and operational grants in the public sector at district level, the MoH also needs to look closely at the funding and distribution of human resources, infrastructure, and equipment. Comprehensive training and mentorship of district authorities and service providers on health services planning and budgeting could help to improve the allocation of resources and budget execution. Further, there is need for consistent advocacy on evidence-based planning and application of the revised resource allocation formula among policymakers and planners. There is need to improve purchasing and value-for-money in the health sector. The Government needs to address the shortage of health workers and low utilization of existing health workers by: (i) distributing the available workforce optimally across all districts, and (ii) increasing the productivity of existing health workers by introducing performance-based financing (PBF) schemes. By using PBF, financing to health facilities could be distributed based on outputs produced rather than inputs. To achieve universal access to quality healthcare, the Government needs to improve quality of healthcare services at Government health facilities since this is where most of the poor people access services. In addition, there is a need to further increase access for poor households to CHAM and private health facilities especially in areas where there are no government health facilities. This could be achieved through the introduction of vouchers in addition to the existing service level agreements between the Government and CHAM. Social Protection Sector Review Malawi’s high poverty rates and population growth, accelerating environmental degradation, and susceptibility to weather shocks make resilience-building through social protection an increasingly important priority. The second Malawi National Social Support Program (MNSSP II) recognizes this and works toward creating a dynamic safety net system that can better respond to persistent poverty, recurrent shocks and a demographic dividend. This social protection expenditure review cuts across four main pillars: 1) social assistance (safety nets), 2) social insurance, 3) labor market programs, and 4) subsidies. Humanitarian assistance to address food insecurity is also analyzed due to the close linkages with social safety nets in Malawi and as a useful comparator to contextualize trends. Expenditure and Financing Between 2011 and 2019, an average of 4.9 percent equivalent of GDP was spent on social protection programs in Malawi, cutting across social safety nets (SSNs), social insurance, labor market programs, and subsidies. This was driven by expenditures on subsidies and pensions. Government expenditure covered most of the subsidies and civil service pensions expenditure, at around 2.8 percent of GDP. However, for safety nets, 94 percent of expenditure is donor financed. Spending on SSNs in Malawi is strikingly low compared to the region. Malawi’s expenditure on SSNs was equivalent to 0.8 percent of GDP between 2011 and 2019, compared to an average of 1.2 percent across the Sub-Saharan Africa region. Malawi implements a narrow range of SSN interventions that focus on the most vulnerable households. Regular and large-scale outlays of humanitarian assistance suggest there is significant room for strengthening the shock sensitivity of Malawi’s SSNs and enhancing their capacity to build resilience. On average, an equivalent of 1.3 percent of GDP was spent yearly in Malawi on humanitarian assistance in response to food insecurity from 2011-2019, substantially higher than comparator countries such as Tanzania (0.09 percent) and Rwanda (0.28 percent). The politically popular Farm Input Subsidy Program (FISP) has also absorbed a large share of social protection spending in recent years. FISP provides subsidized fertilizer and seeds to smallholder farmers. Following reforms in recent years, spending has fallen from 3.5 percent of GDP in 2012 to 0.6 percent of GDP in 2019. However, with the introduction of the Affordable Inputs Program (AIP) in FY2020/21, spending has increased to above 2 percent of GDP. 7 Malawi faces considerable financing requirements for its SSN programs, which rely heavily on development partner funding. Domestic financing accounts for just 6 percent of total safety net expenditure, raising stark questions about the sustainability and affordability of interventions targeting the poorest in Malawi. Limited utilization of Government systems leads to concerns over program ownership and sustainability. Pension expenditures are also likely to be unsustainable in the future. Coverage Social safety net programs in Malawi are estimated to cover 24 percent of the population in 2019, compared to an average of 10 percent across the Africa region. The School Meals Program (SMP) has been Malawi’s largest SSN program in terms of coverage, although coverage of the Social Cash Transfer Programme (SCTP) has quadrupled since 2011 and is now reaching all ultra-poor rural households nationally. The average transfer value for SSNs has increased over time but remains low. The aggregate value of monetary benefits provided by SSN programs represent just 4.8 percent of total income for eligible households. In comparison, social insurance programs constitute almost one quarter of beneficiary welfare. 7 Introduced in the FY2020/21 budget, the AIP increases beneficiaries from 900,000 to 4.2 million—thereby effectively reaching all farming households. It also reduces the price of fertilizer paid by smallholder farmers to MWK 4495 and has increased the subsidy component payable by the Government. The technical analysis for this PER was completed before the introduction of the AIP, therefore it is not fully assessed within the scope of this PER. While the overall coverage of SSNs appears encouraging, numerous concerns arise when the data is disaggregated. SSN programs still exclude a sizeable portion of the population. There are gaps in the coverage and objectives of safety nets, particularly for children and youth. The extensive coverage of humanitarian assistance and FISP also raise questions about the choice of instruments to achieve Government objectives. Humanitarian assistance and FISP have substantially higher coverage than any single SSN intervention, peaking at 34.5 percent in 2016 and 46 percent of the population in 2011, respectively. While this may underscore the need for support, they are not the ideal instruments to respond to such needs. The substantial funding for humanitarian assistance and FISP is likely crowding out investment in better targeted and cheaper social protection programs. Performance Overall, there has been an improvement in recent years in the effectiveness of SSN programs in targeting the poor. This is expected to improve further with the introduction of the UBR. By contrast, the effectiveness of targeting in humanitarian assistance and FISP has fluctuated and their performance in targeting the poorest households is weaker than SSNs. The SCTP has one of the highest incidences of beneficiaries (about 40 percent) that are extremely poor. Recent innovations in targeting under the UBR aim to harmonize targeting across multiple social protection programs, thereby providing an important platform to leverage more effective targeting going forward. While the overall impact on poverty and inequality of SSN programs is negligible given their low coverage and benefit levels, individual programs indicate important evaluative findings. The SCTP has shown some of the most powerful and consistent positive impacts across sub-Saharan Africa. It was ranked first in terms of total consumption and food consumption out of 27 SSN programs in 14 African countries, with beneficiaries spending an equivalent of 179 percent of the transfer amount. Studies also report increased investments in productive agricultural assets, time spent on farm activities, and a lower likelihood of engaging in casual agricultural labor outside the home. SCTP beneficiaries also experienced improvements in school enrollment, attendance and education related expenditure for children. On average, the cost of delivering one unit of benefit is significantly lower for SSNs than for humanitarian assistance. The cost of giving one dollar of benefit is 55 cents for humanitarian assistance, compared to much lower amounts for the SSN programs such as the SMP (7 cents), MASAF PWP (11 cents) and SCTP (18 cents). Similarly, close to 35 percent of expenditure on humanitarian assistance in response to food insecurity in 2015–2016 went to administrative costs, compared to an average (weighted) of 14 percent for SSN programs. Governance and Institutional Issues There are numerous concerns regarding the governance and institutional structures behind Malawi’s safety net system. Provision of SSNs is characterized by fragmentation at the national level, with insufficient capacity to ensure stakeholder accountability or to enforce cross-ministerial coordination. Ineffective coordination at the district level is a major constraint to developing a more coherent system, along with inadequate resourcing, poor infrastructure and limited staffing. The coordination of donor efforts is also a key issue. Malawi’s social safety net programs are largely donor-funded with only some of them using Government systems. The SMP has hardly any Government involvement, whereas the SCTP and MASAF PWP have made considerable use of Government systems. This raises concerns about the ownership and sustainability of programs that exist almost independently of the Government. Donors funding these programs contribute to this fragmentation by adopting different financing structures and implementation modalities, making it difficult to coordinate their initiatives, even for those within the same program, such as the SCTP. Recommendations Programmatic reforms which are crucial to improve the overall effectiveness of the sector include: Reallocate resources from less effective initiatives toward evidence-backed social protection programs. With high levels of spending directed to the FISP (and now AIP) and social insurance, opportunities exist to reallocate resources toward more effective and efficient programs such as the SCTP. Develop new approaches and systems for improved targeting of the poor. Ongoing efforts to improve the targeting of support to the chronic and transitory poor should be sustained. Incorporate human capital-oriented elements in safety net interventions, including through an integrated or multi-sectoral approach. This is critical, given Malawi’s evolving demographic patterns, skill requirements and political environment. Improve the shock sensitivity of social protection through greater linkages between safety nets and humanitarian assistance, and establishing scalable financing mechanisms. Safety nets and humanitarian assistance operate independently from each other, despite having many overlapping functions. Establish a national productive inclusion strategy that ties together existing programming and provides a future vision for the sector. There is an opportunity to work toward a “safety nets plus” model, which builds on existing interventions to promote stronger linkages to economic and social inclusion. Maintain current momentum for investments in core delivery systems. Strengthening core systems is a critical pathway to improve the effectiveness of safety nets and other programs. Support improved analysis around the future needs and sustainability of social insurance. The Government will face a large fiscal burden if it responds to demographic changes by scaling up pension coverage. Financing reforms to support a fiscally sustainable scale up of effective programs in the sector include: Reallocate current fiscal spending to provide a more progressive and effective mix of programs within the social protection portfolio. Reallocating expenditures from the FISP (and now AIP) to the SCTP would result in less leakage of benefits to households in the highest income deciles. Create a concrete national financing strategy for social protection programs. The Government should establish and meet domestic spending budgets for safety net interventions, involving incremental commitments to national budget lines that clearly reflect core sector priorities (e.g. SCTP, UBR). Utilize contingency or reserve financing mechanisms. Malawi can better consolidate safety net, humanitarian, and disaster responses through contingency financing mechanisms. Consolidate existing funding arrangements. Donors should acknowledge the additional burden that separate funding channels place on an already strained implementation system. Engage donors in discussions around funding structure revisions. Government and donors should actively explore the option of multi-year commitments to finance seasonable and scalable safety net interventions in lieu of short-term humanitarian assistance. Intergovernmental Fiscal Transfers and Local Service Delivery Decentralization of service delivery was formally enshrined in the 1995 constitution as part of the transition to multi-party democracy. However, implementation to date has been characterized by a blurring of intergovernmental accountability lines, both between and within each level of government. While decentralization was a key feature of second Malawi Growth Development Strategy (MDGS II), implementation to date in many sectors has been marked by resistance of central line ministries to functionally decentralize power and resources to local authorities. Taking these dynamics into account, recent moves in favor of decentralization have presented a critical opportunity to endorse more workable arrangements between central line ministries and local authorities that could promote incentives for improved service delivery. Such commitments to decentralization, however, must be matched by stronger accountability by local authorities on their use of funds and execution of service delivery responsibilities. Financing of decentralized local service delivery is primarily done using the Malawi’s IGFTS. However, this system has persistently failed to match policy commitments and fiscal outlays for the increased decentralization of service delivery responsibilities in key sectors such as health and education. This is evidenced by the fact that transfers in these two sectors have decreased by 25-45 percent in real per capita terms for almost all local authorities since 2014. The IGFTS in Malawi works through 22 cash-based transfers, where the flow of funds passes through several organizations and institutional processes. This includes 18 sector transfers for heavily earmarked non-wage ORT. Although personal emoluments account for 75 to 80 percent of all budgeted funds at the local level, the funds are not part of the transfer formula and are never actually transferred to local council accounts. Development transfers have become an increasing component of the transfer pool (increasing from 6 percent of the total pool in FY 2013/14 to 38 percent in FY 2018/19), but have also become a dynamic area for political interest as evidenced by the budget conflicts over funding of the District Development Fund (DDF) versus the Constituency Development Fund (CDF). This growing pot has also come with increasing fragmentation, marked by: (i) a disproportionately growing CDF which has increased by 132 percent since 2014; (ii) a dedicated ‘boreholes grant’ which provides funding for four boreholes in every constituency annually; and (iii) in FY18/19, a community-driven fund (COMSEP) which splits MWK 4 billion across all constituencies through Area Development Committees (ADCs). Evaluation of the IGFTS reveals several challenges across five criteria. First, transfers do not adequately meet expenditure needs for service delivery responsibilities assigned to local authorities. There is no policy indicating that ‘finance should follow function’, as vertical sharing is not informed by the minimum expenditure levels required for agreed levels of service delivery in decentralization policy. Second, the predictability of the IGFTS varies across transfers. IGFTS are relatively predictable in sector ORT transfers in health and education (with variance across districts) and in the General Resources Fund (GRF). However, predictability diverges in development transfers, with the CDF growing and recognizing above 99 percent of budgeted totals every year, while the District Development Fund (DDF) was cut by over 60 percent at mid-term in FY18/19. Third, the timeliness of transfers are marked by irregular patterns of delays. Difficult cash flow management decisions therefore have to be made by Local Authorities in order to cope, the implications of which often compound problems being faced at the health and education facility level. Fourth, transparency is low, especially as the complexity of the formula for horizontal shares diminishes transparency and impact varies. Formulas in some sectors, such as health and education, are complex, and recipients of the transfers do not clearly comprehend the formula, its variables, their weights or the application made to determine the shares. Fifth, equalization is weak, so that the impact of transfers does not increase equitable access to services. Malawi’s IGFTS is in need of reform. The present system reduces local governments, de facto, to spending units of the central Government. As such, it generates few of the envisaged benefits of decentralized governance in terms of allocative efficiency, or strengthened accountability, be this between citizens and the state, between local government sector managers and staff at the facility level responsible for delivering services, or between elected leaders and administrators. Put differently, the current system provides little basis for local discretion and as a result, local resource allocation and service delivery remain suboptimal. Multiple but specific purpose transfers leave little room for local budgetary decisions addressing local need. Moreover, Malawi has weak revenue decentralization. The urgency of the COVID-19 response has accentuated the shortcomings of the IGFTS and broader arrangements of decentralized service delivery in Malawi to efficiently send funds to the front lines in a coordinated fashion. At the onset of the pandemic, a blanket transfer of between MWK 30 – 35 million with loose guidelines for implementation were transferred to each local authority directly by the Ministry of Health, bypassing the systems of the formal IGFTS through the National Local Government Finance Committee (NLGFC). As resources continued to flow into the country from donors, there were increased concerns around the lack of transparency and coordination surrounding the funds and equipment earmarked for local authorities and the (lack of) mechanisms for strategic allocation of resources to the respective local authorities. Finally, there have been challenges in bridging the gap between national and local COVID-19 response plans. All 35 local authorities have developed budgeted COVID-19 response plans, but initial indications reveal these are largely disconnected from what has been planned and budgeted for at the national level. Recommendations Areas to prioritize for reform for the IGFTS include: Reduce the fragmentation of development transfers: While the vertical imbalance of the IGFTS remains stark, a key first step to correcting this would be a reduction in the fragmentation of development transfers to local authorities and a commitment to the DDF as a consolidated fund for discretionary development expenditures. Efforts to reduce this fragmentation and increase the quantity of discretionary development funds would be an essential part of any credible signaling by Government that it is committed to reforming inter-governmental relations in Malawi. Improve the predictability of annual and monthly transfers : A commitment should be made to increasing the predictability of the DDF, which continues to be cut by over 50 percent at mid- term on an annual basis and leaves LAs unable to deliver on planned development projects. In addition to year-end predictability, month to month predictability should also be the focus. Floors and ceilings for departures from the budgeted shares should be specified, creating an accountability for the Ministry of Finance as well. NLGFC’s mandate to report on predictability should be reiterated. Reduce the uncertainty in timeliness of transfers: The flow of funds system should be made transparent and the NLGFC should be given a role to monitor and report on timeliness. However, it is crucial to ensure that districts are more accountable for the consequences of their cash management decisions on the operations of schools and clinics. Increase the transparency of the IGFTS: Support for NLGFC capability and incentivized performance for annual collation, analysis, reporting, and public forum of implementation of the IGFTS should be agreed. Improvements in transparency should be feasible through simplification of the transfer formula scheme as well as instituting a role for periodic NLGFC reports on various parameters of the IGFTS. Application of formula for DDF and CDF should be ensured as an immediate priority. A good data regime, including completion of work on IFMIS, should be prioritized. Strengthen the equalization effects of the transfer system: Equalization should be strengthened through revision of formulas and inclusion of rational expenditure needs and fiscal capacity measures in the formulas. This should be achievable over time, starting with an explicit policy on equalization with a strategy for implementation. Education Sector Review Introduction 1. This chapter aims to explore the equity and efficiency of education finance in Malawi, with a special focus on primary and secondary education. This chapter will: (i) assess the adequacy and sustainability of public spending in the sector; (ii) evaluate the efficiency and effectiveness in the use of these resources; (iii) conduct an equity analysis of education expenditures and whether they protect the poor and the most disadvantaged populations; (iv) provide policy recommendations to increase efficiency and equity of spending and outcomes; (v) help inform ongoing sector reforms, including the preparation of the new National Education Sector Investment Plan (NESIP) 2020-2030 and Education Sector Implementation Plan (ESIP) 2020-25; (vi) help establish a baseline for evaluating public spending in the sector; and (vii) inform existing or planned programs of sector support from development partners (DPs), including the Education Sector Implementation Grant (ESPIG) from the Global Partnership for Education (GPE), which supports the new ESIP and which is currently under preparation. Box 1. Malawi’s Education System Malawi’s education system structure Malawi’s education system includes eight years of primary education (Standards 1-8), followed by four years of secondary education (Forms 1-4) and four or five years of university education.8 As of 2017/18, the system included: 5,187,634 primary students, of which 123,717 attend private schools 6,065 primary schools, of which 513 are private 387,569 secondary students, of which 86,863 attend private schools 1,469 secondary schools, of which 353 are private, and 295 semi-public open secondary schools9 Six higher education institutions (HEIs) 2. Limitations of analysis and available data. Education expenditure in Malawi is divided between secondary and higher levels, which are financed by the Ministry of Education, Science and Technology (MoEST); and primary level, where the majority of funds are managed and expended by local government authorities (see Box A1, in Annex 1). Although data on national-level expenditures, and transfers to districts, is relatively complete and transparent, data on district-level expenditures was not available during the preparation of this report, this gap has been partially accounted for through additional data sources.10 Duration of higher education depends on program of study. Early childhood education is managed by the private/NGO sector or Community Based Childcare Centers – CBCCs. The remaining secondary schools are Government boarding and day schools, and community day secondary schools. Authors used sources of empirical data on local-level school conditions and financing, including from surveys such as the Malawi Longitudinal Schools Survey (MLSS), conducted by the World Bank in partnership with the Government. In addition, the team Education Sector Performance Enrollment, completion, and gender parity 3. Key enrolment and completion indicators reflect the need to bolster efforts to increase access at post-primary levels of education. The gross enrollment ratio (GER) for Malawi in 2017 was 118 percent at the primary level but much lower at post-primary levels (42 percent in secondary and 3 percent in tertiary), reflecting problems of access at upper levels (Figure 3). Net enrollment ratios (NER) are significantly lower than GERs, reflecting widespread under-age enrollment at primary level, and over- age enrollment at all levels, exacerbated by high rates of repetition; like GERs, NERs fell rapidly from 87 percent at primary level to 15 percent at the secondary school level in 2017. About 44 percent of students entering the schooling system complete primary school–this rate fell from 54 percent in 2013. The completion rate at the secondary school level on the other hand increased from 23 percent in 2010 to 30 percent in 2017. Problems of access at the secondary school level are more serious for female students: the gender parity index (GPI) is 100 percent at the primary school level, reflecting broadly equal access to primary education for girls and boys, but falls to 81 percent at the secondary level, reflecting challenges in girls’ transition to secondary school. At the tertiary level, the GPI is just 54 percent, reflecting significant problems of access for females due to marriage, childbirth and other family responsibilities (see Box 3). These gender disparities in education attainment contribute to significant gender gaps in poverty in Malawi (pg. 85). Figure 3: Trend in GERs, NERs, completion rates and gender parity by level of education Access, percent 140 2010 2013 2017 115.3 118.6 97.1 93.9 120 88.4 86.0 84.4 76.5 75.5 100 72.2 53.9 80 49.6 38.5 35.3 60 29.6 23.1 15.2 40 11.2 3.1 2.0 20 0 Secondary Secondary Secondary Secondary Primary Primary Primary Primary Tertiary Tertiary Tertiary Tertiary GER NER Completion rate GPI Source: Authors’ calculations based on IHS 2016/17 4. Focusing on post-primary levels, there is a huge disparity in access across areas and wealth quintiles and this has persisted over time. Access across secondary and tertiary education is higher in urban areas than in rural areas. GER at the secondary level in urban areas increased from 70 percent in 2010 to 88 percent in 2017. Likewise, in rural areas, the GER increased from 29 percent in 2010 to 32 percent in 2017. While access for the lowest wealth quintile marginally increased from 12 percent in 2010 to 13 percent in 2017, access for children in the highest quintile increased from 78 percent in 2010 to 102 percent in 2017. Access to tertiary education is biased toward the rich as only youth from wealthier households enroll in post-secondary institutions (see Figure 4). collected qualitative data through targeted district and school case studies in selected education divisions/districts (Zomba, Mzuzu, Mzimba and Mangochi), as well as visits conducted by World Bank staff as part of the preparation of pipeline projects in Malawi. Figure 4: Trend in GERs by level of education, area and wealth status Gross Enrollment Ratio, percent 102.2 120 92.0 87.6 2010 2013 2017 82.0 100 77.9 80 69.9 52.3 48.4 40.6 60 32.5 32.1 31.7 31.3 30.0 29.3 24.8 22.2 20.6 40 15.1 13.1 12.3 10.7 9.1 6.2 20 1.5 0.8 0.6 0.4 0.3 0.3 0.3 0.2 0.2 0.1 0.1 - 0 Middle Middle Rural Richest Richest Urban Rich Rich Poor Poor Poorest Poorest Secondary Tertiary Source: Authors’ calculations based on IHS 2010, 2013 and 2016/17 Repetition and Dropout 5. Repetition is a core feature of the Malawi education system, with high rates particularly at primary level with no significant improvement between 2016 and 2018. Almost one-quarter of primary students repeat in a given year (23 percent for boys and 22 percent for girls in 2018; see Figure 5). However, repetition rates at the secondary level are lower and have remained at 3 percent over the last two years (Figure 5). At the secondary school level, girls are slightly more likely to repeat than boys. Dropout rates are very high at the primary level (28 percent) and secondary level (23 percent), without improvements over the period under review11; these persistent rates of dropout contribute significantly to large shares of out-of-school children (see pg 6).12 Figure 5: Trend in repetition and dropout rates by educational level and gender Repetition and dropout, percent Primary Secondary 28.4 28.4 28.3 28.0 27.8 27.8 27.5 27.1 27.1 24.9 24.5 24.3 24.1 23.9 30 23.4 23.3 23.3 23.2 23.0 22.7 22.4 22.1 21.8 21.7 25 20 15 10 4.0 3.6 3.4 3.2 3.2 3.2 3.1 3.0 2.9 5 - - - 0 2016 2017 2018 2016 2017 2018 2016 2017 2018 2016 2017 2018 2016 2017 2018 2016 2017 2018 Male Female Malawi Male Female Malawi Repetition Dropout Source: Authors’ calculations based on EMIS 2016, 2017, 2018. Dropout defined as enrollment – repetition – promotion. Out-of-school children 6. More than one in 10 school-age students in Malawi are out of school. About 12 percent of primary school-age children are out of school with the majority (10 percent) of them never attending 11 Dropout rates calculated as average across schools and grades. Defined as enrollment – repeating students – promoted students / total enrollment. Direct records of dropped-out students typically show lower figures. 12 2018 dropout rates are not calculated for secondary education due to data limitations. school (see Figure 6). Boys are more likely than girls never to have attended primary school but equally likely to drop out. Children from rural and poor households are less likely to attend school than children from urban areas and wealthy households. At lower secondary school age, the rate of out-of-school children is similar (13 percent) but most are those who have dropped out (11 percent). Dropout rates are higher for girls, and children from rural and poor households. These high rates of out-of-school children represent a threat to Malawi’s achievement of the Sustainable Development Goals as well as to future human capital and workforce development. Across educational levels, children are significantly more likely to be out of school in rural than in urban areas; this may reflect disparities in school conditions and staffing.13 Figure 6: School-age cohort children breakdown by never attended and dropout Out of school rates, percent Source: Authors’ calculations based on IHS 2016/17 Overage enrollment 7. Delayed entry into the school system and overage enrollment are important factors that contribute to high dropout rates. High rates of late enrollment, combined with the high incidence of repetition, means that significant numbers of students are overage (Figure 7). With the official school entry age at 6, only about 30 percent of grade 1 students enter by age 6, about 43 percent enroll between the ages of 7 and 8 and 17 percent enter after at least a 3- year delay. By grades 8, 10 and 12, the share of students entering school on time is reduced to 8 percent and delayed entry by 3+ years is 55 percent.14 Being an overage student may affect the self-esteem and motivation of the student, especially among a younger cohort, which also negatively affects the teachers’ ability to manage the class. Therefore, integrated policy action is required for enrolling children on time and moving them smoothly through Children from poorest households and rural areas are more likely to be out of school. However, the rates have improved between 2010 and 2017. Figure A1 (in Annex 1) shows the distribution of out-of-school rates across wealth quintiles. The out-of-school rate in 2017 for primary school-age children is 18 percent in the poorest quintile compared to 4 percent in the wealthiest quintile. At primary level (ages 6-13), poorer households are more likely both to have school-age students who have never attended school, and those who have dropped out. At lower secondary level (ages 14-15), the total out-of-school rate in 2017 is 17 percent compared to three percent for the wealthiest quintile, representing radically different levels of engagement with secondary school for the wealthiest families. Overall, lower secondary school-age children from the poorest households are more than five times likely to be out of school than those from the wealthiest households, demonstrating severe challenges of equity of access at this level. Dropout rates decreased between 2010 and 2017 for boys and girls, across wealth quintiles and areas of residence. 14 One reason why the rates are relatively stable between grades 8 to 12 is because repetition rates are lower in secondary school compared to primary school. the system. A managed automatic promotion policy could be a feasible proposition to improve efficiencies.15 Figure 7: Enrollment by group: underage, on time, overage by 1-2 years and 3+ years Enrollment by age Under age On time 100% 16.6 80% 54.7 54.5 57.0 60% 42.7 40% 31.8 31.1 29.3 29.8 20% 8.6 7.7 6.7 10.8 4.9 6.7 7.0 0% Grade 1 Grade 8 Grade 10 Grade 12 Source: Authors’ estimations based on IHS 2016/2017 8. Many children who start school do not finish upper secondary school, particularly children from the lowest wealth quintiles. Comparison across wealth quintiles shows significant variations in the levels at which youth leave the school system, with poor households consistently worse off (Figure 8). Of 15-24 year olds from the highest wealth quintile, 20 percent have completed secondary school compared to 3 percent in the lowest three quintiles. These disparities highlight that high rates of dropout and low rates of transition to secondary level have equity as well as efficiency implications. Figure 8: School attendance tree for age 15-24 Percent Never attended school Left primary school Still in primary Left lower secondary Still in lower secondary Left upper secondary Still in upper secondary Completed upper secondary 100 3 3 3 6 3 4 7 7 5 10 20 9 8 80 8 10 20 9 60 37 31 27 21 12 25 40 39 36 14 20 35 31 31 16 0 3 4 4 2 1 3 Poorest Poor er Middle Richer Richest Ma la wi Source: Authors’ estimations based on IHS 2016/2017 Although automatic promotion is generally assumed to be a threat to quality, many impact evaluations show that there is no effect on the chances of survival through the school system. For example, an econometric breakdown of the promotion into two parts, one based on indicators of merit (attendance and achievement in mathematics and language) and the other uncorrelated with those indicators) allow a test of whether parental decisions about enrollment are influenced by merit or non-merit promotions. Results suggest that enrollment decisions are heavily influenced by student academic performance in the previous year, and that promotions that are uncorrelated with merit have a negligible impact on school continuation. (Elizabeth M. Kinga, Peter F. Orazem, Elizabeth M. Paterno (1999)) Literacy and learning outcomes 9. Literacy rates in Malawi improved between 2010 and 2017. About 24 percent of the working population cannot read or write, although literacy rates are higher among younger Malawians. Figure 9 shows self-reported literacy status. In 2017, 76 percent of 15-64 year olds could read or write, compared with 67 percent in 2010. Similarly, 85 percent of 15-24 year olds could read and write in 2017, increasing from 76 percent in 2010. Rates are lower for females in both age groups, although they are considerably closer among 15-24 year olds, suggesting a gradual improvement in gender parity. Not surprisingly, literacy rates are higher in urban areas and in the richest quintile, and this has persisted over time. Figure 9: Literacy rate trends among working age population, by gender, area, wealth status Percent 100 2010 2013 95 95 93 91 91 91 89 89 89 87 90 87 86 86 85 84 83 83 83 80 80 79 78 78 80 76 76 76 75 75 73 73 72 70 70 67 70 66 65 63 63 62 60 57 60 50 Total Urban Total Urban Rural Rural Richest Richest Male Male Female Female Poorest Poorest Youth (age 15-24) Adult (age15-64) Source: Authors’ estimations based on IHS 2016/2017 10. Although the overall literacy of the workforce is increasing, learning outcomes in schools remain poor. Despite the high rate of repetition, students reaching Grade 4 have learned little. Figure A2 (in Annex 1) shows the share of Grade 4 students achieving a range of basic literacy and numeracy tasks in the learning assessments conducted as part of the Malawi Longitudinal Schools Survey (MLSS) in 2018/19. The findings suggest that Grade 4 students struggle to complete even tasks aligned with the Grade 1 and 2 curricula.16 Even by Grade 6, only 44 percent of students reach the minimum proficiency level in reading in the most recent Southern and Eastern Africa Consortium for Monitoring Educational Quality assessments, conducted in 2013 (SACMEQ IV).17 11. Primary learning outcomes are highly inequitable between schools. In the MLSS learning assessments conducted in 2016, the difference in average learning outcomes between the highest- performing 10 percent of schools and lowest-performing 10 percent was equivalent to a difference of almost two years in learning.18 These large disparities persist across Chichewa, Math and English. These disparities in learning between schools do not occur primarily between regions or districts: of 34 education districts, 22 have an average student score in the MLSS learning assessments within 10 percent 16 Only 31 percent of tested Grade 4 students could add simple two and three-digit numbers, while only 39 percent could subtract three-digit numbers. In English, only 24 percent could select the correct form of the verb ‘to be’ to use in a sentence and only 20 percent could correctly write the plural form of the word ‘mango’. 17 6 percent of pupils are at the pre-reading proficiency level, 21 percent are emergent readers, and 29 percent are at the basic reading level. 18 The MLSS ascribes a knowledge score to each student and subject, with a mean of 500. On average, students in the sample group had a composite knowledge score of 502, but the average was 581 or above in the top 10 percent of schools, while the average was below 414 in the bottom 10 percent. of the national average. Instead, variance in school average learning outcomes occurs mostly at a micro- geographic level, suggesting that school characteristics account for most variation. Regression analysis confirms that school conditions, staffing and student characteristics explain most variation in learning. 12. Despite these significant disparities between schools, the majority of variation in learning – 73 percent – is between individual students within a single class. In 86 percent of schools, the gap between the highest and lowest performing students in a single Grade 4 class in the MLSS learning assessments was equivalent to more than two years of schooling, effectively creating multi-grade classrooms. In particular, girls, overage students, and those without literate parents perform significantly less well. A similar pattern is observed in SACMEQ IV with Grade 6 boys and students from the least poor households outperforming girls and poorer students; and in the Primary School Leaving Certificate Examination (PSLCE), conducted in Grade 8, where 72 percent of girls passed in 2019 versus 82 percent of boys19. 13. At secondary level, inequities in learning are primarily between government day and boarding schools, community day secondary schools, and open secondary schools. Public secondary schools take four distinct institutional forms: government boarding schools; government day schools; community day secondary schools (CDSS), operated under public-community partnerships; and open secondary schools (OSS), run in the evening typically on the sites of CDSS. Students with the highest grades in the examination at the end of primary school are selected to join the government boarding and day schools (at national and regional levels) and then CDSSs. Growth in secondary enrollment has been dominated by OSS and CDSS in recent years (see Figure A3, in Annex 1). 14. Student learning at the secondary education level is inequitable by school type, and gender, based on the Malawi School Certificate of Education (MSCE). Figure A4 (in Annex 1) shows that about 60 percent of secondary education students passed the end cycle MSCE exams in 2017, with girls registering much lower performance (53 percent) compared to boys (67 percent). There are also variations by type of public school.20 External efficiency 15. The average educational attainment for Malawi’s working age population is less than a completed primary education, which raises concerns about economic growth and global competitiveness. The working age population tends to have lower levels of educational attainment. In 2016/17, the average working age adult had 6.5 years of schooling, 7.3 among the youth population (age 15-24), and 5.9 among the older cohort. This suggests that the labor market comprises workers with low skills who have not completed primary education and therefore lack adequate literacy and numeracy skills (Figure 10). Malawi National Examination Board (MANEB) 20 Lowest performance was registered by students in the CDSS where 49 percent passed MSCE with significant variations between girls (39 percent) and boys (57 percent). While most OSS are hosted by CDSS, students in the former outperformed the latter by 7.3 percentage points; and the gender gap in performance at the MSCE is also narrower (11.3 percentage points in favor of boys in the former compared to 17.9 percentage points in favor of boys in the latter). Government day schools (71.8 percent) significantly outperform both CDSS (49.4 percent) and OSS (56.7 percent) at the MSCE but with wider gender differentials in student learning (16 percentage points in favor of boys). Figure 10: Average years of schooling Percent 10 9.0 8.1 7.3 7.1 7.1 8 6.5 6.2 6.4 5.8 5.9 5.8 6 4.7 3.9 4 2 0 Total Urban Southern Rural North 15-24 25-34 35-44 45-54 55-64 Male Central Female Age group Age 15-64 Source: Authors’ estimations based on IHS 2016/2017 16. The estimates of the returns to an additional year of schooling in Malawi are positive. At the national level, an additional year of schooling yields a 14 percent return (Figure 11), with slightly higher returns for women than men. This may reflect the fact that there are fewer educated women than men in the labor force. In comparison, returns to an additional year of schooling are 10.5 percent in Sub- Saharan Africa and 8.0 percent in advanced countries.21 The rate of returns on education in the industrial sector is higher than in the other sectors, reflecting a higher overall rate of productivity in that sector. 22 Figure 11: Returns to additional year of education Percent 25 20 20 14 15 15 14 15 10 10 5 0 National Male Female Agriculture Industry Services Source: Authors’ estimations based on IHS 2016/2017 Expenditure Trends and Funding Sources 17. Private households’ out-of-pocket contributions fund 39 percent of the total education spending in Malawi while the Government contributes 46 percent and development partners the remaining 15 percent. Figure 12 presents the sources of finance and Figure 13 presents the breakdown 21 Psahcharopoulos, G. and Patrinos, H. 2018. “Returns to Investment in Education A Decennial Review of the Global Literature.” World Bank Policy Research Working Paper 8402. Available at: http://documents.worldbank.org/curated/en/ 442521523465644318/pdf/WPS8402.pdf. Accessed: 1 May 2020. 22 The rate of return on education increases with each successive level of education attained in Malawi, confirming that education is key in breaking poverty cycles (Figure A5 in Annex 1). At the national level, the rate of return ranges from 33 percent for some primary education to 192 percent for post-secondary education, compared to those with no education. Higher education leads to remarkably high returns in all categories: 199 percent for males, 172 percent for females, 203 percent for those that work in the private sector, and 262 percent in the industrial sector, which may reflect the low overall level of participation in higher education and the resulting scarcity of higher education graduates in the workforce. by level of education. In 2016/17, the total cost of the education sector in Malawi amounted to US$ 609 million (MWK 450 billion), with US$ 287 million of the contribution coming from the Government followed by US$ 239 million from households and the remaining US$ 91 million from development partners. 23 Public spending increased yearly between 2017 and 2019. Total public spending on education increased from US$ 287 million in 2017 to US$ 408 million in 2019. Donor contributions on the other hand increased from US$ 91 to US$ 132 million in 2018 but decreased to US$ 122 million in 2019. Figure 12: Sources of education finance Figure 13: Breakdown by level of education Million US$ Breakdown by level (2016/17) Percent 700 617 2016/17 2017/18 HH Public 90% 82% 600 2018/19 500 80% 408 68% 400 357 70% 287 300 239 60% 52% 200 132122 48% 91 50% 100 40% 32% 0 30% 18% 20% 10% 0% Primary Secondary Tertiary Source: Authors’ calculations based on the Ministry of Finance (MoF) Budget (2016/17-2018/19) and IHS 2016/17. The World Bank, International Monetary Fund and the Reserve Bank of Malawi for exchange rates. Note: Pre-primary level not included. Data on household contributions only available for 2016/17. 18. The breakdown by level of education shows that primary education is most dependent on Government expenditure, while households directed their support to the secondary and tertiary levels. Malawi has had an official policy of free primary education since 1994; unsurprisingly therefore, 82 percent of primary education is funded by the Government; the corresponding contributions to secondary and higher education are 48 percent and 32 percent, respectively. Households’ highest contribution is at the tertiary education level with 68 percent compared to 18 percent in primary and 52 percent in secondary education. A policy of free secondary education, announced in 2018, is likely to contribute to rapid rises in the Government’s share of secondary spending in the next few years. 19. Total Government spending on education increased steadily between 2015 and 2019. Education as a share of GDP increased from 4 percent in 2015 to 5.1 percent in 2019. As a share of total Government spending, education expenditure fluctuated between 16 percent in 2014 and 21 percent in 2019 (Figure 14).24 However, the 2019/20 allocation fell to 18 percent of the total Government budget, and in general, Malawi’s spending has not consistently been in line with the Global Partnership for Education guideline that 20 percent of the national budget be allocated to the education sector. The 23 The main development partners contributing to the education sector are: United States Agency for International Development (USAID, 25 percent of development partner contributions in 2018); the European Union (23 percent); Royal Norwegian Embassy (RNE, 15 percent); Germany (GIZ, 8 percent, and KfW, 5 percent), and the United Kingdom Department for International Development (DfID, 7 percent); the World Bank Group (9 percent); and the Global Partnership for Education (8 percent). 24 The following votes make up the total education budget used in the analysis: 250- MoEST (primary, secondary, teacher training colleges and administration); 275- subventions; 320- ECD; 370- TVET. Malawian Government’s spending on education is slightly higher than the Sub-Saharan Africa average. 25 This is similar to health expenditure (pg. 88), but contrasts with Malawi’s expenditure on social safety nets, which is just over half of low-income countries in Sub-Saharan Africa (pg. 72). Box 2. Impacts of COVID-19 The Government of Malawi responded to the COVID-19 pandemic by closing all schools and universities from late March 2020 to September 2020, with all grades not returning to school until October and schools continuing to operate on a reduced schedule as of November. Vulnerability in the Malawi education system comes in many forms: low-income students, students in remote rural areas, and families that rely on school feeding programs to provide necessary nutrition to children. The Government prepared a National Preparedness and Response Plan for key sectors. The education plan comprised short-term measures including radio and TV- based educational programming; care and support for orphans and vulnerable children, including provision of take-home rations; and remedial classes for returning students following the resumption of schooling. In the medium term the Government worked to provide better hygiene and sanitation facilities to create readiness of the system following the reopening of schools. The direct costs of these activities were largely financed by donors, notably the Global Partnership for Education (GPE), UNICEF and Save the Children. The loss of learning from the closure of schools has the potential to severely impact long-term learning trajectories, and subsequently lifetime earnings. World Bank analysis suggests that globally, school closures of 3-5 months – less than the length of the Malawi closure – would produce a reduction in student learning outcomes of 7 percent, from 7.9 learning-adjusted years of schooling (LAYS) to only 7.3, assuming some successful mitigation of learning loss from distance learning and catch-up learning following the reopening of schools. In Sub-Saharan Africa, where the average child could be expected to complete only 4.9 LAYS prior to the pandemic, it is projected that this could fall to as low as 4.3 LAYS with an extended closure of five months (World Bank Group, 2020). Simulations suggest that COVID-19 will cause 6.8 million children to drop out of school around the world, and the loss of lifetime earnings from dropout and reduced learning achievement will range from between $130 to $375 in Sub-Saharan Africa depending on the length of school closure and effectiveness of remediation. In the medium term education financing is likely to be negatively affected by the economic shock associated with COVID-19. Although the duration and severity of the economic shock caused by the pandemic is uncertain, it is likely to be significantly larger than anything seen since the financial crisis of 2008/09. Real GDP growth in Sub-Saharan Africa (SSA) is expected to be flat or negative in 2020; this slowdown in growth, combined with economic recovery and stimulus packages, will have significant negative impacts on public revenue, which is forecasted at between 12-16 percent lower in SSA in 2020 than 2019. Estimates suggest that slow growth could mean that public education expenditure in SSA could fall by between 4 and 5 percent in 2020. The overall COVID- 19 impact on education spending per child is likely to be larger because these simple estimates do not take into account declines in Government revenue and reprioritization of Government budgets toward health and social protection which will reduce Government education spending further. The pandemic will also result in a massive income and health shock for households, reducing available funding for education expenditure. These pressures are likely to exacerbate the long-term fiscal pressures on the Malawi education system, with particularly negative impacts on schools which already face severe disadvantages as a result of inequities in service delivery. Issues regarding equity – that is, ensuring that the needs of those who are at risk are met – should be at the center of the Government’s response, both during school closures and after schools reopen. Most notably, to narrow the equity gap resources need to be targeted to schools and students that need them the most – for example, ensuring that an adequate number of teachers are available to support students in remote rural areas. There is also an urgent need to protect and ring-fence education budgets and expenditures at the recently devolved district council level, to enable the payment of teacher salaries, and that primary school improvement grants for schools are ready and support preparedness when schools reopen. In the context of Malawi, this will require urgent attention to the lack of transparency in education expenditure at the district level. 25 Figure A6 in Annex 1 shows public education expenditure as a share of GDP and as a share of total public expenditure for 23 SSA countries. Malawi’s spending on education as a share of GDP (5.0 percent) is higher than the Sub-SSSA average (4.2 percent). 20. Development partner sector financing remains substantially off-budget following financial scandals. Prior to 2013, much financing from development partners was channeled through a Sector- Wide Approach (SWAp) fund, through a Memorandum of Understanding (MOU) and Joint Financing Arrangement (JFA), and consequently included within the MoEST budgets. Following the high-profile ‘Cashgate’ scandal, which centered on allegations of financial impropriety involving multiple Government ministries, several donors withdrew their support for the pooled financing mechanism, with the result that development partner support to the sector became primarily off-budget. In 2017, the development partners and the Government agreed a new common financing mechanism, including an Education Services Joint Fund (ESJF) as a channel for coordinated expenditure and a Common Fiduciary Oversight Arrangement (CFOA); however, the ESJF remains off-budget and under the CFOA, expenditure is managed by an independent auditor who acts as a manager and signatory on funds. As a result, the majority of development partner expenditure remains off-budget (81 percent in 2018/19). This situation in education is mirrored in other sectors, notably health where 74 percent of donor funding is off-budget (pg. 53). The development of the common financing mechanism stands in contrast to the social protection sub-sector, where donor support remains highly fragmented (pg. 91). 21. Malawi’s program of Early Child Development (ECD) is particularly dependent on support from development partners. Approximately 80 percent of financing for nutrition in Malawi is provided by donors,26 and is outside the programs of the budget. The Government’s contribution (0.5 percent of education spending) primarily includes the costs of personnel and infrastructure. The ECD program is implemented through a network of national and international nongovernmental organizations (NGOs). Most donor funds to support nutrition are executed through UNICEF, which supports the implementation of the revised integrated ECD policy (2017–2022) and coordination across key ministries. School feeding programs at primary level are also predominantly donor-financed. Figure 14: Trend in share of education budget Allocation in MWK (Billions) (lhs); Percent (rhs) Percent Total education sector allocation(MK) Education share of government budget 350 25% 3% 2% 3% 3% 2% Education share of GDP Management 300 30% 30% 24% 28% 26% 20% Higher 250 Education 20% 13% 14% 200 15% 14% 18% Secondary Education 150 10% Primary 100 53% 49% 53% 56% 58% 5% ECD 50 0.3% 0.2% 0.2% 0.5% 0.0% 0 0% 2014/15 2015/16 2016/17 2017/18 2018/19 2014/15 2015/16 2016/17 2017/18 2018/19 Source: Authors’ calculations based on the Ministry of Finance (MoF) Budget (2015- 2019), The World Bank and The International Monetary Fund for GDP. Note: ECD shown at bottom (right panel) 22. Household spending on education is primarily at the higher education level, accounting for US$ 143 million of the total US$ 239 million in 2016/17 (60 percent). Despite an official policy of free primary education, households invested US$ 34 million in primary schooling (approximately US$ 7 per enrolled student) and US$ 62 million in secondary schooling (approximately US$ 160 per student) 26 Including Irish AID, DFID, JICA, USAID, Germany, European Union, and the World Bank (Figure 15). Around half of household expenditure on education is on private schooling, and particularly high at pre-primary and secondary levels (91 percent and 54 percent respectively), reflecting weaknesses in public provision (Figure 16). Figure 15: Total spending of household by Figure 16: share of household spending in school type private schools Million US$ Percent HH spending 280 100% HH spending in public schools 91% HH spending in private schools 230 Public spending 80% 180 151 143 60% 54% 49% 130 40% 40% 68 80 62 57 20% 34 33 21 21 30 14 0.6 0.5 0.4 0.0 0% -20 Pre-primary Primary Secondary Tertiary Pre-primary Primary Secondary Total Source: Authors’ estimations based on the Ministry of Finance (MoF) Budget (2016/17), IHS 2016/17. Note: The IHS does not classify tertiary education into private or public 23. Public spending allocation by level of education is biased in favor of post-primary and tertiary education. Figure 18 show the enrollment distribution of students across all education levels and the corresponding funding allocation for households and the Government. Only 53 percent of public spending was allocated to primary level in 2016/17 – even though it accommodates about 87 percent of the total enrollment. Secondary was allocated 20 percent, and higher education 24 percent. 27,28 In particular, the share of public expenditure channeled into higher education is disproportionate to the share of enrolled students at that level, especially in contrast to the primary education sector. Indeed, higher education accounted for 24 percent of total public education expenditure with only 1 percent of student enrollment. In contrast, households spend about 15 percent on pre-primary and primary education, 26 percent on secondary and a significant amount on tertiary (60 percent). Comparing household spending on public non-tertiary education, 29 spending is split almost equally between primary and secondary education, but the pattern reverses for spending on private education where households spend about 70 percent on secondary education and 29 percent on primary education. Overall, about 34 percent of total education spending goes to pre-primary and primary followed by 22 percent to secondary and 42 percent to higher education. Sectoral allocations of education spending in Malawi are similar to those in other SSA countries, despite lower levels of post-primary enrollment (Figure A 7A, in Annex 1). At the tertiary level, per student public spending in Malawi is extremely high at 840 percent of GDP per capita, compared to the SSA average of 200 percent. These findings highlight the need for sectoral prioritization in public spending (Figure A7B). 27 Management and administration account for 3.18 percent of the total education budget. This does not include donor external resources. 28 In 2018/19, the share of public expenditure for primary was slightly higher at 58 percent, secondary remained at 20 percent, and tertiary rose slightly to 26 percent. 29 Tertiary education for household spending is not classified into public or private. Figure 17: Share of spending by public and Figure 18: Enrollment distribution by level household Share of spending Distribution of enrollment Share of public spending Share of HH total spending 86.9% Share of HH spending on public schools 100% 100% Share of HH spending on private schools 70.2% 80% 80% 59.6% 52.7% 50.4% 49.6% 60% 60% 28.9% 40% 25.8% 23.8% 40% 20.1% 14.3% 11.6% 20% 20% 0.8% 0.7% 0.9% 0.2% 0.2% 0.1% 0% 0% Pre-primary Primary Secondary Tertiary Pre-primary Primary Secondary Tertiary Source: Authors’ calculations based on the Ministry of Finance (MoF) Budget (2016/17), IHS 2016/17. The IHS does not classify tertiary education into private or public 24. Finance for secondary education has not yet been adequately enhanced to replace lost revenue from household contributions under the free secondary education policy. Since the official introduction of free secondary education in 2018, payment of tuition fees to secondary schools is no longer required, and other sources of household contributions, known as School Development Funds or Parent-Teacher Association (PTA) funds, have been substantially reduced. However, secondary school head teachers responding to interviews as part of this study report that finance from the Government has not fully replaced these lost sources of revenue, exacerbating fiscal challenges with negative impacts on the supply of teaching, learning and other materials. In addition, payments from the Government are more prone to delays than community contributions, increasing the burden on school finances. 25. Supporting the continued expansion of primary enrollment, while raising input standards, is likely to entail rapid increases in development expenditure in the coming years. World Bank modelling of future enrollment trends at the primary level projects public primary enrollment to rise by around 10 percent by 2024/25.30 Accommodating this expansion while improving service standards to meet policy commitments is likely to require significant increases in primary expenditure (Figure 19) and in particular, efforts to reduce infrastructure costs.31,32 This trend is likely to continue in the subsequent years as a result of rapid population growth, with Malawi’s population set to double between 2015 and 2038 (pg. 70). Careful consideration of unit costs and infrastructure investments is required as part of 30 We employ population projections from the United Nations Department of Economic and Social Affairs, adjusted to reflect school- age population as reported in the UNICEF Multiple Indicator Cluster Survey, and recent trends in gross intake ratio, promotion and repetition rates. 31 Cost modelling includes teacher salaries, classroom and latrine construction; but does not include school grant finance, inspections, examinations, district and sub-district officials, administration costs, or teacher training. 32 For example, to lower the national average pupil-classroom ratio to the policy goal of 60 students per classroom at primary level by 2025, while accommodating additional enrollment, would require the construction of 70,000 additional classrooms; at the current unit cost of US$ 20,500 per classroom, this would entail a total cost of US$ 1.4 billion even before accounting for additional teaching staff. However, evidence from a pilot of supplemental school discretionary grant finance as part of the Malawi Education Sector Improvement Project (MESIP) suggests that community-constructed, low-cost learning shelters can be constructed for less than US$ 3,000, enabling a saving of US$ 1,230 million by 2024/25. the implementation of the new National Education Sector Investment Plan in order to ensure a fiscally sustainable path for primary expansion. Figure 19: Annual expenditure, primary education, 2020-25 Billion, MWK 1,000 Recurrent cost Development cost 800 600 400 200 0 2020 2021 2022 2023 2024 2025 Source: Author’s calculation based on fiscal simulation tool (see Annex 2). Maintains current unit costs for salaries and construction of classrooms and latrines 26. At the secondary level, too, the expansion of the system poses a severe fiscal challenge. Maintaining current public unit costs, achieving the expansion of secondary access to 90 percent of secondary school-age children would increase the annual cost of secondary education by around US$ 107 million, leading secondary education costs to account for more than one-third of total education expenditure (Table 1). Such significantly higher resources cannot be raised by the Government alone. It is imperative that multi-pronged approaches are developed for the expansion of secondary education, hence the relevance of this expenditure review. Table 1: Projected Financing Gap - Achievement of Secondary Education for All in Malawi Expenditure % Government Expenditure Scenarios % GDP (US$) expenditure Education budget 2018/19 (all sub-sectors) 468 million 6.61% 23.80% Secondary education budget (2018/19) 53 million 0.08% 2.70% Secondary education spending if 90% enrollment, using current per student 160 million 2.26% 8.15% spending (US$ 115) Secondary education spending if 90% enrollment, using lower bound student 140 million 1.98% 7.10% spending (US$ 115) 27. Enhancements to the efficiency of secondary education spending, particularly with regard to utilization of teachers, can reduce the cost curve. Potential efficiency enhancements include: raising secondary teacher loads, with a minimum teaching load set and enforced; and selective use of double shifts in schools, building on the widespread use of OSS which operate in the evenings. More generally, the secondary sector is likely to benefit from the expansion of the OSS system, supported by new online content delivery systems; and optimization of placement procedures to ensure capacity within existing schools is not wasted. Attention would also need to be paid to disparities in service delivery and outcomes between the Government day and boarding schools and CDSS and OSS. Box 3. Malawi’s early marriage challenge While Malawi has made progress in reducing child marriages, the rate of progress is still much too slow. On average, the prevalence of child marriage decreased by 7.6 percentage points in 15 East and Southern African countries over two and a half decades. For Malawi, the decline in the rate was only slightly lower, at 7.4 percentage points. However, Malawi started from a particularly high base, so that the rate of prevalence in the country is still 10 points above the regional average (Male and Wodon, 2018). The prevalence of early childbearing is also high at 27 percent of girls aged 18-22 (Demographic and Health Survey, 2018). Analysis of the timing of marriage and childbearing suggests that child marriage is the likely cause of around two-thirds of cases of girls having children before the age of 18 (Wodon et al., 2020). Malawi has already taken a number of steps to reduce or eliminate child marriage and to improve opportunities for girls, including an amendment to the Constitution in 2017 to make child marriage illegal. While these measures should be applauded, they are not sufficient. The negative impacts of early marriage and childbirth High rates of early marriage and childbirth are a key driver of low educational attainment for girls. Parents often attribute their daughters dropping out of school to marriage and pregnancies (McConnell and Mupuwaliywa, 2016). Additional econometric work suggests that the causality between child marriage and early childbearing on the one hand, and girls’ educational attainment on the other hand, is both strong and bidirectional (see for example Field and Ambrus, 2008; Nguyen and Wodon, 2014; and Wodon et al., 2016). The effects of early childbearing compound across generations, with children born to young mothers with low levels of educational attainment being significantly more likely to themselves also achieve lower levels of educational attainment. Early marriage and childbearing also negatively affect Malawians’ health outcomes and earning potential. Children of underage mothers are more likely to die in infancy and to be stunted. Being born of a mother younger than 18 increases the risk of dying before the age of 5 by 4 percentage points; ending all early childbirths would reduce under-five mortality by 0.35 percentage points nationally. Being born of a mother younger than 18 increases the risk of under-five stunting by 7 percentage points, and ending all early childbirths would reduce under-five stunting by 0.44 percentage point nationally. Ending child marriage could increase earnings in adulthood for women marrying early by up to 10 percent, and increase earnings and productivity nationally by up to 1.6 percent (Wodon et al., 2018). In Malawi, the total value of potential earnings lost due to early marriages in 2015 stood at US$167 million (PPP). Early marriage and childbearing also negatively affect Malawi’s economic growth by contributing to unsustainable levels of population growth. Estimates suggest that the elimination of child marriage and early childbearing in Malawi could reduce the country’s annual rate of population growth by 0.21 percentage points. Even greater reductions could be achieved with universal secondary education completion for girls, but the effects are not computed here. These effects have major implications for Malawi's ability to reap the benefits from the demographic dividend (for a more detailed discussion of this dividend, see Canning et al., 2015). Combining these impacts with the economic impacts of poor health outcomes and lower earnings suggests that ending child marriage and early childbearing in 2015 would have generated annual benefits of US$ 23 million in purchasing power parity (PPP) immediately, increasing to US$ 0.5 billion by 2030. Thus, the welfare benefits derived from lower population growth due to the elimination of child marriage and early childbearing are extremely significant. Potential policy responses to reduce early marriage and childbirth The best way to end child marriage and early childbearing is through measures to ensure that adolescent girls remain in school. With the strong correlation between educational attainment, child marriage, and early childbearing, it is clear that efforts to ensure that girls remain in school or go back to school if they dropped out appear to be among the most effective interventions to reduce the prevalence of child marriage and early childbearing. In particular, achieving universal secondary completion for girls could dramatically reduce the prevalence of child marriage and early childbearing. Keeping girls in school requires improving school learning environments and may also require incentives. Secondary education completion rates remain low at least in part because there are just not enough secondary schools to support universal completion, especially in rural areas. Schools also need to provide access to water, latrines and hygienic facilities, which may be particularly important for adolescent girls. If schools are not available in a community, it may be necessary to provide modes of transportation to enable girls to attend school. It is essential to ensure that girls do not suffer physical, sexual, or other harassment either at school or while travelling to and from school (see for example, Abramsky et al., 2014, on gender-based violence and the means to reduce it). In many communities, the economic, cultural, and social environment does not provide viable alternatives to marriage for adolescent girls. Once girls drop out of school, possibly because of the poor quality of these schools or their high cost, it may be difficult for parents to consider any option other than marrying off their daughters. In those communities, improving the provision of quality and affordable primary and secondary education may be one of the best ways to delay marriage and childbearing, as parents often see schooling as a viable alternative to marriage for their daughters. Incentives such as bursaries/scholarships, and other programs to keep girls in school, may also lead to “tipping points” in communities whereby more and more girls remain in school and are able to delay marriage. A few interventions have also aimed to delay marriage by providing financial incentives conditional on not marrying early, with additional schooling often an additional benefit (Botea et al., 2017). Empowering girls by providing them with life skills and reproductive health knowledge can be a powerful means to reduce early marriage and childbirth and maintain girls in school. One typical intervention involves the provision of a “safe space club” for adolescent girls. These clubs serve as delivery platforms for convening girls under the guidance of a trusted adult mentor and can be held in a variety of settings, including schools or community centers. Girls meet regularly and are able to discuss a range of issues under the guidance of the mentors, including those related to sexual and reproductive health. The clubs facilitate the delivery of life skills, including “soft” or socio-emotional skills such as critical thinking and problem solving, communication and negotiation (for example within the girl’s household). By increasing girls’ level of knowledge, life skills can raise these girls’ awareness of the risks associated with becoming pregnant at an early age and increase their desire and ability to avoid early pregnancies (through family planning). By increasing girls’ confidence and self-esteem, life skills may also help expand girls’ aspirations, which could further increase motivation to delay marriage and childbearing. Finally, life skills can increase young women’s communication and decision-making skills, leading to increased abilities to negotiate their preferences for delayed marriage and childbearing. The clubs have proven to be effective when they are implemented well. At the same time, while life skills and sexual and reproductive health knowledge may empower girls, they may not be sufficient to delay marriage and childbearing if social norms curtailing agency for girls are not also addressed at the same time. Efficiency of public expenditures Budget allocation and execution 28. There are large discrepancies between the allocated (voted) budget and the executed budget, especially for capital expenditures. Figure 20 shows trends in budget allocation and Figure 21 shows execution rates by category. This figure shows that, while the allocation to capital spending over the period averages 18 percent, the execution rate is low and averages 45 percent.33 This is in contrast to the execution rate for personnel spending, which is about 100 percent, suggesting that the low levels of budget execution are almost solely due to the poor execution rates of the capital spending category. 33 Data on capital budget for 2017 missing due to incomplete data. Figure 20: Budget allocation Figure 21: Execution rate Billion, MWK Percent Total education budget Share of recurrent in total public spending Total executed education budget Personnel budget Share of executed recurrent budget on education 300 Personnel executed budget 160% Share of recurrent in education budget Share of executed capital budget on education 140% 120% 119% 200 100% 100% 89% 90% 95% 93% 88% 84% 83% 80% 78% 76% 77% 73% 71% 75% 60% 100 46% 46% 48% 40% 40% 20% 0 0% 2015 2016 2017 2018 2019 2015 2016 2017 2018 2019 Source: Authors’ calculations based on the Ministry of Finance (MoF) Budget (2014/15 to 2018/19) Note: 2017 omitted from right panel owing to incomplete data 29. The low execution levels of non-personnel spending adversely affect schools’ ability to finance non-personnel costs. Table A1, in Annex 1, shows the functional allocation and execution budget between 2015 and 2019. The budget execution rate for the personnel allocation for the period is high but declined from 140 percent in 2015 to 102 percent in 2019.34,35 This matches a similar trend in health spending (pg. 97). Decentralization reforms, under which teacher payroll has been moved to district levels, are expected to have supported strong execution of personnel budgets beginning in 2019.36 Development allocations increased significantly in 2017 (17 percent) and in 2019 (25 percent). The execution rate for development spending fluctuates between 40 and 48 percent. Low execution of construction spending could affect access to schools and other essential facilities for learning. Current evidence shows that school construction programs that increased school access significantly improved the human capital outcomes of beneficiaries in Indonesia, India, and Nigeria. Evidence from the MLSS demonstrates that the typical (median) Malawian primary school has a pupil-classroom ratio (PCR) of 95, and 29 percent of primary schools conduct at least one class in the open air, suggesting an urgent need to improve execution of, and increase, capital allocations. 30. Low allocations for other areas of spending, combined with high execution rates, could be indicative of the need to increase allocations in these areas. 37 Goods and services accounted for about 7 percent throughout the same period, while the bottom panel shows execution rates hovered between 42-55 percent. Grants to schools account for 19-29 percent of budgets and are highly executed, with execution averaging about 100 percent over the period. 34 For 2016/17 due to budget devolution, Vote 250 did not contain total personnel allocation at the primary level therefore revised estimates were used instead of actual. According to EMIS 2017, execution rate for personnel was 100 percent. For 2017/2018 and 2018/19, personnel allocations for primary education were obtained from budget analysis by KfW. 35 Evidence from the 2016 baseline of the MLSS suggests that teacher salaries (which account for the majority of personnel expenditures) were prone to severe delays at the time, with 78 percent of teachers experiencing at least one delay that year. 36 The revised budget for primary school payroll expenses was used instead of the actual expenditure because of the budget devolution which started mid 2016/17. 37 This excludes other recurrent transactions to district councils for education. Internal inefficiency related to repetition and dropout 31. High rates of repetition and dropout pose large costs to the public exchequer and to households particularly at primary level. Estimates of the cost to education expenditure from school repetition and dropout show lost public expenditure equivalent to 1.4 percent of GDP (MWK 64 billion) and lost household expenditure of 0.6 percent of their total consumption expenditure (MWK 19 billion). The public sector loses the equivalent of about 51 percent of total recurrent spending on primary and secondary education as a result of repetition and dropout.38 Similarly, households lose about 27 percent of their total current spending on education at the primary and secondary levels. This lost income and foregone output, over the lifetime of the affected children, is equivalent to 2.3 percent of GDP in 2017. Households also lose 3.2 percent of their total consumption expenditure in today’s terms as the result of dropout and repetition (Table 2).39 Table 2: Costs of internal efficiency and dropout rates (MWK thousands) As share As share As share of of Repetition Dropout Total of GDP consumption recurrent spending Public spending Primary 5,836,934 27,593,637 53,430,571 52% Secondary 1,010,770 9,351,836 10,362,606 45% Total 6,847,704 36,945,473 63,793,177 1.39% 51% Household spending Primary 6,303,378 6,658,016 12,961,394 48% Secondary 841,480 5,570,705 6,412,185 14% Total 7,144,858 12,228,721 19,373,579 0.58% 27% Forgone earning Primary 11,421,319 12,063,900 23,485,219 Secondary 114,649 758,989 873,638 Total 11,535,968 12,822,889 24,358,857 0.73% Total Primary 3,561,630 46,315,553 89,877,183 Secondary 1,966,899 15,681,530 17,648,429 Total 5,528,529 61,997,083 107,525,612 Share of 0.99% 1.35% 2.34% 2.34% 3.22% GDP Source: Authors’ calculation based on Ministry of Finance (MoF) Budget, IHS 2016/2017, EMIS 2017 38 Consumption refers to total household expenditure. Spending refers to expenditure at different levels of education. 39 This was calculated using the number of dropouts and repetitions per year by level of education based on the unit costs per student payment. The cost was estimated from: (i) the direct cost of schooling which is generated from a total number of repeaters and dropouts and based on per-student annual unit cost, and (ii) the discounted value of foregone opportunity costs of expected earnings - estimated based on wage employment earnings by considering the age of labor market entry. Equity of public spending Role of Government in protecting equity 32. Overall public spending on education in Malawi is biased toward wealthier households as a result of highly inequitable allocations at the upper levels of education . Figure 22 shows the distribution of public spending across household wealth quintiles within each level of education in a Benefit Incidence Analysis (BIA) and Figure 23 show average household size in each quintile. The analysis of public spending across all education levels shows that the poorest quintile receives only 15 percent of total education spending, 5 percent less than its share in the population, while the wealthiest quintile receives 48 percent. At the primary level, public spending appears to be more equitable in the sense that the poorest quintile receives a high share of public benefits (24 percent), 4 percent more than their share of the population while the wealthiest receive 15 percent of the benefits (5 percent less than their population share). The right panel shows the average household size by wealth quintile. Poor households are generally larger (6.1 persons) than rich households (4.5 persons), suggesting that the benefits poor households receive could be reflective of most children in public primary schools coming from poor households. However, at upper levels spending becomes progressively more regressive, with the share of spending going to the richest quintile rising to 41 percent at secondary, and 89 percent at tertiary level. Figure 22: Benefit Incidence Analysis of public Figure 23: Average household size in each expenditure on education quintile Share of education spending by quintile Household size 90 89 Q1 Q2 Q3 7 6.1 Q4 Q5 6 5.4 70 5.1 4.8 5 4.5 48 50 4 41 3 30 24 22 23 21 18 2 15 16 15 16 13 12 13 10 7 7 1 2 2 0 -10 Primary Secondary Tertiary All Q1 Q2 Q3 Q4 Q5 Source: Authors’ estimations based on the Ministry of Finance (MoF) Budget, EMIS, IHS 2016/17 Note: *The primary classification includes preprimary and primary expenditures 33. The regressive nature of education spending at upper levels reflects inequities in enrollment at these levels. Access to public schools in Malawi is relatively pro-poor at primary level: 25 percent of children from the poorest quintile attend public primary schools versus 13 percent of children from the wealthiest quintile, as a result of higher private enrollment for these households (Figure 24). However, attendance is dominated by the wealthiest households at each successive level, and at tertiary level, almost all enrollment – 90 percent – is from the wealthiest households. This inequity of access, combined with comparatively high levels of public investment in secondary and tertiary education, create an imbalance in overall spending toward the wealthiest. Figure 24: School participation by quintile, public schools Percent 90% 100% Q1 Q2 Q3 Q4 Q5 80% 60% 33% 40% 25% 23% 23% 21% 20% 18% 14% 13% 20% 9% 8% 2% 1% 0 0% Primary Secondary Tertiary Source: Authors’ estimations based on IHS 2016/17 34. Richer households are more likely to send their children to private school at all levels of education. For example, 11 percent of children from the poorest quintile attend private primary schools versus 38 percent of children from the wealthiest quintile (Figure 25). Access to private education grows more restricted to the wealthiest households at each successive level, with 33 percent of private secondary enrollment coming from the wealthiest quintile. Figure 25: School participation by quintile, private schools Percent 50% Q1 Q2 Q3 Q4 Q5 38% 33% 40% 30% 23% 20% 18% 18% 14% 14% 20% 11% 9% 10% 0% Primary Secondary Source: Authors’ estimations based on IHS 2016/17 35. The distribution of public expenditure in primary education is relatively more biased toward the poor than the distribution of income in Malawi. However, the inequality becomes higher at upper levels of education. Incomes are substantially unequally distributed with the poorest quintile of households accounting for only 7 percent of income and the wealthiest quintile, 50 percent (IHS 2016/17). A concentration curve is an alternative way to use the BIA to evaluate the targeting of Government subsidies. Figure 26 includes the consumption concentration curve which is a proxy for the general wealth and income inequality across quintiles. It presents the analysis taking into consideration variations in the number of children by quintile. In general, public spending on education is pro-poor if the concentration curve for the particular level of education is above the 45-degree line. 40 Similar to the 40 The Lorenz curve is a graphical interpretation of the cumulative distribution of income on the vertical axis against the cumulative distribution of population on the horizontal axis. The progressivity of spending is pro-poor if the poor receive more of the program’s benefits than the non-poor and more than their share of the population; graphically this line appears above the diagonal since the 45” line indicates that each quintile in the distribution is receiving the same share, or in other words, each quintile (which represents BIA, only the expenditures on primary education are relatively more equitable than the general wealth distribution.41 Therefore, while public spending in primary education is not pro-poor per se, this is somewhat mitigated by the fact that the distribution of spending promotes greater equality than the general observed income distribution. In contrast, secondary education and especially higher education is significantly not pro-poor and is regressive. Given that the wealthiest quintile receives the most benefit from public spending – the distribution of public spending in secondary (for about 70 percent of the population) and higher education is, in fact, worse than the general wealth inequality. Figure 26: Lorenz Curve for public spending on education by level Source: Authors’ estimations based on the Ministry of Finance (MoF) Budget, EMIS, and IHS 2016/17 Unit cost analysis 36. Both public and private schools depend to varying extents on contributions from households to finance the education system. Figure 27 shows the public unit cost (total public expenditure/ total number of students at each education level) and household unit cost by level of education (total household spending on education / total number of students in each education level), the latter broken into spending on public and private schooling.42 Even at primary level, households contribute around one-seventh of unit costs for public schooling (US$ 5 versus US$ 30 public expenditure), higher at secondary level (US$ 52 versus US$ 194). This is surprising given the official policy of free education at primary (and, from 2018 onwards, secondary) schooling, and demonstrates the continuation of hidden costs for schooling such as uniforms, equipment, and informal fees such as those unofficially levied for participation in examinations. Household contributions are significantly larger at post-secondary level but are relatively low as a share of overall cost (US$ 467 versus US$ 2,657 in public spending). 20 percent of the population) would receive 20 percent of spending. Not-pro-poor but progressive is if the non-poor receive more than the poor, but still the poor receive a share larger than their share of consumption; graphically this line appears below the diagonal but above the Lorenz curve. Not-pro-poor and regressive occurs if the non-poor receive more than the poor, and the share of the poor is less than their share of consumption; graphically this line appears below the diagonal and below the Lorenz curve. 41 This can be observed from the concentration curves in primary lying above that of the consumption curve, indicating that spending tends to be more equitable. The curve of lower secondary partly lies below and above the consumption curve. 42 Public unit cost is total Government spending (recurrent and development) at that level of education divided by public student enrolment at each level of education. Figure 27: Total unit cost comparison by level of education and type of schools attended US$ 250 3000 Public unit cost HH unit cost Public unit cost 194 2500 200 HH unit cost in public school 154 2000 2657 150 HH unit cost in private school 1500 100 1000 52 38 467 50 30 500 5 0 0 Primary Secondary Tertiary Source: Authors’ estimations based on Ministry of Finance Budget, EMIS, IHS 2016/17 37. Private school unit costs are higher than those at public schools at primary level. The total unit cost of private primary schooling is US$ 38, versus US$ 35 for public schools (of which US$ 30 from public expenditure and US$ 5 from households). This represents significantly larger per-student expenditure in these schools. However, at secondary level, the unit cost of private schooling is US$ 154 versus a total unit cost in public schools of US$ 256 (US$ 194 from public expenditure and US$ 52 from households). Affordability of schools and the role of households 38. On average, households in Malawi devote 16 percent of their per capita spending to education. Figure 28 shows education spending as share of household per capita spending, and of total household consumption, by household wealth quintile. Poorest households allocate 13 percent of their per capita spending to education. This is a higher rate than the next three wealth quintiles, suggesting the burden on the poor is high relative to their income. Given the low enrollment rate at upper levels from these households, this demonstrates that despite the free primary education policy, primary schooling represents an inequitable cost burden on these households. Wealthiest households, on the other hand, devote 32 percent of their per capita spending to education. This rate could reflect the fact that tertiary education is expensive and is dominated by the wealthiest. Figure 28: Education spending as share of household per capita spending and of total household consumption, by household wealth quintile Percent 40% Share in per capita spending Share in total consumption 32% 30% 20% 16% 13% 12% 11% 10% 10% 10% 4% 2% 2% 2% 2% 0% Malawi Q1 Q2 Q3 Q4 Q5 Source: Authors’ estimations based on IHS 2016/17 39. The household unit cost in higher education is very high making higher education difficult for poor children to attend given the resources available. Figure 29 demonstrates the extent to which tertiary education is dependent on household contributions from the wealthiest three quintiles. The cost of education is significantly driven by fees, especially for post-primary education levels and interventions that reduce fees would create opportunities for the poor to access education services. School fees constitute the majority of household education spending at all levels except primary, and across all quintiles (Figure A8, in Annex 1). Figure 29: Household education expenditures per child by household wealth quintile Total payments Fees only Primary Secondary Tertiary Primary Secondary Tertiary 470 491 467 395 500 380 370 370 400 271 285 300 270 208 168 200 170 114 98 65 100 60 70 47 37 41 36 31 29 30 24 22 16 7 6 5 4 3 3 2 1 1 1 0 -30 Q1 Q2 Q3 Q4 Q5 Total Q1 Q2 Q3 Q4 Q5 Total Source: Authors’ estimations based on IHS 2016/17 40. Survey data suggest that cost is a major driver of children being out of school, although not the largest cause. Financial constraints rank high among the reasons listed for children being out of school across areas of residence, gender, and consumption quintile (Figure A9, in Annex 1). However, the breakdown shows disinterest in school or laziness is the most frequently cited driver of non- participation at all ages; this may in part reflect student responses to issues in school quality, or simply perceptions by parents. Illness or disability is another important factor contributing to children being out of school. It accounts for 18 percent at the primary school level and about 8 percent at the secondary level. This underscores the need for continued attention to nutrition in early years, 43 and functioning and accessible healthcare facilities. In secondary school, teenage pregnancy is another factor that keeps girls out of school, accounting for the third-highest reason among girls of secondary school age. School-level inequities in financing 41. Grants to primary schools may be inadequately needs-based. School Improvement Grants (SIGs), the main form of discretionary finance provided to schools by the Government, includes a base component, top-up finance on a per-student basis for larger schools, and a needs-based component provided to schools with severe shortages of infrastructure, teachers, or basic amenities. However, the persistence of large disparities in school conditions, even within a single zone, suggest that the needs- based component may need to be increased (see Box 7). Delays in delivery of the SIGs are commonplace: of the six primary schools visited for this study in January 2020, none had yet received their SIG allocation for 2019/20 despite expecting transfers/payments in September 2019. Two of the schools had received 43 In early grades, stunted growth in early childhood may undermine the proper development of the child and may affect his/her readiness for school. Such issues can be addressed by the provision of early childhood development programs in combination with other childhood development activities, including nutrition programs. no SIGs at all in 2018/19, without explanation from their district authorities. Schools in remote areas, which typically face the most severe shortages of materials and problems in terms of infrastructure conditions, are more likely to experience delays in SIG finance. In 2014/15, remote schools received SIG transfers 21 days later than the average school, while schools close to a trading center received their SIG on average 14 days earlier than the average school, according to MLSS data. 44 42. Per-student allocations at secondary level are significantly inequitable across education management divisions, the six sub-regions which act as the primary middle management layer for secondary education (Figure 30). Overall allocations per student vary significantly across regions, ranging from a high of MWK 106,061 (US$ 145) for the Central Western division to a low of MWK 37,896 (US$ 51) for the Central Eastern division against a national average of MWK 75,209 (US$ 102). The same applies to subventions to schools, other recurrent transactions and personal emoluments per student. The rationale for the variations is not known but the type of school most students attend within the respective divisions could be one of the sources of variation. 45 Figure 30: Overall public allocations per secondary education student Kwacha per student 120,000 106,062 100,000 87,724 80,000 74,475 59,500 60,000 37,896 40,000 33,148 30,323 30,564 30,282 18,338 16,574 16,574 20,000 14,975 15,282 15,282 15,141 15,141 7,573 0 NED CWED CEED SWED SEED SHED PE per Learner ORT per Learner Unit Cost per Learner Source: EMIS data 2017 44 Given the high rate of execution of Grant budgets, these delays and non-payments suggest inadequate allocations to SIG. 45 Secondary finance is also subject to inequities in reliability and quantity of funding between schools which constitute cost centers and those who do not. Schools which are cost centers receive finance direct from the Treasury, while non-cost centers are funded via educational divisional managements. During field visits for this study it was found that respondents from cost centers are required to comply with a range of MoF requirements such as multiple quotations for goods and services to be procured, obtaining local purchasing orders from division offices, and paying withholding taxes; as a result of these challenges respondents report that only around Institutional arrangements and efficiency of service delivery Transparency of policy, budget, and human resources and accountability framework 43. The transparency of policies and budgets for education at the national level is adequate but with considerable potential for improvement. The framework of policies governing education at national level is fairly well established with key governing documents and sector plans in place. A moderate level of data is available on national budgetary allocations by sector and sub-sector, including on direct subventions to schools and Education Division Managers at secondary level 46 and to higher education institutions. However, the high level of education expenditure by donor partners which remains off-budget (see below) has negative impacts on the transparency of sector spending as a whole. Policies on teacher recruitment, promotion and management have improved, but new teacher hires and deployment processes remain somewhat opaque and prone to unpredictability in terms of the numbers of teachers deployed and criteria for allocations to districts. 44. Budgetary and human resources frameworks at the district level are not transparent despite their growing importance in the context of decentralization. Following decentralization reforms, the majority of education expenditure at primary level, including teacher salaries, is managed by local Government authorities, primarily through direct allocations for education personnel emoluments and other recurrent transactions from the central Government. The formulae governing these transfers are complex (see pg 237), with 12 different variables contributing to calculations on annual allocations per district, making it difficult for LGAs to predict their expected allocations. Although district allocations are typically received by districts in full on an annual basis, there is significant unpredictability at the monthly level for other recurrent transactions in particular which negatively affects planning of development activities.47 District-level expenditures are expected to be tracked using the Government of Malawi’s Integrated Financial Management Information System (IFMIS), with support to the process from the National Local Government Finance Committee (NLGFC). However, the recording of district fund receipts and expenditure in IFMIS is currently incomplete and insufficiently robust, without consistent distinctions between teacher salaries and other forms of personnel emoluments, and between types of other recurrent transactions, preventing robust analysis of per-district expenditures. Assessment of policies on recruitment, deployment, qualification, and retirement 45. Education service delivery in Malawi is governed by the 2016 National Education Policy and implemented according to the National Education Sector Plan (NESP 2009-2017) and accompanying Education Sector Implementation Plan (ESIP-II, 2012-17). A new National Education Sector Investment Plan (NESIP) 2020-2030, a replacement for NESP, is currently being finalized. Education is also implemented in accordance with sectoral Strategy documents including the National Girls’ Education Strategy 2018-23 and Adolescent Girls and Young Women’s (AGYW) Action Plan for Malawi. The new NESIP will be accompanied by a new ESIP, 2020-2025; the implementation of this will be supported by an Education Sector Plan Implementation Grant from GPE, which is currently being developed. 46. Policies on teacher management at the primary level have become more comprehensive and harmonized, but further enhancement is required. Policies regarding teacher management have historically been prone to fragmentation, inconsistency, and ad hoc implementation, resulting in severe inefficiencies including: inequitable distribution of teachers between schools and, within schools, 46 Secondary education is managed by six sub-regional Education Divisions. 47 World Bank. 2020. Malawi Subnational Public Expenditure Review. Mimeo. between grades; a lack of promotion opportunities, with negative effects on motivation and effort; and limited enforcement of policies relating to conduct and discipline. In particular, a lack of clear guidance on the allocation of teachers to schools, and clear incentives for enforcement of rules on teacher placements and transfers, has led to extremely poor distribution of teachers between schools (for more, see pg. 57). 47. At secondary level, a new strategy document governing teacher development, recruitment, deployment and management is currently under preparation. Its preparation is supported by the World Bank-supported Equity with Quality and Learning at Secondary (EQUALS) project. The strategy is expected to explore a range of potential reforms enabling possible enhancements to teacher utilization as described in the previous section, as well as supporting the establishment of a Secondary Teacher Management Information System.48 Within the TVET and higher education system, efforts to improve policies and institutional capacity are ongoing.49 Efficiency of service delivery 48. The median Malawian primary school has an almost adequate number of teachers, but they are poorly distributed between grades. According to the MLSS, the median Malawian school has 10 teachers, of which 3 are female. The median pupil-teacher ratio (PTR) is 65, just above the Government’s target of 60. However, this adequate median masks wide variation as discussed above; furthermore, these teachers are poorly distributed between grades, with severe shortages in lower grades and overstaffing in upper grades.50 Female teachers in particular, are unlikely to be assigned to upper grades. Of the teachers who teach Grade 1 regularly, 55 percent are female, versus just 10 percent in Grade 8. This represents a severe lack of role models for female students in upper grades, which may contribute to high rates of dropout. 49. A minority of teachers remain in the workforce without having completed secondary education. A typical Malawian primary school teacher is 33 years old, with 14 years of education. Of the MLSS sample of teachers, 51 percent are male. The vast majority of teachers – 92 percent – have at least the Malawi high school completion qualification, the MSCE; fewer than 1 percent have a university diploma or higher qualification. Despite their age, the median Malawi teacher has been teaching for only 6 years, but at their current school for just 3 years. Long-running problems with the lack of promotions is reflected by the fact that 69 percent of teachers are at position PT4, the most junior grade (Asim et al., 2019). The average teacher salary is around MWK 100,000 per month, more than double the national median average income.51 However, 88 percent of teachers in the MLSS sample have unpaid or delayed salary claims outstanding, reflecting widespread problems with timely payment of salaries and benefits. 48 Designed to enhance the quality and comprehensiveness of system data on teacher placement and performance 49 A regulatory body, the National Council on Higher Education, has been established and operationalized. Several system-level capacities have been enhanced, including the development of the higher education management information system (HEMIS), quality assurance system (QAS), higher education qualifications system, and student financing management information system. However, some new systems such as HEMIS are not yet operational. Much progress has been made in the TVET subsector with a harmonized assessment system and curriculum developed, but key challenges remain. There were three different skills training certificates in the country, often creating confusion for students and industry in terms of certification for TVET skills. Although currently there are some difficulties associated with the final operationalization of the new curriculum, assessment, and related certifications, much progress has been made in this area. 50 Lower grades have substantially higher enrollments, with the median school having 163 students in Grade 1, and only 47 in Grade 8. However, the distribution of teachers and classrooms does not reflect these varying enrollments: the typical school has fewer than two teachers for each of these grades, while also dedicating more than one teacher to Grade 8, resulting in far higher PTRs in lower grades (average PTR of 108 in Grade 1, compared with 28 in Grade 8 (EMIS, 2018)). 51 National median income was MWK 45,000 in 2016/17 (IHS 2016/17) 50. In general, primary teachers have a barely adequate level of subject knowledge. The MLSS data collection included asking teachers to complete the same assessment exercise given to Standard 4 students, including a range of questions aligned with the curricula for Standards 1, 2, 3 and 4. The median teacher achieved a score of 96 percent in English and 93 percent in math and Chichewa, showing satisfactory though imperfect knowledge of the concepts covered in the lower primary school curricula. However, a minority of teachers scored significantly below the median: the bottom 10 percent of performers scored less than 89 percent in English, 82 percent in math, and 81 percent in Chichewa. 51. Secondary schools have an inadequate number of properly qualified teachers . The overall PTR at secondary level is low at 20.4 students per teacher. However, this overall level of staffing masks the high percentage of unqualified teachers in the system. Of the 14,333 secondary teachers, 43.7 percent lack the requisite skills to instruct at the secondary education level, while 4,000 are primary teachers. The use of primary school teachers in secondary schools has been commonplace in recent years as a means to support expansion of secondary schooling. While the use of primary teachers in the secondary sub-system relieves the system of the huge salary burden of paying secondary teachers, it is a highly inefficient strategy with negative implications for student learning. An upgrading program, under which primary school teachers deployed at the secondary level simultaneously undergo a four-year training at a secondary education training institution, enrolls only 20 trainees per year. Substantial improvements in the flow of qualified secondary teachers in the system are required to catch up to and keep pace with the expansion of the system. 52. Low overall PTRs at secondary level also mask severe shortages of teachers for the Sciences, Math and Agriculture. At the national level, PTRs in these subjects are extremely high: 247:1 in Math, 370:1 in Biology, 527:1 in Physics and a very high 1.720:1 in Agriculture. Evidence from the three CDSS visited for this study shows that a substantial number of the teachers were not specialists in the subjects they were teaching at the secondary education level. 52 Resourcing schools with qualified teachers in the subjects being taught at the secondary education level is a strategic need, coupled with equipping in-service teachers with the right content knowledge for the subjects they teach. 53. Teachers in Malawi are frequently absent, and when in school, spend inadequate time on teaching. On the day of data collection, 13 percent of teachers were absent at the primary schools sampled for the MLSS. In terms of teaching time, teachers’ own estimations suggest the typical Malawian teacher spends just 2.5 hours per day, less than half the school day, teaching their own class. An additional 1.5 hours is spent preparing lessons. Instead of teaching, teachers spend school time carrying out other duties.53 A minority of teachers appear to spend almost no time teaching: while the top 10 percent of teachers said they spend more than 3.5 hours teaching each day, the bottom 10 percent said they spent less than 45 minutes. The MLSS classroom observations were carried out during scheduled lesson time, but in 58 percent of observed classrooms, the teacher was not present when enumerators entered, and in 12 percent of lessons the teacher never arrived and the lesson never took place. 54. At secondary level, teaching loads are low, exacerbating shortages of teachers in key subjects. Information from schools54 indicates that science and mathematics teachers have the highest teaching load. However, their load is estimated at 17 periods a week – each period is 45 minutes – which translate into about 13 hours a week. Global norms stipulate an average of 40 hours per week and 28 52 The percentage of teachers specialized in the subjects they were teaching was as low as 37 percent and 46 percent in two of the three schools; while all the three schools had no teacher specialized in agriculture (Annex Table 1A). 53 1.4 hours per day is spent marking homework, 1.2 hours on administrative duties, and 1.3 hours helping other teachers. (MLSS 2016) 54 Visits to schools for preparation of EQUALS project, 2018/19. periods per week (21 hours) for public servants and secondary education teachers respectively. The assumption is that the remaining hours would be devoted to lessons/experiments preparation, marking and professional development. The Government should set a minimum teaching load for secondary education teachers guided by the global standards for better use of the teacher stock and enhanced efficiency of the wage bill provisions. This should be further underpinned by the secondary education pre-service teacher training norms where the global norm is that a teacher is trained to teach in two (one major) subjects. 55. When present in the classroom, Malawi’s primary school teachers exhibit a range of positive teaching practices, but rarely assign and mark homework. Comparison of MLSS lesson observations with those from other countries using the Service Delivery Indicators, suggests that Malawi’s primary school teachers, when teaching, are as likely as those in comparable countries to exhibit a range of positive behaviors associated with improved learning such as correcting mistakes and giving positive reinforcement, and Malawian teachers are among the most likely to ask questions of students. However, only 10 percent of teachers in observed lessons set homework for students, and only 5 percent collected or reviewed homework previously assigned. The findings suggest that teachers, while performing adequately when addressing their entire classes, avoid tasks which require completion on a per-student basis such as marking homework or providing support after class; this likely reflects the preponderance of large class sizes. 56. Provision of infrastructure in primary schools has struggled to keep pace with the rapid expansion of the system and the typical Malawian primary school lacks adequate infrastructure, facing maintenance challenges. According to the MLSS, the median primary school has 744 students and seven classrooms, a pupil-classroom ratio of 95, not including open-air classrooms. At the median school, around one-third of classrooms (29 percent) typically have walls that are cracked or damaged, and 6 percent have walls that are substantially broken. The typical primary school also lacks certain key teaching and learning materials. At a typical school all classrooms, including open-air classrooms, have a blackboard and chalk, according to the MLSS. However, only 1 in 7 classrooms (16 percent) at a typical school have a corner devoted to learning materials. Many of Malawi’s primary school students lack chairs and desks: at a typical school, only 41 percent of students sit on a chair for lessons, the remainder sitting on the floor.55 However, six out of seven classrooms at a typical school have a chair for the teacher. Equity of service delivery 57. Primary school staffing levels are subject to a wide degree of variation which represents a serious misallocation of resources. Nationally, the 10 percent of schools with the best staff levels have a PTR below 39 – well below the national target – and 10 percent have a PTR above 100. In the case of the six public primary schools visited as part of the preparation of this report, PTRs ranged from as low as 23:1 to as high as 122:1. Across Malawi, this variation occurs on a micro-geographic scale, with PTRs varying by an average of five times within a single sub-district neighborhood. These disparities reflect the weaknesses of the system for the allocation of teachers to schools, in which poor quality data leads to overly broad guidance to local officials and teachers who in turn exploit this discretion to ensure appointments in settings with low hardship (Asim et al., 2019). However, these disparities are also mirrored in the health sector, suggesting broader weaknesses in personnel management at district level (pg. 59). The three district education managers interviewed for this study each used different criteria for the allocation of teachers to schools and had varying methods for identifying the neediest schools, despite efforts by the MoEST in recent years to standardize approaches. 55 Malawi generally employs combined chair/desk units, so students sitting on a chair are typically also able to use a desk. 58. The quality of school environments, and availability of teaching and learning materials, varies widely. The top 10 percent of schools have fewer than 37 pupils per classroom, and the bottom one-tenth more than 156 pupils per classroom56. In the six primary schools visited for this study, pupil- classroom ratios ranged from 42:1 to 314:1. The condition of classrooms also varies widely: in the top 10 percent of schools with the best overall infrastructure quality, the majority of classrooms have permanent walls and roof, and all physical infrastructure is in fine condition. 57 By contrast, in the bottom 10 percent of schools, 58 percent of classrooms have walls that are broken, leaking or cracked. These inequalities extend to teaching and learning materials.58 59. These inequities in staffing and learning environment directly and negatively affect learning outcomes. Evidence from the MLSS baseline study demonstrates that, controlling for a wide range of school characteristics, schools in remote areas achieve significantly lower learning outcomes in Grade 4; as do those with lower than average levels of desks, notebooks and uniforms or a greater reliance on junior teachers. In order to achieve equitable levels of learning, there is an urgent need to increase the targeting of inputs to schools to address these inequities. 60. At secondary level, shortages of qualified teachers are more severe in CDSS than in the Government boarding and day schools. The overall PTR in CDSS is not much higher than in Government boarding and day schools (Figure 31). However, CDSS are most dependent on unqualified teachers, with the result that the pupil-qualified teacher ratio (PqTR) in CDSS is almost double that of the Government boarding and day schools. Figure 31: Deployment of secondary education teachers Pupil teacher ratio Source: EMIS MLSS baseline, 2016 57 The Malawi Longitudinal Schools Survey defines an infrastructure quality index (IQI) to enable school-level analysis of variation in physical classroom infrastructure. The index assigns scores based on the condition of walls, roofs, floors and doors. Here we report conditions in the top and bottom decile schools by IQI. 58 The MLSS defines a classroom facilities index (CFI) which encompasses key indicators of availability of chairs, a blackboard, chalk, and learning materials. Here we present average findings for the top and bottom deciles of schools by CFI. In the best-resourced 10 percent of schools in terms of classroom facilities all classrooms have chalk, a blackboard, a corner dedicated to learning materials, and chairs for the majority of students. By contrast, in the least-resourced 10 percent of schools, only 35 percent have chalk, only 4 percent have a corner for learning materials, and only 11 percent of classrooms have chairs for the majority of students. Conclusions and recommendations 61. Access to education remains low, with more than 1 in 10 children out of school, and disparities across gender and location. This is driven by low Net Enrollment Levels, which worsen for post primary education. In addition, completion rates are low with 44 percent and 23 percent of learners completing primary and secondary schooling, respectively. High dropout rates are mostly attributed to late enrolment as only 30 percent of the students start grade 1 at the age of 6 with the rest starting at 7 or 8. 62. Literacy rates have improved, particularly among the youth. Despite high labor market returns to an additional year of schooling, the average working-age adult only has about 6.5 years of schooling. 63. A large portion of education spending is funded by the Government (46 percent) followed by households (39 percent) and development partners (DPs) (15 percent). Within the public component of education finance, there are large discrepancies between allocated and executed budget, which undermine the efficiency of public spending in the sector. This is particularly the case for capital expenditure. 64. Sustainability: Severe pressure is likely on primary finance in the coming years as a result of rapid population growth. Public primary enrollment is projected to rise by around 10 percent by 2024/25. Accommodating this expansion while improving service standards to meet policy commitments is likely to require significant increases in primary expenditure, in particular infrastructure costs. 65. Efficiency: Despite low Pupil-Teacher Ratios (PTRs), the misallocation of teachers is a critical bottleneck to improving education outcomes. Teachers are poorly distributed across grades in primary schools and the low overall PTR for secondary schools masks the shortage of teachers for science subjects. The shortages of teachers is more rampant in CDSSs compared to Government boarding or day secondary schools. 66. Equity: Overall public expenditure on education largely supports wealthier quintiles. Currently, the poorest quintile receive only 15 percent of total education spending with the wealthy receiving 48 percent. Expenditure patterns, however, are more pro-poor at primary school levels. 67. Transparency: At district level, transparency in budget and human resource frameworks remains wanting compared to the national level. This is attributed to complicated formulae governing transfers that make it hard to track budgetary allocations. This is compounded by the large amount of education spending by DPs that remains off-budget. In addition, School Income Grants are characterized by delays and this is more predominant in rural areas. Policy priorities and recommendations 68. The following priority actions are recommended to improve equity, efficiency, and transparency of education spending. 69. Equity: Low-cost, safe and sustainable classrooms are needed to provide equitable access to quality education at primary level on a fiscally sustainable basis . Shortages of classrooms have grown more severe in recent years despite investment from MoEST budgets and donor partner support, as a result of (a) incomplete execution of budgeted expenditure; and (b) unsustainably high unit costs. The current average unit cost of around US$ 20,000 per classroom raises the cost of achieving pupil- classroom ratios of 1:60 to almost US$ 1 billion over the next five years at primary level alone. Evidence from pilot implementation of additional discretionary finance for schools suggests that low-cost learning shelters, procured and constructed with community management, can be obtained for around US$ 3,000, with substantial positive implications for fiscal sustainability. These shelters provide immediate reduction in class sizes, particularly for overcrowded, disadvantaged schools, reducing prevalence of open-air classrooms. The Government is currently developing an updated standardized design for low-cost classrooms, expected to provide a higher degree of durability and safety than learning shelters while still maintaining costs below US$ 5,000 per classroom which will significantly reduce the large gap in pupil classroom ratios between Government primary schools. 70. Equity: Primary school grant finance needs reform to respond better to school needs. Under the current SIG formula, only around 14 percent of the total budget for SIGs is allocated on the basis of school need. This small needs-based component has limited impact on the amount of SIG received – for example, a school with a PCR of above 120 (representing a severe shortage of classrooms) received only US$ 305 more than a school with a PCR below 60 (representing no shortage) in 2017/18, less than one- tenth the cost of constructing even a low-cost learning shelter using community construction. 59 The persistence of chronic inequities in school conditions, even within a single subdistrict, suggests that there is an urgent need to increase the needs-based allocation. Analysis by the World Bank suggests that the allocation of comparatively small amounts of finance – US$90 million over four years – targeted to the schools with the most severe shortages of classrooms and teachers in lower grades, and used for construction of low-cost classrooms and hiring of temporary contract teachers, could ensure that no primary school continues to have PTRs or PCRs above 90:1 in lower primary. 71. Efficiency: Action is also needed to address delays in the payment of SIG. As the primary form of discretionary finance provided to schools by the Government, SIG is essential for the operation of schooling, yet SIG payments are typically substantially delayed, particularly for schools in remote areas. Reform is required to allocation and payment procedures for SIG to ensure more timely delivery to schools. Evidence from Tanzania suggests that transitioning school grants away from decentralized distribution, and adopting direct release of grants from central Government to schools, can immediately and permanently address chronic delays and inaccuracies in fund release. 60 A similar approach is likely to be the most efficient approach for Malawi to address SIG delays. 72. Efficiency: Misallocation of teachers to primary schools remains the single largest source of inefficiency in the Malawian education system, and requires further reform. With teacher salaries accounting for more than 70 percent of Malawi’s basic education budget, the current poor distribution of teachers – both between schools, and within schools, between grades – represents the single largest source of inefficiency in the system. Allocations of teachers to new schools have improved in recent years following the development of new guidelines and tools, and in 2018, 50 percent of new deployments were to schools with PTRs above 100, representing a significant shift toward appropriate allocations. As such, there is a need to further strengthen policies and procedures regarding teacher deployment and redeployment to improve the allocation of teachers to primary schools. This will help ensure enforcement of allocation decisions, and to empower officials at the district level to resist pressure to change deployments. A new primary Teacher Management Strategy, approved in 2018, provides increased clarity for policies on teacher recruitment, deployment, promotion, and conduct. However, further reform is required to better harmonize this document with other guidance relating to teacher deployments and to reflect changes in responsibility for decision-making as a result of decentralization; and further technical work is required to put the strategy into operation through 59 MoEST, 2018. 60 On-time payment of grants to schools rose from 74% to 100% following the reform in January 2016 (World Bank Group, 2017) preparation and dissemination of tools for evidence-based performance management and promotion, and the introduction of a revised hardship allowance for teachers in remote postings. 73. Efficiency: Improved teacher utilization and more equitable supply of qualified teachers for all subjects are required in order to raise efficiency at secondary level and to ensure their adequacy for the likely rapid expansion of the system. Secondary teaching loads are estimated at an average 13 hours per week or less, well below global norms. Increased double-shift instruction can reduce the need for new teachers and infrastructure. The fact that schools can run open secondary schools after the regular school day implies that double shifts are feasible for a substantial number of schools. Addressing the teaching load for secondary education teachers will also solve some of the teacher deployment issues for schools offering double shifts. For example, teachers deployed to double- shift schools could: (a) be paid teaching allowances if their total periods in the two shifts for the respective schools exceed the set average minimum teacher periods for secondary teachers; and (b) work across shifts to realize the set minimum periods per secondary education teacher. Intensive dialogue and consensus building is required with all stakeholders on teacher issues, including the Teachers Union of Malawi and the Teaching Service Commission, to develop an efficient and sustainable approach to secondary staffing. 74. Efficiency: Government should explore lower cost models for delivery of secondary education. The ongoing revitalization of the open secondary schools system has the potential to increase the overall efficiency of secondary expenditure. The lower cost of open secondary schools, if conducted and managed as designed, would reduce the costs to parents for secondary education and thereby increase access. This process will however require re-examination of how the current education system and structures interface with the service delivery design of instruction in open secondary schools. The delivery of an online secondary education program currently under development should exploit synergies between sectors for efficient use by the students, including school-based ICT laboratories to be constructed by donor-supported projects, and existing regional open distance learning centers for public universities. 75. Efficiency: There is a need to optimize existing placements to increase the efficiency of the secondary sector. This involves timely replacements of all students that do not turn up at the schools for which they have been selected. It calls for flexibility in admissions management at the divisional levels as well as in empowering school principals to take decisions related to student admissions. This applies in situations where some of the students on the official list for schools do not enroll within an agreed number of days, while students within the school community who qualified to enroll were not officially placed. Instead of schools continuing with empty spaces and waiting for the official replacement of such students, schools could be allowed to enroll students from their respective waiting lists based on agreed admissions criteria and inform the divisional management afterwards. 76. Transparency: Urgent action is needed to raise the quality and availability of data on LGA- level education expenditures. As noted above, despite local Government authorities now controlling the majority of primary education expenditure, there is a lack of collated and transparent data on per- local authority spending. In order for the general public to properly assess education service delivery by local authorities there is a need for greater transparency. Furthermore, evidence from the MLSS, field visits, and other sources confirms the existence of substantial inequity in resources between schools within a single district; greater transparency is therefore required at the level of per-school allocations of teachers, infrastructure investment, and other resources. The forthcoming NESIP 2020-30, which is expected to introduce policy frameworks for targeted investment in the most disadvantaged schools, may present a suitable opportunity to embed transparency into the system on the level of facilities. Health Sector Review Introduction 77. This chapter presents an analysis of health sector expenditure between 2014/15 and 2018/19. While the initial parts of the analysis consider total current health expenditure from all financing sources, the latter parts of the review focus on public spending in the health sector, specifically for resources that are channeled through the Government system. The review covers the volume and composition of public spending in the health sector, efficiency in spending, and equity in financing. An assessment of how the existing PFM system is supporting health service delivery is also provided. In particular, the PFM assessment highlights the viability of the PFM system in facilitating service delivery at district hospitals and health centers. A preliminary analysis of the potential impact of the COVID-19 pandemic on service provision and financing of health services in Malawi is also provided. 78. The data for the analysis was mainly from Government sources. Budget and expenditure data covering the fiscal years 2014/15–2018/19 were drawn from annual Government financial reports at both national and district levels. These reports are generated directly from Malawi’s IFMIS. Apart from Government financial reports, additional information was obtained from a number of sources. For the analysis on total expenditures in the health sector, data were obtained from national health accounts (NHA) surveys while data on health inputs (for instance human resources, drugs and medical supplies) were collected from various departments at the Ministry of Health of Malawi (MoH). Data on demographics, performance of the health system, and health outcomes were collected from the MoH’s Health Management Information System (HMIS), demographic and health surveys, household surveys, the Institute for Health Metrics and Evaluation, and the World Bank. Macroeconomic and fiscal data were obtained from the World Bank and the International Monetary Fund. Health Sector Performance 79. Over the past two decades, Malawi has made remarkable progress in improving maternal and child health outcomes. It not only more than halved the maternal, infant, neonatal and under-5 mortality rates, but also surpassed the averages for regional and peer countries (Table 3). However, Malawi is still lagging behind in certain health outcomes, including stunting, HIV, and malaria prevalence, which contribute to the low human capital index (HCI) of 0.41 (Figure 32). This means that a child born in Malawi today will be 41 percent as productive when she grows up as she could be if she enjoyed complete education and full health. Out of the five indicators which are aggregated in the HCI index, the high level of stunting is one of the main underlying reasons for the low HCI in Malawi. Approximately 40 percent of Malawi’s children below the age of five are stunted, 3 percent are wasted, and 12 percent are underweight. The adult survival rate in Malawi is also worse than the averages for regional and peer countries. Only 73 percent of the children aged 15 in Malawi today will survive until the age of 60. One of the underlying factors for poor maternal, child health and nutrition outcomes in Malawi is the high Adolescent fertility rate which was estimated at 132 births per 1,000 women aged 15-19 in 2018 which is higher than the averages for regional and peer countries. Table 3: Key Health Outcome Indicators Malawi Avg Avg Indicators 2000 2010 2016 2018 LIC SSA Life expectancy at birth, total (years) 45.1 55.6 62.7 63.8 63.5 61.3 Adolescent fertility rate (births per 1,000 women ages 15-19) 158 148 135 132 94 101 Maternal mortality ratio (per 100,000 live births) 749 444 358 349 462 534 Mortality rate, infant (per 1,000 live births) 99.8 53.2 35.3 32.1 49.2 53.0 Mortality rate, neonatal (per 1,000 live births) 38.7 28.2 21.9 20.4 27.2 28.0 Mortality rate, under-5 (per 1,000 live births) 172.6 84.9 50.0 43.9 69.9 78.1 Prevalence of stunting, height for age (% of children under 5) 54.7 47.3 38.3 39.0 34.8 33.5 Prevalence of HIV, total (% of population ages 15-49) 14.4 10.6 9.7 9.2 2.0 3.8 Prevalence of anemia among children (% of children under 5) 74.4 64.8 59.2 59.2 59.9 Incidence of tuberculosis (per 100,000 people) 386 310 193 181 206 231 Incidence of malaria (per 1,000 population at risk) 427 386 211 214 191 219 Source: Authors’ calculations from the World Development Indicators (WDI) dataset LIC = Low-income country; SSA = Sub- Saharan Africa Figure 32: Human Capital Index Source: Authors’ calculations from WDI data 80. Progress on health outcomes over the past two decades could be reversed by the COVID- 19 pandemic. As at October 25, 2020, there were 5,887 confirmed cases of COVID-19 in Malawi with 183 deaths.61 These figures are significantly lower than earlier projected,62 and this suggests that the COVID- 19 pandemic is still within manageable levels in Malawi. But despite the low number of COVID-19 cases, utilization of essential health services has gone down since April 2, 2020 when the first case of COVID-19 was registered in Malawi. A detailed review of health facility data from Malawi’s HMIS shows that utilization of some of the reproductive, maternal and child health services has reduced since the advent of the COVID-19 in the country.63 The number of consultations with new users of modern family planning 61 https://covid19.who.int/region/afro/country/mw 62 https://www.imperial.ac.uk/media/imperial-college/medicine/mrc-gida/2020-03-26-COVID19-Report-12.pdf 63 World Bank Development Research Group (2020). “Monitoring disruptions to essential health services in Malawi” methods declined by 26 percent on average between April-May 2019 and April-May 2020. Over the same period, the number of pregnant women who attended the first antenatal care visit declined by 15 percent while the number of institutional deliveries reduced by 4 percent. Further, the number of children who were given the third dose of Pentavalent vaccine reduced by 10 percent during the same period. These findings are corroborated by a study by the Global Financing Facility, 64 which shows that the COVID-19 pandemic is likely to disrupt the supply and demand of maternal and child health services in Malawi. The reduction in utilization could be attributed to disruptions in the procurement and distribution of medicines and other essential medical commodities, greater emphasis on COVID-19 leading to reduced hours for essential services, suspension of outreach services (e.g. antenatal and under-5 clinics), and re- allocation of resources toward COVID-related activities. On the other hand, demand for health services may have been affected by patients’ fears of contracting the disease if they go into the health facilities, and reduced ability to pay for medical costs at private and other fee-charging facilities. However, these findings should be interpreted with caution as other factors could have also affected utilization. But whatever the case, maintaining and/or scaling-up the provision of essential health services during the COVID-19 outbreak is critical to reducing morbidity and mortality due to the COVID-19 itself and from other diseases. 81. The COVID-19 pandemic is likely to affect the financing and provision of health services in Malawi whether the number of cases remain low or increase. Foremost, though the number of cases and deaths as at October 25, 2020 are low, the disease is far from over as no cure or vaccine has been discovered. This means that there is need for additional investments in the health system because the existing human, material, and financial resources in the health sector in Malawi are insufficient to effectively respond to the COVID-19 virus. Malawi is ranked at 154 out of 195 countries on the 2019 Global Health Security Assessment (GHSA), 65 and this suggests that its ability to prevent, detect, and respond to epidemics and/or pandemics is limited. Evidence from Europe and Asia highlight the need to adequately invest in the health system to manage a rise in COVID-19 patients. In the event that there is a surge in the number of COVID-19 cases requiring outpatient and inpatient services in Malawi, health authorities will have to deal with an increased disease burden due to the COVID-19 as well as other communicable diseases such as HIV/AIDS, malaria, and tuberculosis, which are already high. This will require additional domestic and external funding over and above what is currently being provided from all sources. But even if the number of cases remains low, the disease will most likely have a negative effect on economic growth due to reduced trade with other countries. Consequently, GDP growth in Malawi is projected at 2.0 percent in 2020, which would be the lowest since 2012. And given the already high fiscal deficit, expenditure on health and other social sectors is likely to shrink. 64 Global Financing Facility (2020). Preserve Essential Health Services During the COVID-19 Pandemic: Malawi. Washington DC, The World Bank. https://www.globalfinancingfacility.org/sites/gff_new/files/documents/Malawi-Covid-Brief_GFF.pdf 65 Worldwide, the average score on the GHSA index is 40.2% while the average score for the African region is 30.8%. Malawi scored 28.0 in 2019. https://www.ghsindex.org/wp-content/uploads/2019/10/2019-Global-Health-Security-Index.pdf Box 4: High Levels of Fertility and Stunting are Impeding Progress on Human Capital Investments in people can help build the human capital that is key to ending poverty and ensuring shared prosperity. A healthy, well-nourished, well-educated, and skilled labor force enables economies to grow faster and compete in the dynamic, technology-led global economy. While Malawi has made good progress on many aspects of human capital, including reduction of fertility and childhood stunting over the past two decades, there is need to deepen and accelerate efforts. This is because Malawi has a low Human Capital Index (HCI) of 0.41 and is ranked 125 out of 157 countries. High levels of childhood stunting (estimated at 39 percent of the children below the age of five), and a low adult survival rate (0.74) are the main reasons for the low HCI in Malawi. High levels of stunting are a serious impediment to human capital development. It is estimated that poor childhood nutrition outcomes in Malawi contribute to losses of US$597 million (2012 terms) or US$39 per capita per year. Therefore, undernutrition among the children and adolescents is a critical development challenge in Malawi. The underlying causes of childhood stunting in Malawi are associated with factors that children are exposed to at individual, household, and community levels. Consequently, only 74 percent of the children aged 15 in Malawi today will survive until the age of 60. Further, at 349 deaths per 100,000 live births, Malawi still has a very high maternal mortality ratio. Poor maternal, child health and nutrition outcomes in Malawi could be attributed to the high adolescent fertility rate estimated at 132 births per 1,000 women aged 15-19 in 2018 which is higher than the average for low income countries. Nearly one in every three adolescent girls aged between 15 and 19 years in Malawi has already began childbearing. Adolescent pregnancies contribute about 29 percent of all deliveries in Malawi, and given that most of them are undernourished and their bodies are not fully developed, they are predisposed to delivering low birth weight babies. This increases the number of neonatal deaths, which accounts for 47 percent of the total number of under-five deaths in Malawi. Further, about 15 percent of all the maternal deaths in Malawi are from adolescent girls. Figure 33: Total Current Health Expenditure by Disease: 2017/18 HIV/AIDS and STIs 37% Reproductive Health 17% Malaria 13% Nutritional deficiencies 9% Non-communicable diseases 7% Vaccine preventable diseases 5% Injuries 5% Tuberculosis 2% Diarrhoeal diseases 1% 0% 10% 20% 30% 40% At the health system level, there are a number of challenges that hinder the provision of quality health and nutrition services. As revealed in the Harmonized Health Facility Assessment (World Bank et al. 2019), there is limited health worker capacity, inadequate medical supplies and equipment, and weaknesses in governance and management. Results from this PER also highlight gaps in the PFM arrangements at Government hospitals and health centers; and that allocation of the available financial and human resources is not aligned to the existing disease burden or health outcomes at district level. For instance, even though there are significant disparities in the prevalence of stunting by district, there is higher expenditure in districts with lower levels of stunting (see figure 44a). Breakdown of the total current health expenditure by disease also shows that nutritional deficiencies (underweight, wasted, stunted) are not adequately prioritized despite nutrition being key to human development (see figure 33). The patterns of health spending in Malawi also suggests that there is an association between high fertility, education attainment, poverty, undernutrition, and health status. In particular, high fertility is a risk factor for HIV/AIDS, poor maternal health outcomes, and undernutrition; and this perpetuates poverty. Malawian women from the poorest communities and those with low education have the highest levels of fertility and undernutrition which limits their economic prospects and traps them deeply in poverty across generations. In lieu of the above, it is evident that Malawi cannot end extreme poverty and ensure shared prosperity without effectively tackling the high levels of fertility and stunting. Prospects for economic growth and poverty reduction in Malawi could be substantially enhanced by investments in the key determinants of health such as agriculture, nutrition, health, education, water and sanitation, social protection, good governance and so forth. Furthermore, given the challenges posed by the COVID-19, it will be critical for Malawi to ensure that provision of essential health and nutrition services is sustained especially for the poor and vulnerable populations. Health Expenditure Trends and Funding Sources Level of Total Health Expenditure 82. Spending on health in Malawi averaged US$645 million per annum over the period 2009/10 and 2017/18 (Figure 34). The level or volume of spending grew consistently between 2009/10 and 2012/13, but declined in 2013/14, due to the Cashgate scandal, 66 after which it has increased in absolute terms but has never reached the 2012/13 level. Expenditure peaked in 2012/13—the year before the Cashgate scandal—both in US$ per capita, and as a share of GDP. After the Cashgate scandal, total per capita health spending has largely been between US$39 and US$40, while as a share of GDP, health spending has been declining since 2012/13. In comparison to other low-income countries, Malawi’s total Current Health Expenditure (CHE) as a share of GDP and total CHE per capita are higher, largely due to high donor funding. Malawi’s total CHE as a share of GDP was about 10 percent per annum on average over the period 2009/10 and 2017/18 which is almost twice the 5.2 percent average for low-income countries. In per capita terms, total CHE per capita estimated at US$39 per year over the same period is also relatively higher than the average of US$34 for low-income countries. 83. Though per capita health spending in Malawi is relatively higher than other low-income countries, it is below the estimated need. Firstly, the US$39 per capita is lower than the funding that is required for Malawi to provide essential healthcare as outlined in the national Essential Health Package (EHP).67 Consequently, only 44 percent of the health facilities in the country are able to comprehensively deliver health services outlined in the EHP; and this is significantly below the 80 percent target outlined in the Health Sector Strategic Plan (HSSP II) 2017-2022. 68 Secondly, for low-income countries like Malawi, estimates by the World Bank through the third edition of the Disease Control Priorities 69 shows that 80 percent coverage for a package of essential universal healthcare requires US$72 per capita per annum. This suggests that more funding to the health sector is required. However, given that the share of GDP committed to the health sector is already high, there is need to explore efficiency enhancing interventions within the existing resources as getting additional (new) financing will be difficult. 66 In 2013, there was a high-level corruption scandal in which an estimated US$32 million was stolen from Government coffers, commonly referred to as the “Cashgate” scandal. The Cashgate scandal has eroded donor confidence and trust in the Government system and is now the main reference point for corruption in Malawi. The EHP was first implemented in 2004. It prioritizes interventions and services on which to allocate the available financial resources. The EHP was revised in 2011 and later in 2016/2017 in order to make it commensurate with the health maximization principle, equity, continuum of care, complementarity of service provision, and available resources. 68 The HSSP II is Malawi’s medium term strategic plan in the health sector which outlines the national objectives, strategies and activities to improve health outcomes. The goal of the HSSP II is to move toward Universal Health Coverage (UHC) of quality, equitable and affordable quality health care with the aim of improving health status, financial risk protection and client satisfaction. 69 Jamison, D. T., H. Gelband, S. Horton, P. Jha, R. Laxminarayan, C. N. Mock, and R. Nugent, editors. 2018. Disease Control Priorities: Improving Health and Reducing Poverty. Disease Control Priorities (third edition), Volume 9. Washington, DC: World Bank. Figure 34: Trends in Total CHE, 2009/10-2017/18 US$ Millions (LHS), Percent (RHS) 800 70 697 693 685 679 670 625 623 612 Per Capita (US$) and Share of 700 Expenditure (US$, millions) 60 35 520 600 50 43 40 40 40 39 39 39 38 500 40 GDP (%) 400 30 300 11.6 11.3 11.3 11.1 10.7 20 9.8 200 8.2 7.4 9 100 10 0 0 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 2016/17 2017/18 Total CHE (US$ millions) Total CHE Per Capita (US$) Total CHE as a Share of GDP (%) Source: Authors’ construction from Malawi National Health Accounts (MoHP 2014; 2016; 2020) Composition and Growth of Health Expenditures 84. Donor funding is the largest source of funding to the health sector in Malawi. As shown in Figure 35, donor funding is significantly more than all the other sources of health funding combined. In reduced to 2 to 3 times. It is also important to note that since 2013/14, the contribution of households to the total spending in the health sector has been increasing consistently and this could have implications for equity of access to healthcare by the poor if the increased spending is from the poor. The results also show that the annual nominal growth in donor and Government expenditure on health has been lower than households and employers or NGOs since the 2013 Cashgate scandal (Figure 36). But though the households, employers and local NGOs have significantly increased their spending on health post- Cashgate, the level or volume of donor funding to the health sector is still very high. 85. High dependency on donors to finance the health sector in Malawi poses a potential risk of making health financing unsustainable and this could cause disruptions in service delivery. Since the Cashgate scandal, most of the donors have opted to provide funding to the health sector in Malawi through vertical programs and projects. By 2017/18, about 74 percent of donor funding to the health sector was off-budget and only 24 percent was pooled under the Government budget. 70 The increasing volume of off-budget donor funds has led to a proliferation of agencies and NGOs that manage financial resources on behalf of the donors. Most of the aid agencies and NGOs use their own planning, financing, procurement, and monitoring systems to manage donor funds. Use of vertical programs and parallel systems negates the five principles on aid effectiveness 71 and is a missed opportunity to improve the PFM system in the country. This data was obtained from the Aid Management Platform managed at the Ministry of Finance. 70 In line with the Paris Declaration on Aid Effectiveness, the five principles that make aid more effective are: Ownership, Alignment, 71 Harmonization, Managing for Results, and Mutual Accountability. Several donors that operate in Malawi are signatories to the Paris Declaration on Aid Effectiveness. For more information see https://www.oecd.org/dac/effectiveness/34428351.pdf Figure 35: Composition of Total CHE, 2006/07-2017/18 Percent 80% 14% 70% 12% Households and Employers Donors and Government 60% 10% 50% 8% 40% 6% 30% 20% 4% 10% 2% 0% 0% 2006/07 2007/08 2008/09 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 2016/17 2017/18 Donors Government Households Employers/Local NGOs Source: Authors’ construction from Malawi National Health Accounts (MoHP, 2020) Figure 36: Nominal Growth in Health Expenditures: 2007/08-2017/18 50% 48% 40% 30% 35% 34% 28% 27% 20% 21% 19% 18% 10% 0% Pre-Cashgate (2007/08-2013/14) Post-Cashgate (2014/15-2017/18) Government Donors Households Employers and Local NGOs Source: Authors’ construction from Malawi National Health Accounts (MoHP, 2020) 86. Health expenditures from all the main financing sources are generally aligned to the disease burden but priorities differ substantially. Spending priorities by disease differ by funding sources i.e. donor, Government, and households (Figure 37a). Of the total resources from donors in 2017/18, about 43 percent was spent on HIV/AIDS, 21 percent on Reproductive Health, 9 percent on Nutritional Deficiencies, 8 percent on Malaria, and 5 percent on non-communicable diseases (NCDs). Meanwhile, the order of prioritization of spending by the Government and households was HIV/AIDS, NCDs, Malaria, Reproductive Health, and Nutritional Deficiencies. While all the three financiers focused their spending on the top 10 causes of DALYs in Malawi, Government and household funding was more aligned to the order of priority of the disease burden (i.e. compare Figure 37a with 37b). This suggests that although health spending in Malawi is linked to the disease burden, the order of prioritization is aligned to the preferences or interests of the financiers. Meanwhile, between 2007 and 2017, there has been a reduction in the prevalence of the top 10 diseases except for ischemic health disease and sexually transmitted infections (STIs) (Figure 37b). And while the prevalence of HIV/AIDS reduced by about 70 percent, the prevalence of STIs increased by over 55 percent. This could be attributed to the substantial funding for HIV/AIDS and opportunistic infections as compared to STIs. Of the total donor funding on HIV/AIDS and STIs, only one percent was spent on STIs while the Government and households each used 7 percent of their funding for HIV/AIDS and STIs. This suggests that focusing funding on the top causes of DALYs could help to reduce the overall disease burden. 87. There is need to improve predictability of donor funding and overall allocation of resources by regions, level of healthcare, and functions. Identifying and resolving inefficiencies in the allocation and use of donor funds is critically important because donor funding is the largest source of funding to the health sector in Malawi. Interviews with MoH officials revealed that donor funds are often released late while the absorption of some of the donor funds is low due to a number of reporting requirements. Results from the 2020 National Health Accounts72 also show that about 22.8 percent of the total funds in the health sector are spent at public hospitals while only 13.8 percent of the expenditure is spent at primary healthcare facilities. Further, there is larger funding to urban than rural areas. And given that public health services are free in Malawi, the increasing share of households’ expenditure on health suggests that there are gaps in service provision. This has implications on equity of access, as discussed later. Figure 37: Spending on the Top Causes of DALYs by Financing Sources in Malawi (a) Spending on Diseases as a Share of Total (b) Top 10 causes of DALYs in 2017 and percent Spending on Health by Financing Source, 2017/18 change, 2007-2017 -70% -50% -30% -10% 10% 30% 50% 1 HIV/AIDS 7% 7% 12% 3% 3% 10% 2% 10% 2 Neonatal disorders 9% 10% 10% 3 Lower respiratory infection 21% 20% 20% 8% 4 Malaria 5% 21% 20% 5 Diarrheal diseases 6 Congenital defects 43% 29% 29% 7 Tuberculosis 8 Meningitis GOVERNMENT DONORS HOUSEHOLDS HIV/AIDS and STIs NCDs (including injuries) 9 Ischemic heart disease Malaria Reproductive Health Nutritional deficiencies TB 10 STIs All Other Diseases Source: Authors’ construction from Malawi National Health Source: Institute for Health Metrics and Evaluation Accounts (MoHP, 2020) http://www.healthdata.org/malawi Government’s Commitment to Funding the Health Sector 88. Though Malawi has one of the lowest GDP per capita in the world ($390 in 2018), the share of Government spending on health is relatively higher than other low-income countries. As shown in Table 4, Government Health Expenditure (GHE) as a share of the total General Government Expenditure (GGE) was about 8.4 percent on average over the period 2014/15 and 2018/19 while GHE as a share of GDP was 2.4 percent on average over the same period (Table 4). Compared to low-income 72 MoHP. 2020. Malawi National Heath Accounts Report for Fiscal Years 2015/16–2017/18. Lilongwe: MoHP. countries and other countries in SSA, Government’s total spending on health as a share of the total GGE and GDP in Malawi are higher (Figure 38). Specifically, GHE as a share of GDP at 2.4 percent in Malawi is higher than the average for low-income countries (1.2 percent) and the average for SSA countries [excluding high income] (1.9 percent). Similarly, in per capita terms, GHE per capita is estimated at US$9.6 for low-income countries. These findings suggest that the level of Government commitment or prioritization of the health sector in Malawi is already high and prospects for additional funding could be through improvements in efficiency in allocation and use of the available resources. Table 4: GHE as a Share of Total GGE and GDP, 2014/15-2018/19 2014/15 2015/16 2016/17 2017/18 2018/19 Average GHE as % GDP 2.5% 2.2% 2.5% 2.4% 2.6% 2.4% Source: Authors’ calculation based on Government financial reports Figure 38: GHE by Low Income Countries and other Countries in SSA, 2017 Source: Authors’ calculations from WDI data Public Expenditure on Health 89. As discussed above, the public health sector in Malawi is funded from a number of sources with the bulk of the funds coming from external sources. In this section, we only analyze public expenditures (both domestic and external) that are reflected on the national budget. Trends in Public Expenditure on Health 90. While the level of public spending on health increased in nominal terms during the period under review, in real terms the value of spending was virtually constant (Figure 39a). In nominal 2018/19 but due to high inflation, in real terms, the level of public expenditure on health has been Further, given high population growth, per capita public expenditure on health in real terms has remained virtually the same at around MWK 2,132 per annum during the period under review. Compounding the problem is that most of the key health system delivery inputs (i.e. medicines, medical supplies and equipment, ambulances, and utility vehicles) are purchased outside the country. This has decreased the ability to buy sufficient quantities of drugs and medical supplies that are paid in foreign currencies from international suppliers. 91. There are inconsistencies between expenditures and budgeted amounts. As shown in Figure 39b, public expenditure on health as a share of the total public budget for health was consistently above budget with the exception of 2016/17 (90 percent) and 2018/19 (97 percent). However, at district level, the expenditures were consistently above the budget averaging 179 percent per annum over the period unpredictable, interventions are not adequately prioritized, and that overall health services planning is weak. Figure 39: Level of Public Spending on Health and Budget Execution Rates: 2014/15-2018/19 B. Budget Execution Rate 160B 140% 2010=100) 200B 140B 120% Percentage of Budget Disbursed 126% 120B 116% 100% 150B 100B 101% 97% 80% Amount (MK) Amount (MK) 90% 80B 100B 60% 60B 145B 40% 129B 40B 105B 50B 97B 20B 20% 76B 37B 39B 35B 38B 38B 0B 0% 0B 2014/15 2015/16 2016/17 2017/18 2018/19 Nominal Real Budget Source: Authors’ calculations based on Government data Figure 40: Level of Public Spending on Health in Per Capita Nominal and Real Terms 8,000 10.7 12 9.9 9.8 7,000 9.4 8.4 10 6,000 8 5,000 7,878 4,000 6 7,194 6,049 5,715 3,000 4,573 4 2,000 2,224 2,280 1,983 2,114 2,059 2 1,000 0 0 2014/15 2015/16 2016/17 2017/18 2018/19 MK Per Capita (Nominal) MK Per Capita (Real) US$ Per Capita (Nominal) Figure 41: Public Spending on Health; and Drugs and medical supplies 200 2.21 2.5 2.14 1.93 1.73 2.0 150 1.57 Total Exp. US$ millions Per Capita (US$) 1.5 196.01 100 176.15 163.40 159.36 145.67 1.0 50 0.5 36.58 36.30 33.65 28.20 31.80 0 0.0 2014/15 2015/16 2016/17 2017/18 2018/19 Total Public Expenditure on Health Drugs & Medical Supplies Per capita Public Expenditure on Health by Level of Healthcare 92. The Malawian Government has been spending more money at district level. As shown in Figure 42, the share of the total public expenditure on health utilized at district level has been the largest with an average of 61 percent per annum over the period under review. And despite the slight decline in the share that was spent at district level in 2017/18 and 2018/19, expenditure at district level is still significantly higher than the other levels. This suggests that the Malawian Government is committed to its primary healthcare strategy. Expenditure at the five central hospitals 73 in the country has been fairly low, averaging about 17 percent of the total public expenditure on health per year over the period under review. This could have negative implications on the provision of quality specialized hospital services including management of COVID-19 cases. This is because provision of healthcare at hospital level is relatively much more expensive than primary healthcare. With the advent of the COVID-19 and rising burden of NCDs, the need for more specialized hospital services has come to the fore and this means that the Government has to balance between primary and hospital care. Assuming that the number of COVID-19 patients requiring specialized hospital care increases, there will be immense pressure on the health system, especially on the hospital sub-sector which is ill-prepared to cope with the pandemic. Currently, Malawi only has 1.3 hospital beds per 1,000 people, 74 six Intensive Care Units (ICUs), and 25 ICU beds serving a population of 18.6 million.75 73 The five central hospitals in Malawi are: Queen Elizabeth Central Hospital (Blantyre), Zomba Central Hospital and Zomba Mental Hospital (Zomba), Kamuzu Central Hospital (Lilongwe), and Mzuzu Central Hospital (Mzimba). 74 https://data.worldbank.org/indicator/SH.MED.BEDS.ZS?locations=MW 75 ICU data from Manda-Taylor, L., Mndolo, S., & Baker, T. (2017). Critical care in Malawi: The ethics of beneficence and justice. Malawi Medical Journal 29 (3), 268–271. https://doi.org/10.4314/mmj.v29i3.8 Figure 42: Public Expenditure on Health by Level of Healthcare 100% 16% 23% 16% 29% 25% 80% 60% 65% 65% 61% 58% 40% 57% 20% 19% 16% 18% 14% 17% 0% 2014/15 2015/16 2016/17 2017/18 2018/19 Central Hospitals Districts National Source: Authors’ calculations based on Government data Growth in Public Expenditures on Health by Individual Districts 93. While overall spending at district level is high, there are variations in the level and growth in spending across the districts. Review of the distribution of public expenditure on health by individual districts shows high nominal increases in spending across all the districts, ranging from 29 percent to 105 percent between 2014/15 and 2018/19 (Figure 43). Though this could be attributed to an increase in the population, this is not necessarily the case. For Lilongwe, which has the largest population in the country, the growth in health expenditures in 2018/19 as compared to 2014/15 was only 61 percent. Further, in per capita terms, public expenditure on health in Lilongwe was lower than all the other districts in Malawi. This can be attributed to the large resident population and frequent movement of people to the city to access healthcare from Kamuzu Central Hospital (Figure 43). However, there are other factors such as high disease morbidity and mortality that could also influence variances in health spending across the districts. Most of these factors are captured in the 2008 formula for allocating financial resources from the center to the districts but this formula has not been effectively applied. This is discussed below. Figure 43: Trends in Public Spending on Health by Districts Percentage Growth Spending: 2014/15 vs 2018/19 Per Capita Spending By District: 2018/19 6B 120% 6B 10000 105% 8795 99% 9000 5B 100% Total Spending - MK Billions 5B Total Spending - MK Billions 90% 8000 Growth in Spending 6636 4B 80% 6330 7000 68% Per Capita Spending 4B 61% 6000 3B 60% 3B 5000 43% 3521 4000 2B 36% 40% 2218 2B 2218 3000 29% 1B 20% 2000 1B 1930 1000 0B 0% 0B 0 Phalombe Mulanje Blantyre Nsanje Dowa Machinga Chikwawa Zomba Dedza Balaka Chitipa Salima Mwanza Chiradzulu Mchinji Ntcheu Mangochi Kasungu Thyolo Neno Lilongwe Mzimba Karonga Ntchisi Rumphi Nkhata Bay Nkhotakota Nsanje Mulanje Phalombe Likoma Mwanza Chitipa Zomba Balaka Chikwawa Dowa Dedza Chiradzulu Ntcheu Mchinji Mangochi Neno Thyolo Blantyre Lilongwe Salima Karonga Mzimba Machinga Rumphi Ntchisi Kasungu Nkhata-Bay Nkhotakota Total 2014/15 Total 2018/19 Total 2018/19 Percentage Growth 2014/15 vs 2018/19 Per Capita Spending 2018/19 Source: Authors’ calculations based on Government data 94. The existing needs-based formula for allocating financial resources to the districts is not being used. Malawi has over the years developed four formulas for distributing financial resources from the center to the districts aimed at achieving efficiency and equity objectives. The first formula was developed in 2000/01 by using population size as the key variable. This formula was revised in 2002 by including variables on poverty and under-5 mortality to strengthen its ability to reflect need. In 2008, the formula was revised to include more variables on need and cost differentials in service provision across the districts. The variables which were added to the 2008 formula are: stunting, an index for differential costs of service provision, bed capacity and outpatient utilization.76 However, in practice, this formula has not been in use and distribution of public funds to the districts is based on historical precedence. Specifically, the amount allocated to each district annually is based on the previous year’s allocation which is increased (or decreased) in line with the available budget. This approach perpetuates inequities and explains why health expenditures across the districts are not reflective of the disease burden (need) as highlighted below. 95. There are wide variations across the districts between per capita public health spending and health outcomes. As shown in Figure 44, variations in per capita spending across the districts are not reflective of the burden of stunting and under-5 mortality rates. Ironically, stunting and under-5 mortality are among the key variables which were added to the 2008 resources allocation formula. For instance, the percentage of under-5 children who are stunted in Malawi are highest in Mangochi, Neno, and Mchinji and yet the per capita sending is lower in these districts as compared to Lilongwe, Blantyre, Nkhotakota and Phalombe where the levels of stunting are relatively lower but the capita spending is higher. Similarly, while the under-5 mortality rate (U5MR) per 1,000 live births in Malawi is highest in Mchinji, Mulanje, and Phalombe districts, the per capita sending is lower than other districts with lower U5MR. Case in point is Likoma district which has lower U5MR and stunting but its per capita health spending is much more than other districts and cities with worse indicators on U5MR and stunting. These variations are also not explained by social or economic needs. Among the possible explanations on why the 2008 formula has not been applied are: (i) redistribution of the existing financial resources was not possible, (ii) the total resource envelope did not grow enough to achieve an optimal distribution of the additional resources to operationalize the formula, and (iii) wavering political support due to regular changes at ministerial level at the MoHP. In 2019, a new formula was developed by updating data on the population size, disease burden and coverage rates, unit costs of treatment, and cost variations across the districts.77 The 2019 formula is also aligned to the country’s health benefit package—the EHP—aimed at advancing the principles of health maximization, cost-effectiveness, and equity. 78 96. The MoH has to ensure that the 2019 formula is fully applied to both Government and donor resources to achieve its goal of increasing efficiency and equity in resource allocation. Currently, the 2019 formula is only intended for Government resources which constitute 24 percent of the total spending in the health sector, leaving out the largest financier, the donors, who contribute 58 percent of the total spending in the health sector. Furthermore, considering that this formula only focuses on the allocation of financial resources for drugs and operational grants in the public sector at district level; the MoH also needs to look closely at the funding and distribution of human resources, infrastructure, and equipment. As presented in the next sections, consolidated funding on these items is 76 Manthalu, G., Nkhoma, D., & Kuyeli, S. 2010. Simple versus composite indicators of socioeconomic status in resource allocation formulae: the case of the district resource allocation formula in Malawi. BMC health services research, 10, 6. https://doi.org/10.1186/1472-6963-10-6 77 Twea P, Manthalu G, Mohan S. 2020. Allocating resources to support universal health coverage: policy processes and implementation in Malawi. BMJ Global Health; 5:e002766. https://gh.bmj.com/content/bmjgh/5/8/e002766.full.pdf high, especially for human resources. For instance, if this formula had been applied in 2018/19, it could only have influenced 27 percent of the total public funds at district level (Figure 44) with no impact on the funding for human resources. Therefore, optimal funding and distribution of human resources, infrastructure, and equipment are key to promoting equity of access. Figure 44: Per capita expenditure by district vs health outcomes Stunting vs Per capita Spending U5MR vs Per capita Spending 10,000 50 10,000 140 Per capita Spending (MK) U5MR per 1,000 live births 120 8,000 40 8,000 U5 Children Stunted (%) Per capita Spending (MK) 100 6,000 30 6,000 80 60 4,000 20 4,000 40 2,000 10 2,000 20 0 0 0 0 Dedza Dowa Chikwawa Likoma Mangochi Mchinji Chiradzulu Lilongwe Mulanje Blantyre Nsanje Phalombe Mzimba Machinga Zomba Salima Chitipa Balaka Mwanza Karonga Ntcheu Ntchisi Kasungu Rumphi Neno Thyolo Nkhata-Bay Nkhotakota Mulanje Phalombe Nsanje Balaka Dedza Dowa Chikwawa Likoma Zomba Chitipa Mwanza Mchinji Ntcheu Mangochi Chiradzulu Neno Thyolo Lilongwe Blantyre Machinga Salima Karonga Mzimba Ntchisi Rumphi Kasungu Nkhata-Bay Nkhotakota Per capita Spending - 2018/19 Stunting Per capita Spending - 2018/19 U5MR Source: Authors’ calculations from Government data and Malawi Demographic and Health Survey 2015/16 Public Expenditure on Health by Key Health Systems Inputs – National Picture 97. Half of the total public funds in the health sector in Malawi are spent on personnel emoluments (Figure 45). Over the period under review, an annual average of 51 percent of the total public expenditure on health was on personnel emoluments. The public sector health wage bill as a share of the total public expenditure on health in Malawi is fairly low except if compared to some of the lower middle-income countries in the African region i.e. Angola, Tanzania, Kenya, and Lesotho (Table 5). After personnel emoluments, expenditure on ORT comes second followed by expenditures on drugs and medical supplies, and lastly infrastructure development. During the last two years of the review (2017/18 and 2018/19), expenditure on drugs and medical supplies as a share of the total public expenditure on health remained constant at 16 percent (Figure 45). Compared to the share of spending on drugs in other African countries, the share of spending in Malawi is lower. The current level of funding only caters for about six months’ supply and this leads to persistent shortages at the hospitals and health centers. 98. Development expenditure in Malawi’s health sector has varied considerably in recent years. As shown in Figure 45, development expenditure as a share of the total public expenditure on health has varied from 5 percent in 2014/15 to 16 percent in 2015/16, and has been between 12 and 13 percent in 2017/18 and 2018/19. The main areas for the expenditure on development were on physical structures and medical equipment which constituted about 98 percent of the total expenditure on infrastructure development. Increased spending on infrastructure is one of the goals outlined in Malawi’s HSSP II which among other things seeks to increase physical access to health facilities by rehabilitating and expanding health infrastructure countrywide. Figure 45: Public Expenditure by Key Health Systems Inputs 100% 5% 3% 16% 13% 12% 22% 23% 80% 16% 16% 23% 17% 21% 21% 60% 22% 13% 40% 55% 52% 51% 49% 49% 20% 0% 2014/15 2015/16 2016/17 2017/18 2018/19 Personnel Emoluments ORT Drugs & Medical Supplies Development Source: Authors’ calculations based on Government data Table 5: Public Health Sector Wage Bill as a share of the total Public Expenditure on Health Country Wage bill as a share of GHE Source Zambia 62 PER 2018 Zimbabwe 60 HFSA 2019 Seychelles 54 PER 2014 Malawi 51 PER 2020 Angola 44 PER 2017 Tanzania 43 PER 2020 Kenya 41 PER 2014 Lesotho 35 PER 2017 Source: Authors’ review of health PERs in Sub-Saharan Africa Public Expenditure on Health by Key Health Systems Inputs – District Level 99. Similar to the overall picture on public spending on health by key health systems inputs, the share of spending on personnel emoluments at district level is high. As highlighted in Figure 46, the share of spending on personnel emoluments increased from 65 percent in 2016/17 to 73 percent in 2018/19. Due to the large expenditure on personnel emoluments, expenditure on ORT is very low. This means that service delivery activities such as outreach and supportive supervision at district level are affected. And since the bulk of the resources at district level are spent on personnel emoluments, it will be increasingly difficult for the Government to recruit additional health workers in the public health sector and/or to make efficiency gains through the existing funding. Furthermore, at 18 percent of the total public expenditure on health at district level, spending on drugs and medical supplies is low. This has negative implications on the supply and access to quality and efficacious medicines in the country. Figure 46: Expenditure by Key Health Systems Inputs at District Level 100% 23% 19% 18% 29% 25% 80% 9% 10% 13% 10% 9% 60% 40% 71% 73% 64% 62% 65% 20% 0% 2014/15 2015/16 2016/17 2017/18 2018/19 Personnel Emoluments ORT Drugs & Medical Supplies Source: Authors’ calculations based on Government data Efficiency and Value for Money 100. This section analyzes the pathways through which resources in the health sector could be lost or wasted are traced from financing to health outcomes. The analysis was undertaken by looking at three pathways: (a) potential inefficiencies in the allocation of funds and/or purchase of key health systems inputs (such as ORT, human resources, drugs and medical supplies, medical equipment and physical infrastructure); (b) potential inefficiencies in transforming the available inputs into quality health services and outputs given the country’s total CHE per capita; and (c) the country’s ability to translate the available health services and outputs into better health outcomes. The latter is in fact a value-for-money or effectiveness analysis. This process is depicted in Figure 47. While the previous sections provided information on the level and composition of the health expenditures, and allocation of funds by level of healthcare and key health systems inputs (first pathway); the subsequent analyses provide further insights on the first pathway as well as the second and third pathways. Figure 47: Pathways for potential inefficiencies in the Health Sector Source: Authors’ Construction Public Financial Management 101. One of the major causes of inefficiencies in the health sector in Malawi is the weak PFM system in the public sector. Effective management of public expenditures on health is essential to increasing coverage and achieving better health outcomes in Africa. 79 As part of this PER, a comprehensive assessment of the ability of the PFM system to facilitate quality health service delivery was conducted at district hospitals and health centers.80 The results show inadequate compliance with existing guidelines on the management of public funds, especially at Government health facilities. For example, while health expenditures at district level are in excess of the budgeted amounts, there are persistent delays in the transfer of the funds, and inter-sectoral borrowing of earmarked funds. Delays in funding means that the money is not remitted as planned or when it is needed i.e. monthly disbursements are unpredictable. In addition, budgetary releases to the districts are not usually communicated to the health providers and this often leads to poor accountability. 102. Planning and budgeting processes are in place at district level but they do not effectively support prioritization of activities. Service providers at district hospitals and health centers are usually not informed about the available funding or ‘in-kind’ support for the following year. In addition, participation of service providers in the planning and budgeting processes is marginal, and this weakens the prioritization process. Further, by using vertical programs, donors also contribute to fragmentation of the planning and budgeting, delivery, and monitoring and evaluation (M&E) systems in the health sector. For instance, in addition to the existing Government system, there are multiple financial management and M&E systems that are managed by donors. This practice is in conflict with the Paris declaration on aid effectiveness and the Accra Agenda for Action. 103. Service providers are not recognized in the budget. The bulk of the health services in Malawi are delivered at primary healthcare facilities (district hospitals and health centers). These facilities are managed by district health management teams (DHMTs) and are not categorized as spending units or cost centers. The DHMTs are in charge of accounting, financial management, and M&E, which is a conflict of interest. While financial management systems are in place to identify spending by economic classification at the DHMT level, it is not possible to determine the amounts spent by individual hospitals and health centers or how this relates to the volume of services delivered, quality or need. 104. There are no comprehensive financial reports in the health sector. Financial reporting at district level is done through the Navision accounting software while the central Government’s financial management information system (FMIS) uses a different application called Epicor. These two systems are not integrated, making it difficult to generate comprehensive financial reports across all levels of Government. Furthermore, donor financing is generally not captured in the FMIS, which leads to only partial financial reporting. Given that about 60 percent of the total health expenditure in Malawi is provided by donors, this means that a large part of the health expenditures are not routinely reported. Therefore, it is difficult to establish a link between total expenditures (financing), inputs, service delivery outputs, and outcomes. Consequently, evaluating the effectiveness of spending in the health sector in Malawi is challenging. 105. Execution protocols emphasize control over flexibility. Budget execution protocols at Government health facilities at district level require input-based controls of the line item budget with Kutzin, J. 2016. Why Does Public Finance Matter for UHC? http://www.who.int/health_financing/events/JosephKutzin-why-does- public-finance-matter-for-uhc.pdf?ua=1 80 In line with the Local Government Act of 1998, operational budgets for the health sector were devolved to the district councils in 2005. As such, health service delivery at community, primary and secondary levels is under the mandate of the district councils while provision of tertiary level services is the responsibility of the MoHP. limited opportunity for virement. Major adjustments, such as shifting funds across budget lines, require a supplementary budget. These rigidities have prompted DHMTs to avoid using the electronic system, preferring instead manual accounting which bypasses internal controls. While this provides greater flexibility, it undermines the PFM Act. Furthermore, manual accounting raises accountability concerns, and contributes to the accumulation of arrears. Therefore, there is need for greater flexibility for budget execution at the district level. Lessons can be drawn from health facilities which are managed by the Christian Health Association of Malawi (CHAM). CHAM facilities receive a global budget and have greater flexibility with regards to how they spend money. As highlighted in Table 16, budget formulation, execution, and evaluation at CHAM facilities is relatively better than at the Government facilities. Therefore, Government execution protocols could follow the CHAM approach without introducing significant fiduciary risks. Table 6: Budget Formulation, Execution and Evaluation by Facility Ownership HEALTH CENTRES/DISPENSARIES HOSPITALS Budget Service Delivery Measures Budget Service Delivery Measures Phase E1 E2 Q A Phase E1 E2 Q A Govt D D D+ D+ Govt D+ D C D Formulation Formulation CHAM D D C D+ CHAM B+ C B B Govt D+ D D D+ Govt C+ D D+ C+ Execution Execution CHAM D+ C C D+ CHAM B B A A Govt D D D D+ Govt D D D C Evaluation Evaluation CHAM C C D C CHAM A A A A E1=Efficiency, E2=Equity, Q=Quality, A=Accountability Technical Efficiency 106. Despite the relatively high expenditure on personnel emoluments in Malawi, there is still a critical shortage of clinical health workers. As shown in Figure 48, in 2018/19, the number of doctors per 1,000 population was less than one across all the 28 districts in Malawi. Secondly, the distribution of the doctors varies by district with most of the doctors in the country working in Blantyre which has 0.029 doctors per 1,000 people compared to Machinga and Chitipa districts which had a paltry 0.004 doctors per 1,000 people. The staffing levels in the other two cities in Malawi (Lilongwe and Zomba) are lower than Blantyre but better than most of the other districts. Considering that most of the doctors in Malawi work in hospitals, this means that hospitals in rural districts find it difficult to attract doctors. The uneven distribution of doctors in Malawi can also be due to the high number of Government and privately-owned tertiary hospitals in the three cities (Blantyre, Lilongwe, and Zomba). Figure 48: Number of Doctors per 1,000 Population, 2018/19 0.04 0.029 0.024 0.022 0.03 0.020 0.017 0.03 0.015 0.013 0.012 0.012 0.012 0.02 0.011 0.010 0.009 0.008 0.008 0.008 0.007 0.007 0.02 0.006 0.005 0.005 0.005 0.005 0.005 0.005 0.004 0.004 0.01 0.01 0.00 Salima Nkhotakota Chitipa Rumphi Kasungu Mchinji Ntchisi Karonga Dowa Ntcheu Nkhata Zomba Thyolo Neno Mulanje Blantyre Nsanje Lilongwe Mangochi Phalombe Mzimba Chiradzulu Mwanza Machinga Chikwawa Dedza Balaka Source: Authors’ calculations based on Government data 107. There is a direct relationship between public health expenditure per capita and availability of clinical staff (Figure 49a). This means that the distribution of health workers is a key factor in how financial resources are distributed in the public health sector in Malawi. Secondly, Figure 49b shows that there is an inverse relationship between the availability of health workers and total outpatient visits. Generally, districts with lower staffing levels see more outpatients than those with higher staffing levels. This scenario is prevalent in cities and districts where the size of the population is large because areas with larger populations often have scale efficiency i.e. lower unit costs of service delivery. In districts with high staffing levels but low outpatient visits, the implication is that some of the health workers are being underutilized. The other possible explanation is that the quality of outpatient services is poor in the districts with high staffing levels, and hence the low utilization. As revealed in previous sections, high expenditure on personnel emoluments but low spending on medicines and other medical supplies contributes to the limited access to quality and efficacious medicines in the country. To address this problem, there is need to improve resource allocation so that there is optimal distribution of financial resources across all the key health systems inputs. In particular, distribution of human resources, medicines and other essential health commodities should be aligned to the need of the districts. Figure 49: Productivity of Health Workers by Districts - 2018/19 Per Capita Spending vs Staffing Levels Staffing Levels vs Outpatient Visits Source: Authors’ calculations based on Government data 108. Malawi performs better than most of the low-income countries in SSA in service coverage (Figure 50). To benchmark Malawi against other countries, the Universal Health Coverage (UHC) service coverage index81 was used. The UHC service coverage index is a single indicator that is computed from 14 tracer indicators to monitor coverage of essential health services as part of the Sustainable Development Goal (SDG) number 3: “Ensure healthy lives and promote well-being for all at all ages .”82 With a score of 46 in 2017, Malawi stands out among countries with similar per capita total current health expenditure such as Tanzania, Togo, and Burkina Faso (Figure 50). Malawi’s UHC index is also slightly higher than the SSA regional average of 44.83 Malawi needs to sustain this good performance and improve further through efficient utilization of the available resources. Figure 50: UHC Index for Low Income Countries in SSA, 2017 Source: Authors’ calculations based on WDI data Value for Money 109. Malawi performs better than most of the low-income countries in SSA in transforming the available health services and outputs into better child health outcomes. Comparison of the UHC index scores to the level of under-5 mortality (Figure 51) shows that the performance in Malawi is better than the other low-income countries in SSA. However, Malawi has relatively higher maternal mortality than other low-income countries (i.e. Mozambique, Burkina Faso, and Madagascar) despite having a higher UHC index score. In other words, Malawi is not effective at translating the available services and outputs into better maternal health outcomes. This could be attributed to low quality of maternal healthcare as outlined in the 2015-16 Demographic and Health Survey (DHS). The DHS shows that while the percentage of births occurring at a health facility (91 percent) and percentage of births attended by a skilled provider (90 percent) are high in Malawi, the quality of antenatal and maternal delivery services are poor.84 Further, critical shortage of key health systems inputs (human resources, medicines and medical supplies, poor/inadequate infrastructure), and poor governance and accountability also 81 https://www.who.int/healthinfo/universal_health_coverage/report/uhc_report_2019.pdf?ua=1 82 ibid. Essential health services are services that all countries, regardless of their demographic, epidemiological or economic profile, are expected to provide. The package of essential health services includes reproductive, maternal, newborn and child health; infectious diseases; non-communicable diseases; and service capacity and access. 83 https://data.worldbank.org/indicator/SH.UHC.SRVS.CV.XD National Statistical Office and ICF International. 2016. Malawi Demographic and Health Survey 2015-16: Key Indicators Report . Zomba, Malawi, and Rockville, Maryland, USA. NSO and ICF International contribute to provision of low quality maternal healthcare in Malawi. 85 Consequently, Malawi’s score of 32.2 on the health access and quality (HAQ) index86 is low. The HAQ index measures healthcare access and quality for diseases and injuries considered amenable 87 to health care. The top 10 causes of amenable death and disability in Malawi are presented in Figure 37b above. 110. For service coverage to translate into improved outcomes, there is need for greater focus on quality of healthcare. This could be achieved by reprioritizing public spending on health so that there is an optimal balance between spending on personnel emoluments and other key health systems inputs. Having a better mix of service inputs and reconfiguring the financing mechanism from an input- based to a results-based financing system is also critical. Figure 51: Translation of Outputs in Health Outcomes UHC Index vs Under-5 Mortality Rate UHC Index vs Maternal Mortality Rate Source: Authors’ calculations based on WDI data Equity Analysis 111. Attainment of equitable access to quality healthcare services is one of the key features of the national health policy of Malawi. The national health policy of Malawi is operationalized through the HSSP II which aims to create “a situation where everyone—irrespective of their ability-to-pay—gets the health services they need in a timely fashion without suffering any undue financial hardship because of receiving care.” Having a national health policy and strategic plan in place affirms the Government’s commitment to achieving the health-related SDG, and UHC. 88 Aligned to the HSSP II is the EHP where Malawi has defined a list of priority interventions and services on which financial resources are allocated. The EHP is commensurate with the health maximization principle, equity, continuum of care, 85 Ministry of Health and ICF International. (2014). Malawi Service Provision Assessment 2013-14. Lilongwe, Malawi, and Rockville, Maryland, USA: MoH and ICF International 86 Fullman, N., J. Yearwood, S. M. Abay, C. Abbafati, F. Abd-Allah, J. Abdela, A. Abdelalim, Z. Abebe, T. A. Abebo, and V. Aboyans. 2018. “Measuring Performance on the Healthcare Access and Quality Index for 195 Countries and Territories and Selected Subnational Locations: A Systematic Analysis from the Global Burden of Disease Study 2016.” The Lancet 391: 2236–2271. 87 The HAQ index incorporates 32 causes of disease and injury which in the presence of timely and quality healthcare should not result in death. These include maternal and perinatal mortality; infectious diseases; neoplasms; nutritional, endocrine, and metabolic diseases; neurologic disorders; cardiovascular diseases; respiratory and digestive system diseases; genitourinary system diseases; and external causes. https://www.paho.org/hq/dmdocuments/2013/annex-basic-indicators-2013.pdf.pdf. 88 SDG 3, target 3.8 requires all countries to “achieve universal health coverage, including financial risk protection, access to quality essential health-care services and access to safe, effective, quality and affordable essential medicines and vaccines for all.” complementarity of service provision, and available resources. The Government provides free EHP services through all the Government health facilities, and where there is no Government facility, CHAM health facilities are contracted (as outlined in Service Level Agreements [SLAs]) to provide a package of selected health services for free. 112. This section uses two population-based surveys, the Integrated Household Surveys 3 and 4 which were produced in 2010/2011 and 2016/17, respectively, to answer the following equity- related questions: (i) What is the composition of health spending at household level and how do the spending patterns differ across households from different socio-economic backgrounds? (ii) Do certain households experience hardship or catastrophic expenditure when accessing health services? and (iii) How are the health services utilized across households from different socio-economic backgrounds? The analysis focuses on the financing and distributional impact between 2010/2011 and 2016/17 across districts and income groups to determine whether regions with poor health outcomes and poor households have benefitted from public resources over the years. The study does not look at each individual health strategy or reform per se but seeks to establish if the health system as a whole is pro- poor. Spending on Health Services at Household Level What is the composition of health spending at household level and how do the spending patterns differ across households from different socio-economic backgrounds? 113. Though out-of-pocket expenditure on health as a share of the total current health expenditure has been increasing consistently since 2012/13 (Figure 52), poor households have not been affected. This is highlighted in Figure 53a which shows a decline in the total households spending on health as a share of the total household expenditure among the poorest households. On the other hand, for the more affluent households (quintiles 2 to 5), there has been an increase in the total households spending on health as a share of total household expenditure between 2010/11 and 2016/17. However, when faced with illnesses requiring medicines, the burden on poorer households has increased, although it has remained steady for the wealthiest quintile of households (Figure 53c). The increasing household spending on medicines between 2010/11 and 2016/17 for most of the income groups could be attributed to the inadequate public spending on medicines over the years as highlighted in earlier sections. Persistent shortages of medicines at Government health facilities prompts households to buy the medicines from private drug stores and pharmacies. Figure 52: Out-of-Pocket Expenditure on Health as a Share of total CHE 14% 12.7% 12.7% 12.4% 12% 10.9% 10.8% 10.2% 10% 9.0% 8.1% 8% 6.6% 6% 4% 2% 0% 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 2016/17 2017/18 Source: Authors’ construction from Malawi National Health Accounts (MoH, 2020) Figure 53: Level and Composition of Households Spending on Health by Income Quintile A. Total Households Spending on Health as B. Outpatient as a Share of Total a Share of Total Household Expenditure Households Spending on Health 2.5 35 30 2.0 25 PERCENT 1.5 20 PERCENT 15 1.0 10 0.5 5 0.0 0 Poorest Q2 Q3 Q4 Richest Poorest Q2 Q3 Q4 Richest 2010/11 2016/17 2010/2011 2016/2017 C. Medicines as a Share of Total D. Hospitalization as a Share of Total Households Spending on Health Households Spending on Health 80 25 20 60 15 PERCENT PERCENT 40 10 20 5 0 0 Poorest Q2 Q3 Q4 Richest Poorest Q2 Q3 Q4 Richest 2010/2011 2016/2017 2010/2011 2016/2017 Source: Authors’ calculations from Malawi IHS 3 and 4 Catastrophic Health Expenditure Do certain households experience hardship or catastrophic expenditure when accessing health services? 114. Studies on catastrophic health expenditure are anchored on the notion that health care payments should not exceed a given threshold such that the households experience financial hardship when assessing healthcare services (Wagstaff and van Doorslaer, 2003). Households with expenditures on healthcare above an acceptable threshold are regarded to be experiencing ‘catastrophic health expenditures.’ In the literature, the commonly used threshold values for assessing catastrophic health expenditures are 10 percent for total household expenditure and 40 percent for non-food expenditure (O'Donnell et al. 2008). The incidence of catastrophic health expenditure is computed as a proportion of households, expressed by head count, that incur catastrophic payments for healthcare. In our analysis, we used the 40 percent threshold which means that we estimated the proportion of households spending more than 40 percent of their total non-food expenditure on health. 115. The results show that between 2010/11 and 2016/17, the proportion of households in the lowest quintile (poorest) incurring catastrophic health payments decreased (Figure 54a). On the other hand, the proportion of households incurring catastrophic health payments for all the other households (quintile 2 to 5) increased between 2010/11 and 2016/17 (Figure 54a). Further, between 2010/11 and 2016/17, the incidence of catastrophic health payments also increased significantly in rural areas while there was a reduction in urban areas (Figure 54b). This is not surprising because most of the settings in Malawi are predominantly rural and as such, the findings in Figure 54a are triangulated. Thus, we conclude that despite having in place a free healthcare policy and existence of SLAs with CHAM facilities, catastrophic health expenditures are still prevalent in rural areas in Malawi. Therefore, the chances of poor households being exposed to financial hardships when accessing healthcare and sliding into poverty are very likely in Malawi. Figure 54: Percentage of households spending above 40 percent of their total non-food expenditure on health (a) Income groups (b) Urban-Rural 2010/2011 2016/2017 2010/2011 2016/2017 2.5 2.0 2.0 1.6 1.5 1.2 Percent Percent 1.0 0.8 0.5 0.4 1st 2nd 3rd 4th 5th 1st 2nd 3rd 4th 5th Rural Urban Rural Urban 0.0 0.0 Source: Authors’ calculations from Malawi IHS 3 and 4 Equity of Access to Health Services How are the health services utilized across households from different socio-economic backgrounds? 116. Equity of access to health services in Malawi has improved over the years. Between 2010/11 and 2016/17, there was a slight increase in spending on outpatient visits as a share of total households spending on health among the poorest and richest households but overall, the poorest households continued to experience a lower burden (Figure 53b). Furthermore, households spending on hospitalization declined across all the different income groups during the period under review (Figure 53d). This could be attributed to increased utilization of free health services at Government and CHAM health facilities by the poorest households. As shown on Figure 55, there was an increase in the utilization of health services at Government and CHAM health facilities by the poorest households between 2010/11 and 2016/17 while there was a decline in utilization for the richest households. However, the poorest households were consuming more health services at the Government health facilities as compared to the CHAM health facilities. Further, despite the decline in utilization for the richest households in 2016/17, they still continued consuming more health services at CHAM facilities as compared to the poorest households (Figure 55b). Low utilization of health services at CHAM facilities by the poorest households could be associated with the user fees which are applicable at CHAM facilities for services which are not covered under the SLAs and for CHAM facilities that do not have SLAs with the Government. There is need to increase the number of SLAs between the Government and CHAM. Figure 55: Use of health services by income quintiles 25 A. Government Health Facilities 35 B. CHAM Health Facilities 30 20 25 15 20 PERCENT PERCENT 15 10 10 5 5 0 0 Poorest Q2 Q3 Q4 Richest Poorest Q2 Q3 Q4 Richest 2010/2011 2016/2017 2010/2011 2016/2017 Source: Authors’ calculations from Malawi IHS 3 and 4 117. Utilization of outpatient and inpatient services at Government, CHAM and private health facilities is more among the rich than the poor (Figure 55). This is despite the increased utilization of health services among the poorest households at Government and CHAM health facilities between 2010/11 and 2016/17. This is because the rich households continued consuming more healthcare services at CHAM and private health facilities in both 2010/11 and 2016/17 in addition to what they consume from the government facilities. Further, while utilization of health services increased among the poorest households, there is still a large number of poor households in quantile 2 who may not have benefited. Malawi is the sixth poorest country in the world and 69.8 percent of the population is below the international poverty line of US$1.90 per day. 89 Therefore, though the provision of free healthcare services at Government and selected CHAM health facilities has contributed to improved equity of access, overall utilization of outpatient and inpatient services at Government, CHAM and private health facilities is still higher among the rich than the poor. To attain its long-term goal of attaining equity of access to quality healthcare, the Malawian Government needs to improve quality of healthcare services at Government health facilities; and also increase access to CHAM and private health facilities in areas where there are no Government health facilities. Full application of the existing needs-based resources allocation formula at district level is also important. 89 ibid Figure 56: Utilization Incidence of Outpatient and Inpatient Services: 2010/11 vs 2016/17 (a) Outpatient (b) Inpatient Source: Authors’ calculations from Malawi IHS 3 and 4 Conclusions and recommendations 118. Malawi performs better than most of the low-income countries in SSA in service coverage, and in transforming the available health services and outputs into better child health outcomes. However, Malawi is not as effective in translating the available services and outputs into better maternal health outcomes. This could be attributed to the low quality of maternal healthcare services. 119. The COVID-19 pandemic, if not addressed, could reverse the gains in maternal and child health outcomes that Malawi has registered since 2000. Since the advent of the COVID-19 in Malawi, there has been a reduction in the utilization of some of the key reproductive, maternal and child health services. This is mainly due to: (i) disruptions in the procurement and distribution of medicines and other essential medical commodities, (ii) greater emphasis on COVID-19 as compared to other essential services, and (iii) patients’ fears of contracting the disease if they go into the health facilities. 120. Financing and provision of health services in Malawi are likely to be affected by the COVID- 19 pandemic whether the number of cases remains low or increases. Additional investments in the health system are required to: (i) prevent further spread of the disease, (ii) to treat the sick, and (iii) to maintain the provision of other essential services. This will require additional domestic and external funding but given the negative impact of the COVID-19 outbreak on economic growth and resource mobilization worldwide, expenditure on health and other social sectors could shrink rather than increase. 121. Malawi’s total spending on health in per capita terms and as a share of GDP is higher than other low-income countries. However, total health spending per capita estimated at US$39 per year is insufficient to provide essential healthcare as outlined in the country’s health benefit package—the EHP. This has contributed to gaps in service delivery. 122. The Malawian Government’s commitment to funding the health sector is already high. Compared to other low-income countries, Government’s spending on health as a share of the GDP and total Government spending in Malawi are higher. This suggests that the probability of obtaining additional resources from the Government is low. Further, in real terms, public expenditure on health has been shrinking over the years while the population and disease burden have been rising. 123. Donor funding is still the largest source of funding to the health sector but the annual nominal growth in donor expenditure on health has been low since the Cashgate scandal. Households, employers and local NGOs have significantly increased their spending on health post- Cashgate as compared to the growth in donor and Government spending. 124. High dependency on donors to finance the health sector in Malawi poses a potential risk of health financing becoming unsustainable and this could cause disruptions in service delivery. Since the Cashgate scandal, most of the donors have opted to provide funding to the health sector through vertical programs and projects. By 2017/18, about 74 percent of the donor funding to the health sector was off-budget. Moreover, most of the aid agencies and NGOs use their own planning, financing, procurement, and monitoring systems to manage donor funds. Use of vertical programs and parallel systems negates the five principles on aid effectiveness90 and is a missed opportunity to improve the PFM system in the country. 125. While all the three main financiers (donors, Government, and households) focus their spending on the top 10 causes of DALYs, Government and households’ funding is more aligned to the order of priority of the disease burden. Furthermore, donor funds are often released late while the absorption is low due to a number of reporting requirements. Resolving inefficiencies in the allocation and use of donor funds is critically important because donor funding is the largest source of funding to the health sector in Malawi. 126. In line with the national health policy, the bulk of the financial resources in the public health sector are spent at district level. However, more than half of the total public funds in the health sector in Malawi are spent on personnel emoluments. Due to the large expenditure on personnel emoluments, expenditure on drugs and medical supplies, and ORT is low. Compared to the share of spending on drugs in other African countries, the share of spending on drugs in Malawi, at 16 percent, is lower. The current level of funding only caters for about six months’ supply and this leads to persistent shortages at the hospitals and health centers. 127. Despite the relatively high expenditure on personnel emoluments in Malawi, there is still a critical shortage of clinical health workers while some of the existing health workers are underutilized. Most of the doctors in the country work in Blantyre which has 0.03 doctors per 1,000 people. This is significantly less than 1 doctor per 1,000 population. However, districts with lower staffing levels see more outpatients than those with higher staffing levels which suggests that some health workers are not being fully utilized. 128. Predictability of funding is low. Spending is not aligned to the budget and this could be due to weaknesses in domestic resource mobilization at national level, and gaps in health services planning. Constant below and over budget expenditures raise concerns about the credibility of the budget as a planning and resource allocation tool in the health sector. 129. Malawi has over the years developed four needs-based formulas for distributing financial resources from the center to the districts aimed at achieving efficiency and equity objectives . However, the current formula is not being used, leading to wide variations across the districts between per capita public health spending and health outcomes. Further, the existing formula only focuses on the In line with the Paris Declaration on Aid Effectiveness, the five principles that make aid more effective are: Ownership, Alignment, 90 Harmonization, Managing for Results, and Mutual Accountability. Several donors that operate in Malawi are signatories to the Paris Declaration on Aid Effectiveness. For more information see https://www.oecd.org/dac/effectiveness/34428351.pdf allocation of Government funds, and financial resources for drugs and ORT. These constitute a very small portion of the overall resource envelope in the health sector. 130. Gaps in the PFM arrangements have made it difficult to deliver health services at Government hospitals and health centers. The PFM system is characterized by inadequate compliance with guidelines, poor accountability, and limited communication and dialogue. Budget execution also emphasizes control over flexibility, which has led to district authorities sidestepping the system. Moreover, poor integration of the district and central Government financial reporting systems limits budget monitoring and reporting. 131. Although out-of-pocket expenditure on health as a share of the total current health expenditure has been increasing consistently since 2012/13, poor households have not been affected. The increasing burden of out-of-pocket spending has been on the wealthy households. Further, the proportion of poor households incurring catastrophic health payments has decreased. However, geographically, the incidence of catastrophic health payments has increased in rural areas while there has been a reduction in urban areas. 132. Equity of access to health services in Malawi has improved over the years but there is room for further improvement. Increased equity of access to health services could be attributed to increased utilization of free health services at government and CHAM health facilities by poor households. However, poor households still consume more health services at government health facilities (where quality of healthcare is low) as compared to the CHAM and private health facilities. Furthermore, overall utilization of health services at government, CHAM and private health facilities is still higher among the rich than the poor. Recommendations 133. There is need to sustain the gains made in transforming the available health services and outputs into better child health outcomes while scaling-up the delivery of quality maternal health services. This will require addressing underlying issues such as the high adolescent pregnancy rate, high number of teenage marriages, low education status for the mothers, poor maternal nutrition, and low access to social protection services. 134. Additional financial and material resources will be required to prevent further spread of the COVID-19, to treat the sick, and to maintain the provision of other essential services. At a minimum, the Government needs to ensure that the existing financial resources in the health sector are not reduced. If an economic crisis forces the issue, the Government could consider cutting certain areas of the national and/or health budget to sustain the provision of health services with the least possible impact on health and other social outcomes—e.g. cutting capital spending or temporarily suspending some non-essential services. 135. The Government should continuously monitor how the COVID-19 is affecting supply and demand for health services, consistently map how the available resources are allocated and spent, and undertake timely procurement and distribution of medicines and other essential medical commodities. Scaling-up risk communication on the COVID-19 and maintaining the provision of essential health services during the COVID-19 outbreak is critical to reducing morbidity and mortality due to the COVID-19 itself and from other diseases. 136. The MoHP needs to improve efficiency in the allocation and use of available resources by: (i) re-prioritizing Government spending as highlighted above, and (ii) developing a financial sustainability plan that could extend available Government and donor funding for an additional 3-5 years. This is because the likelihood of increasing domestic and external funding is very low given the already high Government and donor spending in the health sector; and reduced revenue generation capacity worldwide. 137. The Government could consider developing and implementing a comprehensive health financing strategy (HFS) to guide resource mobilization, pooling, allocation, and purchasing of health services. The strategy needs to encompass the financial sustainability plan (alluded to above) and viable strategies for promoting financial resilience and efficiency. The HFS being proposed in Malawi could focus more on resilience and efficiency enhancing measures. 138. In order to increase its effectiveness, donor funding needs to be aligned to Government systems at both district and national levels. Aligning donor funding to the Government system is critical to improving the overall allocation of funds, governance and accountability in the health sector. Immediate actions include: (i) development of a system for routine mapping and tracking of external funds at both central and district levels, (ii) aligning donor funding to the order of priority of the disease burden, and (iii) increasing predictability of donor funding through the use of joint budgeting, disbursement, financial management, procurement, and reporting mechanisms. 139. The Government needs to enforce use of the existing PFM guidelines, and facilitate improvements in communication and dialogue on planning and budgeting at district level. There is also need for greater flexibility on budget execution at Government health facilities similar to the flexibility enjoyed at CHAM health facilities. Health budgets at district level also need to be ring-fenced to avoid inter-governmental transfers when the funds are disbursed to the district councils. There is also a need to integrate accounting systems at the district and central Government levels to enable the generation of integrated financial reports in the health sector. 140. To achieve its efficiency and equity objectives, the Government needs to fully apply the revised (2019) district-level resources allocation formula to both Government and donor resources across all the districts. Furthermore, considering that this formula only focuses on the allocation of financial resources for drugs and operational grants in the public sector at district level; the MoH also needs to look closely at the funding and distribution of human resources, infrastructure, and equipment. Focusing on the resources allocation formula alone will not lead to the desired improvements in efficiency and equity. 141. Comprehensive training and mentorship of district authorities and service providers on health services planning and budgeting could help to improve the allocation of resources and budget execution. Further, there is need for consistent advocacy on evidence-based planning and application of the revised resource allocation formula among policymakers and planners. 142. There is need to improve purchasing and value-for-money in the health sector. Currently, the health sector in Malawi is characterized by a high expenditure on personnel emoluments, critical shortage of clinical health workers, and low utilization of the existing health workers. The Government needs to address this problem by: (i) distributing the available workforce optimally across all the districts, and (ii) increasing the productivity of the existing health workers by introducing performance-based financing (PBF) schemes. By using PBF, financing to health facilities could be distributed on the basis of outputs produced rather than inputs. PBF schemes are currently being used in Rwanda, Tanzania, Uganda, Zambia, and Zimbabwe. 143. Catastrophic health expenditures are still prevalent in rural areas even though there has been some improvement in financing and access to health services by the poor. Therefore, to achieve universal access to quality healthcare, the Government needs to improve quality of healthcare services at Government health facilities since this is where most of the poor people access services. In addition, there is need to further increase access for poor households to CHAM and private health facilities especially in areas where there are no Government health facilities. This could be achieved by introducing vouchers in addition to the existing SLAs between the Government and CHAM. Social Protection Sector Review Introduction 144. Social protection continues to grow in prominence across the world, playing a critical role in supporting individuals and households to manage risk and access opportunities. The same is true of Malawi, one of the world’s poorest countries and trapped in a low income and low productivity equilibrium. Around 70 percent of the population live below the international poverty line of US$ 1.90 per day. Many households remain highly vulnerable, especially the poorest, who are typically more exposed to risk and less able to access opportunities. 145. Malawi is at an important juncture in its efforts to strengthen its social protection system and has several building blocks in place. At a policy level, the second Malawi National Social Support Program (MNSSP II) 2018-2023 envisages the creation of a dynamic safety net system that is better positioned to respond to persistent poverty, recurrent shocks and a demographic dividend. The Social Cash Transfer Program (SCTP) demonstrates some of the strongest and most consistent positive impacts of any such program across sub-Saharan Africa. Malawi has also invested in the social protection system, including a social registry, e-payments and grievance redress mechanisms. 146. The social protection sector is facing an array of challenges. Almost the entire safety net system is financed by external development partners. Its future financing remains opaque, heightening ongoing concerns about accountability, oversight and sustainability. The Government has made little progress in addressing fiscal space challenges, which could help reduce crowding out by programs such as the Farm Input Subsidy Scheme (FISP) and increase the Government’s fiscal contributions to social protection. In addition, extreme weather and other types of climate risks have been rising in recent years. Demographic predictions that Malawi’s population could double in two decades underscore that social safety net programs will need to evolve to meet the needs of current and future generations. 147. This review seeks to contribute through a detailed analysis of social protection expenditure in Malawi. First, it outlines the objective, scope and methodology. It then describes the economic and poverty context that underpins social protection discussions in Malawi. It subsequently studies expenditure and financing trends from 2011 to 2019 across the entire sector and within each social protection pillar. Then it considers sector performance issues by unpacking themes such as coverage, targeting, effectiveness, cost of delivery, generosity of benefits, evidence and institutional capacity. Finally, it concludes with recommendations. Objective, Methodology and Scope 148. The review is intended to generate evidence and analysis on the efficiency, effectiveness and sustainability of social protection expenditures. The PER is intended to inform the future design and reform priorities of social protection interventions in Malawi, especially social safety net programming. The PER is anchored around two key objectives. First, to assess the financing structure of social protection programs. Second, to set out a range of options to improve the effectiveness and efficiency of spending and fiscal sustainability of the social protection system. By identifying gaps, challenges and opportunities for reform, the PER aims to contribute to the policy dialogue that facilitates more informed decision-making and strategic planning by the Government and donors. The review analyzes data covering the 2011-12 to 2019-20 (2011-19) period. 149. In this review, social protection includes four main pillars: social assistance (social safety nets), social insurance, labor markets and subsidies. These are the categories used by the Atlas of Social Protection: Indicators of Resilience and Equity (ASPIRE). In the context of Malawi, these pillars are: (i) Social assistance or social safety nets (SSN), which comprise an array of non-contributory programs, primarily for immediate and medium-term consumption support through cash transfers or in-kind, and include the SCTP, the School Meals Program (SMP), the Food for Assets (FFA) program, and the Malawi Social Action Fund (MASAF) Public Works Program (PWP); (ii) Social insurance to protect workers and dependents against exposure to risks, such as the National Pension System (NPS) and the Civil Service Pension Scheme (CSPS). (iii) Labor market programs to improve employability and create labor market linkages, such as the Technical and Vocational Education and Training (TEVET) and Village Savings and Loans programs (VSL); (iv) General subsidies for varying purposes including improved agricultural productivity and access to housing, such as the Farm Input Subsidy Program (FISP) and the Decent and Affordable Housing Subsidy Program (DAHSP). 150. Humanitarian assistance to address food insecurity is also included in this review although it is not directly part of the social protection ecosystem. Due to the close overlap with SSNs in Malawi, humanitarian assistance which seeks to improve food security is a useful comparator to contextualize social protection trends. While a wide range of humanitarian assistance has been provided over recent years, the review only includes the short-term consumption support provided to households in response to assessments of food insecurity conducted by the Malawi Vulnerability Assessment Committee (MVAC). 151. The review analyzes 10 core social protection programs implemented from 2011 to 2019 . Despite gaps in coverage and expenditure data, these programs reflect the core interventions and mix of social protection programs in Malawi. Annexes 1 and 2 provide a breakdown of program characteristics, detailing program objectives, targeting details and implementation arrangements. 152. These programs share a number of important characteristics. They are mature and operate at scale, clearly underpinned by national policy frameworks. The SSN programs, for example, fall within the policy priorities of the Malawi National Social Support Policy (NSSP) and the thematic priorities of the MNSSP II91. The featured programs are federally-managed and predominantly implemented by national agencies—often with financial support from international bodies. Most programs have been operational for a significant number of years, although some have been more recently established by development partners in collaboration with the Government. 153. This analysis excludes social protection interventions and other programs managed by NGOs. Programs supported by non-state actors are scattered throughout the country, are typically small in investment and coverage, but have an important aggregate impact. Substantial NGO programs include the considerable financing received by Mary’s Meals to implement the SMP, although requests for this data were mostly not met. The review’s exclusion of programs by NGOs and other non-state actors has likely resulted in an underestimation of overall social protection financing and coverage in Malawi. There is also no data on the coverage of humanitarian assistance between 2017 and 2019. 91 Government policy documents, including the NSSP and MNSSP II, generally refer to the social protection sector as “Social Support”. 154. This review analyzes data from a number of administrative sources and surveys. Financing and expenditure data has been taken from program-specific administrative sources. 92 Sector performance data has been drawn primarily from the IHS from 2010/11, 2013 and 2016/17. This survey data has been coupled with evaluative evidence when available, and supplementary administrative data where applicable. Using these data sources, the review applies an analysis of coverage, targeting effectiveness, level of benefits and impact.93,94 Key Risks and Vulnerabilities, and the Role of Social Protection 155. Social protection interventions have a strong potential to promote equity, enhance resilience and strengthen longer-term economic and social opportunities in Malawi. Malawi, the sixth poorest country in the world, has recently witnessed both a deterioration in growth and an increase in poverty. GDP growth has slowed since 2011, when it was about 7 percent, to 3.5 percent in 2018 (World Development Indicators). Real GDP per capita grew at an average 1.5 percent between 1995 and 2015, well below the average of 2.7 percent in non-resource-rich Sub-Saharan African economies during the last 20 years (World Bank 2018b). 156. The persistence of poverty highlights the need for safety nets as programs of last resort. The proportion of the population living below the national poverty line95 has remained largely constant since 2004, and was measured at 51.5 percent in 2016. This trend is at odds with those of many countries in the region, although extreme poverty in Malawi (at the national poverty line) decreased by about 4.5 percentage points between 2011 and 2016 to reach 20 percent. There are sizeable differences in poverty levels between urban centers and the rural areas where nearly 95 percent of the poor live. Some underlying factors for the persistence of poverty include low educational and health outcomes, and limited diversification of income sources. 157. Safety net investments demonstrate strong potential to realize important human capital outcomes. Malawi has made progress recently in reducing child mortality and fertility rates while increasing primary school enrollment. However, overall progress remains limited on key aspects of human capital development. With a 2017 HCI of 0.41, a child born today in Malawi will only be 41 percent as productive compared to a situation where the child had access to complete education and full health. The low levels are due to factors including the large burden of stunting and fertility, low enrollment and poor quality and access to education, and inadequate domestic financing for human capital. Malawi has one of the highest rates of adolescent fertility in the world, with 135 births per 1,000 women aged 15- 92 These were made available by the Government including the National Local Governance and Finance Committee (NLGFC), the former Local Development Fund, Ministry of Finance (MoF), Ministry of Agriculture, (MoA) and the Ministry of Gender, Children, Disability and Social Welfare (MoGCDSW); and development partners primarily the United Nations Children’s Fund (UNICEF), the World Food Programme (WFP), the European Union (EU), Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ), and international NGOs, whenever available. 93 The performance analysis uses ASPIRE methodology; indicators are generated with World Bank ADePT software. See: www.worldbank.org/ASPIRE 94 The most recent IHS survey (IHS4) includes a module that captures information on SSN programs such as free maize and other food, cash transfers from the Government and others, public works, school feeding, the Targeted Nutrition Programme (TNP) and Supplementary Feeding for Malnourished Children. Even though not all questions in this module refer to a specific program, robust assumptions can be made about what the module is referring to. It is assumed that IHS questions about whether households receive free maize and other free food capture the humanitarian assistance provided on the basis of MVAC assessments, and this data is therefore categorized as humanitarian assistance. Likewise, the IHS question about receipt of direct cash transfers from the Government is assumed to capture the SCTP (which is referred to as Mtukula Pakhomo). 95 The poverty line in 2004/05 IHS was MWK 37,002 per year per person, being adjusted for change in cost of living in 2010/11 and 2016/17 193. Early marriage is widespread, and women comprise 60 percent of those living with HIV/AIDS. The MNSSP II provides a strong entry point for building robust links between social protection and other human development sectors, especially education, health, and nutrition. 158. There are significant gender gaps in Malawi. Poverty headcount rates of female-headed households are around 5 percentage points higher than for male-headed households. Female-headed households tend to be poorer and have less endowments (own smaller cropland, less profitable and informal businesses, limited assets and access to infrastructure and basic services) than male-headed households. Lack of access to improved water and sanitation facilities also affects female school attendance, given that adolescent girls are disproportionally involved in collecting water. Narrowing the gender gap could have a profound effect on agricultural productivity, human capital development and resilience to shocks. 159. Resilience is an increasingly important issue, given Malawi’s vulnerability to shocks throughout the seasonal cycle. Poverty fluctuates substantially across the year, both in urban and rural areas, driven largely by the harvest season which runs from April to June (World Bank 2018a). Poverty rates are lowest during the harvest season and increase thereafter.96 160. Malawi has a history of large-scale weather shocks which increase the vulnerability of households and may force many into poverty either permanently or temporarily. Accelerated by high population growth and environmental degradation, weather shocks in Malawi are expected to worsen over time. In comparison with almost every other country in the region, Malawi faced the highest sensitivity to extreme dry events from 1980 to 2014 (Hallegate, 2016). One reason for this vulnerability is the high dependence on its staple crop, maize, for both production and consumption. 97 Flood and drought events in recent years, and most recently Cyclone Idai, have exposed such vulnerabilities. 98 Poor harvests are associated with falls in GDP per capita (Lea & Hanmer, 2009). Severe weather conditions further affect food prices and agricultural outputs, with significant impacts on poverty. Temporal shocks have the potential to keep chronically poor households locked in poverty and force a further two in every five households into poverty (Dang and Dabalen, 2017). Between 2010 and 2013, a third of Malawians fell into or moved out of poverty, and a fourth remained chronically poor. 161. Rapid population growth is leading to an unprecedented increase in Malawi’s young population, which is strengthening the importance of the social protection sector, especially with regard to future labor and social insurance demands. Like many of its neighbors, Malawi’s high fertility and low life expectancy have resulted in a large youth population (World Bank, 2017). The country’s population is expected to double in approximately two decades from 17.2 million in 2015 to 34.4 million in 2038 (Dabalen et al 2017). There is an opportunity for Malawi to take advantage of its demographic dividend, yet the country faces an enormous challenge in creating enough jobs for its rapidly growing working age population (MoF, 2016b). Overall unemployment is already high at 21 percent in 2013, with rates of approximately 26 percent for women and 14 percent for men (ILO, 2017). 96 This was true both in 2010-11 and 2016-17. There were, however, some differences between 2010-11 and 2016-17, and across rural and urban areas. First, poverty rates in those quarters outside the harvest season were higher in 2016-17 than in 2010-11, perhaps due to the damaging floods and drought in 2015 and 2016. Second, urban areas experienced a decline in poverty after the second quarter of 2016-17, whereas there was a consistent increase in rural areas. 97 Maize accounts for roughly 90 percent of all land under cereal production and 54 percent of caloric intake by households. 98 Severe floods in late 2014 and early 2015 affected over one million people and displaced 200,000 (UNDAC, 2015) causing damage estimated at US$ 335 million, or 5.2 percent of GDP (Government Post-Disaster Needs Assessment, 2015). A subsequent drought led to food shortages for an estimated 2.8 million people (WFP, 2015), and an El Nino weather event in 2015-16 led to crop failures that left over 40 percent of the population (6.5 million people) food insecure and in need of humanitarian assistance (Government, 2016a). Cyclone Idai affected an estimated 975,000 in March 2019, with the total needs for recovery and reconstruction estimated at US$ 370.5 million (Government, 2019a). Youth unemployment (people aged 15–34) is similarly high at 23 percent (as of 2013). Providing access to income-generating opportunities can contribute to more productive citizens by ensuring greater economic inclusion, especially in an environment where the pace of job creation, at an annual growth rate of 1.5 percent, has been slower than annual population growth at 3 percent. 162. High structural and temporary poverty combined with projected future population growth have contributed to evolving social protection systems. The MGDS 2006-2011 gave rise to some of Malawi’s largest social protection programs, including SCTP, FFA, SMP and FISP (ILO, 2017). Since 2012, the enabling environment for social protection has been strengthened through various policy frameworks. The National Social Support Policy (NSSP) provides an overarching vision and framework for social protection and was made operational by the MNSSP, which ran from 2012-2016 and sought to unite the multitude of social protection programs under one coherent framework. It also laid the foundation for a national social protection floor, in line with ILO Social Protection Floors Recommendation, 2012 No. 202 (ILO 2016a). 163. The MNSSP II sets an ambitious reform agenda to improve Malawi’s social safety net system through to 2023. The MNSSP II focuses on moving from individual social protection programs toward a system of interventions with increased coherence, integration and harmonization. It outlines five thematic areas of support, including: (1) Consumption Support aimed at providing transfers to extremely poor households allowing them to meet basic needs; (2) Resilient Livelihoods, which seeks to address multiple causes of poverty by promoting a mix of programs meeting different needs throughout the life cycle; (3) Shock Sensitive Social Protection, which focuses on empowering people to prepare for and respond to shocks and crises by re-orienting the focus of social protection beyond short term consumption support; (4) Linkages between social protection and other programs, which recognizes that poverty is multidimensional and social protection alone cannot fully address all aspects of poverty; and (5) Social protection systems strengthening focusing on coherence and harmonization, including coordination committees at district and community levels and integrated implementation mechanisms. Expenditure Trends and Sustainability Overview of total social protection expenditure 164. An average of 4.9 percent of GDP was spent on social protection programs between 2011 and 2019, driven by expenditures on subsidies and pensions. As shown in Figure 57, spending was highest in 2012 and 2013, primarily due to an increase in the FISP and SSN program expenditure, but fell overall in 2014 as a result of reforms to the FISP. Since 2014-15, further reforms to the FISP including reallocation of funding to SSN programs has been offset by higher spending and reforms on pensions, which has thereby maintained overall spending levels. Figure 57: Expenditure on Social Protection Program by Type (% of GDP), 2011 – 2019 165. Government spending on social protection is primarily through subsidies and civil service pensions, while funding for social safety nets relies heavily on donor partners. Since the reforms to the FISP in recent years, the Government’s allocations to subsidies have substantially reduced. However, its funding for pensions is expected to balloon further due to the costly formula that creates significant liabilities for the Government. Projected future increases in pension liabilities could further constrain fiscal space for social protection. On the other hand, minimal levels of the government’s financing for SSNs have generated concerns about the sustainability of interventions targeting the poorest. Social Safety Nets Expenditure 166. The average expenditure on SSNs in Malawi was 0.8 percent of GDP between 2011 and 2019, a strikingly low amount compared to an average of 1.2 percent across the Sub-Saharan Africa region. Expenditure peaked at 1.3 percent in 2016, as a result of flood and drought crises. Malawi’s average expenditure on SSNs is just over half the 1.4 percent spent on SSNs by low-income countries in Sub-Saharan Africa (Beegle et al, 2017). Expenditures on SSNs comprise only a small fraction of overall social protection spending in Malawi (see Figure 57). Between 2011 and 2019, SSN spending comprised an average of 19.1 percent of total social protection expenditures. 167. Social safety net expenditures are concentrated in a small number of key programs. Compared to other countries in the region, Malawi implements a narrow range of SSN interventions that target the most vulnerable households. In other low-income African countries, the average is 18 programs (Beegle et al, 2018). In Malawi, the key SSN programs have included three Government-led programs: SCTP, MASAF PWP (concluded in 2019), and Nutrition and Access to Primary Education (NAPE). There are two further interventions which are led by development partners: Food for Assets (FFA) Public Works Program (PWP), led by WFP; and the School Meals Program (SMP), led primarily by Mary’s Meals and the WFP. Although a small number of programs may lower overall administrative costs and reduce the number of channels through which to conduct oversight, the design of individual programs may limit the reach of the system as a whole. For example, the SCTP provides cash transfers for labor-constrained ultra-poor households. Taken alone, it would not be able to address vulnerabilities associated with, say, information asymmetries in the job market. Figure 58: Spending (% of GDP) Trends by Social Safety Net Program (2011 – 2019) 1.4% 1.2% 1.0% 0.8% 0.6% 0.4% 0.2% 0.0% 2011 2012 2013 2014 2015 2016 2017 2018 2019 NAPE MASAF SCTP FFA SMP Source: Administrative data 168. Historically, school meals and public works interventions have comprised the largest share of social safety nets expenditure, although this has declined. As Figure 58 shows, funding for the SMP grew from 0.2 to 0.3 percent of GDP between 2011 and 2017, before declining to 0.1 percent in 2019. The MASAF PWP has also experienced substantial expenditure fluctuations over the same period, ranging from an 0.04 percent of GDP (US$ 5 million) in 2012 to upwards of 0.4 percent (US$ 22.3 million) in 2016 before declining to 0.02 percent (US$ 1.34 million) in 2019 when it was closed after facing a number of challenges.99,100 In addition, the FFA PWP was among the lowest funded SSN programs between 2011 and 2015, at less than US$ 2 million annually. Due to the effects of weather shocks, additional funding scaled up FFA PWP in 2015-16, increasing to 0.3 percent of GDP in 2016, which was maintained until 2018. Funding for FFA then fell to 0.2 percent of GDP in 2019, due in part to the replacement of in-kind support with cash. 169. The mix of safety net programs is currently in transition, with an expansion of cash transfers in recent years. A more balanced distribution across other safety net programs began to emerge in 2016, likely due to the expansion of SSN programs responding to floods and droughts, as well as the progressive rollout of the SCTP to full national coverage. Expenditure on the SCTP has been increased from 0.04 to 0.42 percent of GDP between 2011 and 2019. As of February 2020, the SCTP was operating at the national level, covering 291,315 households and a total of 1,239,809 beneficiaries. Notably, expenditure on SCTP has overtaken that of the SMP and far outweighs the former MASAF PWP. 99 This funding variation for MASAF PWP is likely due to three factors. First, in late 2012, the Government scaled up its public works activities with US$ 2.5 million of additional funding (UNU-WIDER 2017). Second, policy changes to the MASAF PWP in 2015 led to reductions in the number of beneficiaries (ILO 2016). Third, the program was scaled down significantly in 2019 and then closed. 90 percent of MASAF PWP funding came from the World Bank and the EU. The Government provided the remaining 10 percent. 100 An impact evaluation conducted in 2012-13 found no significant improvements in the food security of beneficiaries and highlighted several challenges related to the program’s design, such as poor targeting, significant rationing, the relatively small size of the transfers, and the infrequency of payments (Beegle et al, 2015). A series of design modifications were introduced, culminating in a shift toward a smaller public works pilot focused on addressing fiduciary challenges, asset creation and improved program timing and duration. In 2019, SCTP expenditure was US$ 34.5 million, compared to SMP (US$ 7.4 million) and MASAF PWP (US$ 1.3 million). 170. Humanitarian assistance spending in Malawi is substantially higher than comparator countries, even in years when there were no extreme weather events. On average, 1.3 percent of GDP was spent yearly on humanitarian assistance in response to food insecurity from 2011-2019. The effects of the weather events were particularly acute in 2015–2016, provoking outlays to address food insecurity of 1.8 percent (2015) and 5.8 percent (2016) of GDP. In 2016, humanitarian assistance coverage peaked at 34.5 percent of the population. However, even in years where there were no extreme weather events, such as in 2012-2014, Malawi still spent of 0.6 percent of GDP on humanitarian assistance to address food insecurity. This is significantly more than average levels in Mozambique (0.15 percent), Rwanda (0.28 percent), Tanzania (0.09 percent), Uganda (0.38 percent) and Zambia (0.04 percent) (Beegle et al, 2017). 171. Regular large-scale outlays of humanitarian assistance suggest there is significant room to strengthen the shock sensitivity of Malawi’s social safety net and enhance its capacity to build resilience. At present, SSNs do not adequately address the predictable nature of seasonal food insecurity that can be drastically amplified by climate stressors (Holmes et al, 2017). Safety nets and humanitarian aid largely operate separately from each other, despite having many overlapping functions. Safety nets need to be designed so that they can be scaled up during temporary crises to help households be resilient and to reduce dependence on humanitarian aid. There have been some positive developments in this regard over recent years, however, and at the time of writing, Malawi’s social protection system was being prepared to help respond to the COVID 19 pandemic. This is elaborated on further in Box 5. Box 5: Responsiveness of Social Protection to COVID 19 in Malawi There is an increase in the number of social protection interventions globally, mainly in the form of cash transfers, in response to COVID 19: As of end May 2020, a total of 190 countries and territories worldwide have planned, introduced or adapted 937 social protection measures in response to COVID-19 101. Cash transfer programs account for half of the social assistance measures by governments in response to COVID 19. Cash transfer programmes dominate the interventions mainly due to their ease of adaptability, markets functionality, delivery cost efficiency and fungibility of cash. The Malawi Social Protection response to COVID 19 is fully aligned with the country’s social support policy framework: Malawi’s COVID 19 social protection response is piggybacking on the SCTP by increasing coverage to target new beneficiaries among poor urban households and to provide cash top ups to existing social cash transfer program beneficiaries, in line with the Shock Sensitive Social Protection pillar of the second MNSSP. The following are some of the key considerations for designing and implementing the COVID 19 response, especially the urban component: Utilizing existing social protection delivery systems such as the social registry (Unified Beneficiary Registry (UBR)) and SCTP Management Information System (MIS) can facilitate a quick and timely response. Rapid adaptation of both the UBR and SCTP MIS for prompt registration and enrolment of beneficiaries could be critical for a swift response. Targeting of new urban geographical areas could include an emergency roll-out of an abridged version of the social registry’s data collection tool in urban hotspots. The utility of other data sources such as Integrated Household Survey and satellite data can also be explored in order to enhance the objectivity and transparency of the identification and targeting process. Adhering to social distancing restrictions requires rapid deployment of a digital payment method for benefits. Safety net programs mostly operate through physical delivery of cash, which would present 101 Gentilini, U et al. “Draft Living paper” v9 (May 15, 2020) Social Protection and Jobs Responses to COVID 19; A Real Time Review of Country Measures. numerous challenges in the COVID-19 environment.102 The current crisis presents an opportunity to expedite the introduction of electronic payments. Partnerships with mobile money operators could be explored as a swift and more secure means of delivering safety nets benefits, particularly in urban areas. Building in strong accountability mechanisms across all key implementation processes would help offset the potential for administrative inefficiencies, political capture, fraud and corruption. The rushed nature of emergency responses is an opportunity for political capture, fraud, and corruption, as well as implementation inefficiencies, including high administration costs. It is therefore critical that all key implementation processes, including registration, targeting, enrolment and payment, should be managed transparently, with strong accountability mechanisms built into each key process. Digitization of these key processes through the UBR and SCTP MIS could be one way of minimizing these potential risks, complemented by remote monitoring and grievance redress mechanisms. Defining optimal benefit levels and duration of the program are paramount for an effective response: The Government has already set the urban response benefit level at MWK 35,000 per household per month. Considering that most urban poor are engaged in some form of employment paying at least the minimum wage – for example vending, or ganyu/piecework – the urban transfer level is logical as it is aligned to the minimum wage. Equally important is the need to define the optimal implementation period consistent with the COVID 19 in-country dynamics. Securing both national and external financing is critical for the response: Financial commitments to the social protection response are estimated at more than MWK 40 Billion from both Government and development partners. Securing both national and external financing is a good practice to ensure that the response is adequately financed, and the domestic contribution also shows strong Government commitment and ownership. Continuous learning and feedback are essential in emergency interventions: The COVID 19 social protection response will be reviewed monthly, with a larger and more significant review after 3 months of implementation to draw lessons and determine the efficacy of the overall program, as well as to inform medium-to-long term urban safety net design and overall adaptability of social safety net systems in Malawi. Coverage 172. Social safety net programs in Malawi are estimated to cover 24 percent of the population in 2019-20, compared to an average of 10 percent across Sub-Saharan Africa. This is among the highest coverage rates in the region relative to poverty levels. High levels of coverage do not necessarily translate to high impact, however, as levels of benefits can be inadequate, with the amounts transferred insufficient to make a meaningful impact on household consumption and other objectives. For example, the SMP is a significant contributor to high coverage in Malawi, but its impact on education or nutrition is unclear. Furthermore, SSN coverage still excludes a sizeable portion of the population. The coverage of Malawi’s SSNs is less than that of other priority poverty reduction and relief programs, namely the FISP and humanitarian assistance (discussed below). 173. From a life cycle approach, there are gaps in the coverage and objectives of safety nets, particularly for children. A very small proportion of poor Malawian children benefit from cash transfer programs. Although the SCTP is now operational in every district, only 12 percent of children living in multidimensional poverty are in households that are benefiting from the SCTP. As Figure 59 shows, there are prominent shortcomings in coverage during early childhood. Twelve percent of SCTP beneficiaries are children under 5 – a highly vulnerable period characterized by dependence on others and important development stages. Malawi suffers from high rates of malnutrition, and the first 1,000 days of life 102 As of 2015, only 8 percent of Malawians were using mobile money platforms, mostly in urban areas (at 27 percent vs 5 percent in rural areas). However, the prevalence and usage of mobile-money platforms for both rural and urban poor is expected to have grown over the past few years. comprise the most critical window for physical growth, brain and psychosocial development (World Bank, 2019b). Yet this cohort of the population is not specifically targeted through the SCTP and is in fact substantially under-represented in comparison to national averages. Children aged under two make up just 1.9 percent103 of all children (aged 0-17) in SCTP beneficiary households, whereas children aged under two make up an average of 11.6 percent of all children in the broader population (Government of Malawi, 2019b). There are considerable opportunities to link safety net investments to early childhood interventions. A strategy is in the early stages of adoption supported by the World Bank financed Investing in Early Years for Growth and Productivity Project. The government is also leading efforts to develop guidance on nutrition sensitivity for implementers of MNSSP II. Figure 59: The Malawi social safety net system through the life cycle lens * Emerging ** Latent *** Moderate **** Strong 174. The School Meals Program (SMP) is Malawi’s largest SSN program in terms of coverage. It provides a meal to all children in targeted schools (GoM 2016b) in food insecure districts, particularly those that have low enrolment and high dropout and repetition rates. The SMP covered 12.6 percent of the child population in 2019, an increase from 10.1 percent in 2011. During the lean season, WFP provides a monthly ration of food to girls and orphan boys in grades 5 to 8 on the condition that they attend 80 percent of school days (GoM 2016b). NAPE is one of the smallest SSN programs, providing school meals for students at select primary schools in seven districts, and covering 1.3 percent of the child population in 2019. 175. Coverage of the SCTP has quadrupled since 2011 and now reaches eligible ultra-poor households in every district. The SCTP started as a pilot project in Mchinji district in 2006. It provides the poorest 10 percent of labor-constrained households with cash grants averaging MWK 7,000/month. Favorable results from the pilot project attracted additional funding, enabling coverage to increase nine- fold in seven years, from 0.7 to 6.5 percent of the population from 2011 to 2018. With clear policy guidance from the MNSSP II, the SCTP has undergone a rapid scale-up to full national coverage. Children 103 Source: SCTP MIS as at February 2020. and young adults form the single largest group of SCTP beneficiaries, directly or through household members. 176. Coverage of public works programs have declined as MASAF has scaled down, although the FFA PWP has expanded. MASAF PWP was previously one of the largest SSN programs, but it has recently been scaled down. MASAF PWP targeted ultra-poor households without labor constraints, and was one of the oldest safety net programs, having commenced in 1996. The program operated in all 28 districts, and there was a significant upward trend in coverage between 2011 and 2016 doubling from 7.4 to 16.8 percent. However, by the time it was scaled down in 2019, it reached just 0.1 percent of the total population. On the other hand, FFA PWP coverage has undergone substantial expansion over the period covered by this review. It reached no more than 0.1 percent of the population in 2011, and by 2018 it had expanded to 4.2 percent, before reducing slightly to 3.7 percent in 2019. FFA PWP is a food security- focused public works initiative intended for food insecure individuals. Financing and Sustainability 177. Malawi faces considerable financing needs for social safety nets, raising questions about sustainability. Even under conservative estimates, financing is not sustainable and gaps are likely to remain with Malawi’s SSN programs largely financed by development partners. Domestic financing accounted for about 6 percent of total safety net expenditure in 2019. This raises stark questions about sustainability and affordability. 178. While Malawi’s unsustainable financing of safety nets is not unique, a number of other countries in the region have managed to self-finance their SSN programs, including Kenya, Ghana, Lesotho, Mozambique and Tanzania. Lessons from these countries suggest that it is possible to (i) increase the level and strengthen the sustainability of financial resources; (ii) identify the most appropriate mix of domestic, foreign, public, and non-public funding sources; and (iii) deploy a flexible financing strategy to respond to shocks and crises (Beegle et al, 2018). Given the current landscape and evidence, this will require long-term incremental planning in Malawi, but there are several steps that can be taken in the immediate term. This is elaborated on further in Section 4. Figure 60: Financing Sources for Social Safety Nets across Africa Source: Beegle et al (2018) Social Insurance Expenditure 179. Malawi’s allocation to social insurance programs have increased from 2.3 to 3.3 percent of GDP between 2016 and 2019, despite only covering 2 percent of the population. Social insurance consists of a bifurcated system comprising the National Pension System (NPS) and the Civil Service Pension Scheme (CSPS). This duality is characteristic of old age pension systems in the region—33 of 44 countries have separate systems for the public and private sector (Abels and Guven, 2016). While the NPS is for the private sector and designed as a funded defined contribution 104 scheme, CSPS is, as the name suggests, for the public sector and designed as a pay-as-you-go105 defined benefit106 scheme. There are important synergies to be had by fully or partly integrating the two systems (Abels and Guven, 2016), and some recent reforms have involved the transition of some CSPS beneficiaries to the NPS contributory scheme. Expenditure on pensions has increased as a share of GDP, divided across both schemes, with particular increases between 2017 and 2018 (see Figure 61). Figure 61: Expenditure on Social Insurance as % of GDP (2011 – 2019)* 3.5% 3.0% 2.5% 1.5% 1.4% 2.0% 1.1% 1.0% 1.1% 1.5% 0.8% 0.8% 0.7% 1.0% 1.9% 1.9% 1.2% 1.2% 1.3% 1.2% 1.4% 0.5% 1.0% 0.8% 0.0% 0.0% 2011 2012 2013 2014 2015 2016 2017 2018 2019 National Pension System (NPS) Civil Service Pension Scheme (CSPS) Source: Administrative data * Data for National Pension System is not included for 2017-19 as it was not provided by Government. 180. Social insurance expenditure in Malawi is slightly higher than the regional average of 1.6 percent of GDP (Figure 62). This level is comparable to countries such as Namibia, Tanzania, Mauritius, Togo and Madagascar (Abels and Guven, 2016). Notably, expenditure on the NPS is comparatively higher than other countries in the region, despite covering a small percentage of the population. 104 A defined contribution pension plan is a pension plan in which the periodic contribution is prescribed, and the benefit depends on the contribution from the investment return. It can be fully funded or notional and non-financial. 105 This is a method of pension system financing whereby current outlays on pension benefits are paid out of current revenues from an earmarked tax, often a payroll tax. 106 This is a pension plan with a guarantee by the insurer or pension agency that a benefit based on a prescribed formula will be paid; it can be fully funded or unfunded and notional. Figure 62: Pensions expenditures in civil service pension scheme, % GDP 3.5% 3.0% 2.5% 2.0% 1.5% 1.0% 0.5% 0.0% Source: Abels and Guven, 2016 181. Mandatory pensions through the Pension Act 2011 and demographic trends are likely to put upward pressure on pension expenditures. The Pension Act 2011 mandates every employer to whom the Act applies to provide a pension to all employees. In addition, the size of the population above 60 years is projected to quadruple from 793,000 in 2017 (4.3 percent of total population) to 3.2 million in 2050 (7.7 percent), under the medium variant of projections (UNDESA 2017a, 2017b). The increased coverage resulting from these two factors will significantly increase pension expenditure. An expansion of coverage may eventually lead to a greater share of the poor receiving pension benefits. However, 83 percent of the workforce, mostly poor, is employed in the informal sector (ILO 2018), which would pose a substantial impediment to their entry into the predominantly ‘formal’ pension system. Coverage 182. Despite the huge costs, only a tiny proportion of Malawians are covered by pensions. An estimated 256,000 people are covered by the CSPS and 77,000 by the NPS, including all who are enrolled in either program, even those not yet eligible for payouts. This is equivalent to about 2 percent of the total population or about 44 percent of the population above 60 years, (but potentially a substantially lower share if calculations do not include contributors younger than 60 years old). This is higher than trends observed across Africa, where coverage of pension recipients is less than 10 percent of the population above 60 years for the majority of countries (Abels and Guven, 2016). Moreover, the majority of these pension beneficiaries belong to the top two income quintiles—civil servants and formal sector employees. Financing and Sustainability 183. Pension schemes are financed by both the government and the private sector, the latter seemingly providing some temporary fiscal relief. The NPS is funded through contributions from employers and employees, thus currently putting little pressure on the government’s fiscal position. In the past, the CSPS has been financed directly from the Government budget, running the risk of creating unsustainable fiscal pressures. However, in recent years, a contributory mechanism has been introduced to the CSPS, which will help reduce the fiscal burden in the future. 184. Future pension expenditures are likely to be unsustainable, particularly in the face of demand for expansion. As noted above, demographic trends indicate an increasing old age population. Therefore, scaling up pension coverage through both the CSPS and NPS in response to these trends is likely to pose a large fiscal burden on the government. While NPS expansion to the formal sector is unlikely to result in additional burden on the government, expansion to the poorer informal sector would likely require government contributions. Given the size of the informal economy, the government would be subject to fiscal pressure to provide a safety net for old age citizens. If a proportion of the informal economy is absorbed into the formal economy in future, this would provide some relief to the government through higher employer and employee contributions to the NPS. 185. Any expansion of the CSPS potentially has a direct impact on Government finances and on other budget priorities. As greater pension liabilities for the government will increase the fiscal burden, regardless of the introduction of a contributory component, this can potentially crowd out spending on key areas which are essential for human development and poverty reduction such as health, education and SSNs. The generosity of the CSPS also has broader implications for the private sector as it may guide expectations for what private employers are expected to provide. 186. Some caution should be exercised in comparing the role and need for non-contributory safety nets versus contributory social insurance for formal workers. While expenditure on pensions is currently the largest form of social protection spending in Malawi, this largely reflects financing commitments to civil servants in the public sector, where pension benefits may be considered an important deferred salary benefit. Labor Market Programs Expenditure 187. Expenditure on labor market programs is small compared to other social protection programs. Labor market programs are extremely small and/or fragmented and are generally implemented by NGOs. Due to difficulties in aggregating information on beneficiaries and costs, there are data gaps in labor market programs. As a result, the review acknowledges that it is unable to report figures for most labor market programs. The review was able to identify that US$ 3.4 million was spent on TEVET in 2016, about 0.07 percent of GDP, and that a 2015 mapping exercise by MoF and the charity CARE Malawi found VSL programs were implemented by 67 organizations constituting 37,461 groups and 610,596 members (3.4 percent of Malawians) (ILO 2016c). Coverage 188. Labor market programs are in the nascent stages of development in Malawi, covering a small share of the population and costing a small amount compared to other social protection programs. Prominent labor market programs include Village Savings and Loans (VSL), Microfinance, Technical and Vocational Education and Training (TEVET), the Malawi Innovation Challenge Fund, and the More Income and Employment in Rural Areas of Malawi (MIERA) program, though these are extremely small and/or fragmented and are generally implemented by NGOs. Financing and Sustainability 189. These programs rely heavily on development partner funding, making sustainability a concern. For instance, a prominent program with VSL components includes the Community Savings and Investment Promotion (COMSIP), funded by the World Bank. Similarly, TEVET is funded by the EU. In future, it would be important for the government to start exploring strategic and financial mechanisms to make a more substantial expansion into labor market interventions. This is particularly important as Malawi has a large demographic dividend to be realized in the future, and such market linkages may lead to increased productivity of household enterprises. General Subsidies Expenditure 190. The politically popular FISP has dominated social protection spending in recent years. The FISP, introduced in 2006 after food shortages in previous years, provides subsidized farm inputs (fertilizer and seeds) to smallholder farmers nationwide. The program’s primary objective is to improve smallholders’ income and food security, providing farmers with two vouchers that can be redeemed for two 50 kilogram bags of fertilizer at a base price. The subsidy equates to about two-thirds or more of the fertilizer market price (Arndt et al 2016). Farmers also receive a voucher for a 5-10 kilogram bag of hybrid maize seed or open pollinated varieties and a flexible voucher that can be redeemed for a free 1 kilogram bag of legumes or groundnut seeds (FAO 2017; IFPRI 2014). 191. Although expenditure on the FISP once comprised the highest share of all social protection programs, spending has been on a downward trend since peaking in 2012. The FISP absorbed 75 percent of the total agricultural budget in 2012, representing 60 percent of all social protection spending and 3.5 percent of GDP (Chinsinga, 2012). Programmatic inefficiencies and ballooning costs caused the program to overshoot its allotted budget by 41-105 percent (Chinsinga, 2012). Expenditure on the FISP averaged 1.7 percent of GDP and consumed over 40 percent of Malawi’s overall spending on social protection between 2011 and 2019. Recent reforms have seen spending fall to about 0.6 percent of GDP in 2019. However, with the introduction of the Affordable Inputs Program (AIP) in FY2020/21, this is expected to increase to 2.4 percent of GDP.107 Coverage 192. While close to half the country accessed the FISP in 2011 this declined to 25 percent of Malawians in 2016. Nevertheless, it continues to be the largest social protection program. As noted earlier, the reduction in coverage is in response to recent reforms undertaken by the Government to improve the functioning of agriculture markets and expand the SCTP. Key reforms included: (i) increasing the share of fertilizers sold through the private sector; (ii) reducing the level of subsidy to beneficiaries by introducing a fixed value coupon whereby the difference between the coupon and market price was paid by beneficiaries; and (iii) improving targeting of beneficiaries to benefit the poorest. Financing and sustainability 193. Unlike most other social protection programs, the FISP is Government-owned, funded and implemented. It has been primarily funded by the Government with some initial donor funding provided in the form of overall budgetary support (Chirwa, Matita and Dorwad, 2011). The Ministry of Agriculture, Irrigation and Water Development remains responsible for designing and implementing the program. 194. The FISP financing is undergoing review and may reduce over time. Reducing the fiscal burden of the FISP has been at the heart of a larger Government reform program to lay the foundation for a more resilient economy by improving agricultural productivity and is complemented by various pro- market policy actions and increased funding to expand the coverage of SSNs. The World Bank has been supporting this transition through both technical and budgetary support. The reforms are in keeping with international and national literature that suggest subsidies are not a preferred option for targeting the poorest households (Coady et al, 2015; Pace et al, 2017). The policy implication is that promoting 107 Introduced in the FY2020/21 budget, the AIP increases beneficiaries from 900,000 to 4.2 million—thereby effectively reaching all farming households. It also reduces the price of fertilizer paid by smallholder farmers to MWK 4495 and has increased the subsidy component payable by the Government. The technical analysis for this PER was completed before the introduction of the AIP, therefore it is not fully assessed within the scope of this PER. simultaneous reductions in expenditure on the FISP while investing in pro-poor policies such as cash transfers and agricultural extension services is likely to achieve a more pro-poor financing strategy. Recent World Bank analysis estimated that such reallocations would increase income levels among the bottom population decile by 9 percent (World Bank 2018c), potentially reducing benefits leakage to higher income deciles, which was as high as 50 percent under the FISP (Kilic, Whitney & Winters, 2013). Sector Coverage Total sector coverage108 195. The share of population covered by various social protection programs varies significantly , with the FISP decreasing over time but remaining the most prominent. FISP coverage peaked in 2011, reaching 46 percent of the population, and fell by almost half since 2017. Other SSN programs show some variability but cover a much lower share of the population. Despite high levels of expenditure, social insurance covers a small share of the population at just 2 percent in total. This low figure is unsurprising given that social insurance programs only target civil servants or workers in the formal sector, who primarily reside in urban areas. Figure 63: Share of population covered by various social protection programs (2011-2019) 50% 40% 30% 20% 10% 0% NPS NAPE SMP SCT FFA PWP MASAF PWP DAH FISP 2011 2012 2013 2014 2015 2016 2017 2018 2019 196. Social protection coverage (including social insurance and SSNs) increased from 18.7 percent in 2010 to 32.9 percent in 2013 before decreasing to 26.5 percent in 2016/17. 109 Most SSN 108 Coverage measures the percentage of the population benefiting from social protection programs, humanitarian assistance and subsidies. This includes direct and indirect beneficiaries (for example individuals who live in a household where at least one other member is covered by a program). Targeting effectiveness analyses the distribution of program beneficiaries across the welfare distribution. The level of benefits measures the amount of the transfer as a percentage of total household expenditure. Indicators are likely to be an underestimate given that the IHS does not comprehensively capture the universe of social protection programs existing in Malawi. In addition, the value of benefits for some of the programs is not included in the IHS results, primarily because it is not easy to quantify (for instance in the case of the SMP). 109 Information for this analysis is largely drawn from the IHS, instead of the administrative data underlying other sections. Performance indicators provide important complementary information to administrative program data. The IHS is a nationally representative survey hence its findings are applicable to the national population. While administrative data may contain robust results on coverage, indicators relating to performance (for instance incidence of benefits by income quintiles) are less common in a context where robust monitoring and evaluation of programs is not feasible and/or regularly undertaken. There are nonetheless some gaps between IHS4 and administrative data in terms of total coverage of the population which highlights significant differences in how the data is collected. These are primarily observed for MASAF PWP and SCTP, and are explained in the subsequent text. programs increased moderately from 2010 to 2016/17, however, the reduction in recent years was primarily driven by the MASAF PWP. For MASAF PWP, coverage in the poorest quintile reduced from 14.3 percent in 2013 to 8.1 percent in 2016/17, which is likely due to a key change in the program’s participation rules.110 Notwithstanding the reduction in overall coverage to 26.5 percent in 2016/17, Malawi still has one of the highest coverage rates of social protection programs in Sub-Saharan Africa. Coverage in humanitarian assistance in response to food insecurity has also increased consistently over the years, from only 2.6 percent in 2010 to 26 percent in 2016/17 (see Figure 64). On the other hand, the coverage of the FISP has decreased from 47 percent in 2010, to 41 percent in 2011 and then saw a significant drop to 18 percent in 2016/17. Figure 64: Percent of Total Population Covered by Social Protection, Humanitarian Assistance and Subsidies (2010-2017) 50 47.4 41.1 40 32.9 30 26.5 22.3 18.7 18.3 20 11.5 10 2.6 0 Humanitarian assistance (MVAC) All social protection Subsidies (FISP) 2010/11 2013 2016/17 Source: ASPIRE, based on IHS 2010/11, 2013 and 2016/17. Notes: (1) Programs included per area: Social protection includes social insurance (contributory pensions) and SSN programs (Targeted Nutrition Program (TNP), supplemental feeding for malnourished children, school feeding, cash from the Government, cash from development partners, MASAF, public works other than MASAF, and scholarships for secondary and tertiary education). Humanitarian assistance includes: free maize and food other than maize. (2) Coverage is: (Number of individuals in the total population or poorest quintile who live in a household where at least one member receives the transfer)/(number of individuals in the total population). Therefore, it includes direct and indirect beneficiaries. 197. Most individual safety net programs cover less than 5 percent of Malawi’s total population, except for SMP and MASAF PWP. Figure 65 disaggregates coverage rates by each of the programs included in social protection (social insurance and SSN), humanitarian assistance and subsidies, as they are collected in the 2016/17 IHS4. The SMP and MASAF PWP, with coverage rates of 13.8 and 7.5 percent of the total population respectively, are driving the overall coverage of SSNs. The other SSN programs have very modest coverage rates, which are lower than 5 percent when measured as a share of both the poorest quintile and the total population. 198. Humanitarian assistance and the FISP have substantially higher coverage than any single SSN intervention, despite their high cost and weak targeting of the poor. Humanitarian assistance, which is measured on provision of free maize and free food other than maize, covers 16 and 21 percent of the total population, and 19 and 26 percent of the poorest 111 quintile, respectively. The FISP covers 18 percent of the total population and 17 percent of the poorest quintile. These statistics are illustrated in 110 Commencing in 2016, the same households needed to be enrolled across multiple program cycles in the same year, which reduced the number of unique beneficiaries from one year to the next. However, this adjustment was not captured in administrative data. The program continued to use the same accounting procedures of aggregating the total number of beneficiaries across different cycles in a given year, even though this resulted in many beneficiaries being counted more than once. 111 We note IHS data refers to a later time period than for the administrative data reported in section 3, hence the variability Figure 65. While the substantial coverage of humanitarian assistance is in response to the damaging effects of climate shocks between 2015 and 2016, that of the FISP is related to its historical role in cushioning smallholder farmers against food shortages. Though these figures underscore the need for support in the Malawian context, the substantial investment in humanitarian assistance and the FISP is crowding out the investment in better targeted and cheaper112 social protection programs. Figure 65: Percentage of Total Population and Poorest Quintile Covered by Social Protection, Humanitarian Assistance and Subsidies (2016/17) Free maize 26 FISP MVAC 21 Free food other than maize 19 16 FISP coupons 17 18 School feeding 19 14 MASAF 8 8 Cash from development partners 3 2 Cash from government SCTP 4 2 SSN Targeted Nutrition Program TNP 3 2 Public works other than MASAF 3 2 Scholarships -secondary & tertiary education 1 1 Supplemental feeding for malnourished children 1 0 Pensions 0 SI 1 0 5 10 15 20 25 30 Poorest quintile Total population Source: ASPIRE, based on Malawi's Fourth IHS 2016/17. Notes: (1) Coverage is: (Number of individuals in the total population or poorest quintile who live in a household where at least one member receives the transfer)/(Number of individuals in the total population). Therefore, it includes direct and indirect beneficiaries. (2) The poorest quintile is calculated using per capita pre- transfer consumption. Equity of Spending and Targeting Effectiveness 199. While SSN programs have broadly increased the equity of spending in recent years, this has fluctuated for humanitarian assistance and subsidies. SSN programs have more effectively targeted the poor in recent years. Beneficiaries from the poorest quintile has increased steadily from 19.4 percent in 2010/11, to 21.5 percent in 2013 and to 27 percent in 2016/17 (see Figure 66). By contrast, the effectiveness of targeting through humanitarian assistance and subsidies has fluctuated. As shown in Figure 66, there was a dip in the targeting performance of both programs between 2010 and 2013, although this improved in 2016/17. However, the performance of the FISP in 2016/17 remains worse than in 2010. With the caveat that humanitarian assistance, FISP and SSN programs have slightly different objectives and hence some variation in targeting objectives, both humanitarian assistance and the FISP do not perform as well as SSNs to target the poorest households. 112 The relatively lower administrative costs of social protection programs are discussed below. Figure 66: Percentage of Beneficiaries that belong to the Poorest Quintile - Social Protection, Humanitarian Assistance and Subsidies (2010-2017) 18.8 Subsidies (FISP) 17.1 19.9 . 24.3 Humanitarian assistance (MVAC) 19.5 22.7 27.1 Social safety nets 21.5 Protection 19.4 Social 9.1 Social insurance 11.9 3.7 0 5 10 15 20 25 30 2016/17 2013 2010/11 Source: ASPIRE, based on Malawi's IHS 2010/11, 2013 and 2016/17. Notes: (1) Programs included per area: Social protection includes social insurance (contributory pensions) and SSN programs (Targeted Nutrition Program TNP, supplemental feeding for malnourished children, school feeding, cash from Government, cash from development partners, MASAF, public works other than MASAF, and scholarships for secondary and tertiary education). Humanitarian assistance includes: free maize and food other than maize. (2) Beneficiaries' incidence is: (Number of direct and indirect beneficiaries [people who live in a household where at least one member receives the transfer] in a given quintile)/ (Total number of direct and indirect beneficiaries). 200. Within SSN programs, however, targeting effectiveness varies. As shown in Figure 67, the SCTP has one of the highest incidences of beneficiaries that are extremely poor, at 38.4 percent, with almost 70 percent of beneficiaries considered poor. MASAF PWP is less pro-poor than most SSN programs in Malawi, with only 54.5 percent of its beneficiaries classified poor or extremely poor. Scholarships for secondary and tertiary education are also not particularly pro-poor: 57.6 percent of scholarships beneficiaries are not deemed poor, while only 18 percent are extremely poor and 24.4 percent poor. However, poor people are also unlikely to be beneficiaries of a pension, with 74.8 of pension beneficiaries deemed not poor. This is unsurprising, as poor people are less likely to work in the formal sector and contribute to a social insurance scheme. Figure 67. Distribution of Social Protection Beneficiaries - by National Poverty Lines, 2016/17 Free maize 25 38 36 FISP MVAC Free food other than maize 25 38 37 FISP coupons 19 36 44 Supplemental feeding for malnourished children 47 27 26 Cash from government SCTP 38 34 27 Targeted Nutrition Program TNP 30 39 31 Public works other than MASAF 29 34 36 SSN School feeding 29 32 40 Cash from development partners 27 38 34 MASAF 22 32 46 Scholarships secondary and tertiary education 18 24 58 Pensions 9 16 75 SI 0% 20% 40% 60% 80% 100% Extreme poor Poor Non-poor Source: ASPIRE, based on Malawi's IHS4 2016/17. Notes: (1) Beneficiaries' incidence is: (Number of direct and indirect beneficiaries [people who live in a household where at least one member receives the transfer] in a given quintile)/(Total number of direct and indirect beneficiaries). (2) The sum of percentages across quintiles per given instrument equals 100%. 201. The recently launched UBR can be leveraged for more effective targeting. The UBR was launched in 2017 in 11 districts across Malawi. It was created to serve as a consolidated source of harmonized information on the socio-economic status of households to determine their eligibility for social programs. In doing so, it serves as a social registry by supporting the processes of registering households and determining their eligibility for multiple user programs in a coordinated way. 202. The UBR was designed to support registration and eligibility processes for the SCTP and PWP, but is also a potential source of information on the socio-economic status of households for use for other programs, such as the FISP, humanitarian and shock-response interventions, VSL, microfinance, nutrition programs, and bursary scholarships. The standardization of information could reduce targeting costs across programs and minimize errors that would likely result from multiple rounds of targeting across different programs. A recent assessment (Lindert et al, 2018) showed that the UBR has performed well in capturing the poorest households in the first 11 districts. Expanding the UBR to all districts would enable about 70 percent of the population to be registered, making it one of the highest coverage rates in the world (Lindert et al, 2018). Efficiency and Effectiveness of Spending Transfer Values 203. The average transfer value113 for SSNs has increased in nominal terms, but remains low. For SSN programs that provide monetary benefits and are identified in the IHS (these include cash 113 Program administrators set the level of benefits based on program objectives and budget constraints. The review calculates the transfer amount as a percentage of household pre-transfer consumption (welfare). However, the IHSs do not collect monetary information for some of the main programs, including in-kind transfers. For this reason, the analysis does not include this indicator transfers, public works and scholarships), the value of benefits as a share of total household welfare/income was only 4.8 percent in 2016/17. Although this figure is very small, it nonetheless represents an increase of 2.8 percentage points from 2013 (see Figure 68). It is important to underscore that these are average aggregated results for the total beneficiaries of all SSN programs. Analysis based on individual programs and disaggregating for the extreme poor shows more encouraging results. Figure 68. Social Protection Benefits as a Percentage of Pre-Transfer Household Consumption (2010-2017) 22.5 14.0 10.1 4.8 2.6 2.0 Source: ASPIRE, based on Malawi's IHSs 2010/11, 2013 and 2016/17. Notes: (1) Programs included per area: Social protection: social insurance (contributory pensions) and SSN programs (cash from Government, cash from development partners, MASAF, public works other than MASAF, and scholarships for secondary and tertiary education). (2) The share of transfers is: (transfer amount received by all direct and indirect beneficiaries in a population group)/(total welfare aggregate of the direct and indirect beneficiaries in that population group). 204. Most SSN benefits are an important share of the welfare of the poor, yet still fall short of the recommended benchmark to achieve sustained impacts. Evidence from impact evaluations suggest that successful cash transfer programs generally make up at least 20 percent of household consumption (Handa et al. 2013). SSN programs benefits in Malawi, apart from scholarships, constitute less than 20 percent of household welfare. For example, the SCTP represents 18.4 percent and 9.2 percent of the welfare of extremely poor and poor people, respectively. The corresponding figures for cash transfers from development partners are 16.2 percent and 7.4 percent respectively (see Figure 69). 205. Public works programs show the lowest benefits in terms of share of household welfare . The 2016/17 data show that MASAF PWP constituted only 5.5 percent of welfare of the extremely poor and just 2.1 percent of the welfare of the poor. However, these figures may be the result of benefits being annualized in this analysis, whereas MASAF PWP and other public works programs do not necessarily represent a steady flow of benefits throughout the year. For example, beneficiaries may work for only a single program cycle in a year. 206. Social insurance programs constitute almost one quarter of beneficiary welfare. The share of the pension with respect to beneficiary household consumption constituted 22.5 percent in 2016/17, an increase from 10.1 percent in 2013. The overall level is not unexpected, as pensions are intended to replace employment earnings, and such transfers therefore tend to represent a higher share of welfare compared to SSN benefits. For the poor, the share of pensions with respect to household welfare is even higher. The 2016/17 data shows that pensions represent 44.9 and 58.3 percent of the welfare of the extreme poor and poor, respectively (see Figure 69). for FISP, humanitarian assistance (free maize and other food), SMP, the Targeted Nutrition Program and the Supplemental Feeding for Malnourished Children. Figure 69. Social Protection Benefits as a Percentage of pre-transfer Household Consumption - by National Poverty Lines (2010-2017) Pensions 58 SI 45 Scholarships secondary and tertiary 14 education 26 Cash from government SCTP 9 18 7 SSN Cash from development partners 16 Public works other than MASAF 4 6 MASAF 2 3 0 20 40 60 80 Poor Extreme Poor Source: ASPIRE, based on Malawi's Fourth IHS 2016/17. Notes: (1) The share of transfers is: (transfer amount received by all direct and indirect beneficiaries in a population group)/(total welfare aggregate of the direct and indirect beneficiaries in that population group). Program Delivery Cost 207. On average, the cost of delivering one unit of benefit is significantly lower for SSNs than for humanitarian assistance. Figure 70 shows the average administrative expenditure by program as a share of total program costs between 2015 and 2016. The cost of giving one dollar of benefit is 55 cents for humanitarian assistance, compared to much lower amounts for the following SSN programs: SMP (7 cents), NAPE (18 cents), MASAF PWP (11 cents) and SCTP (18 cents). Alternatively, close to 35 percent of expenditure on humanitarian assistance in response to food insecurity in 2015–2016 went to administrative costs, compared to an average (weighted) administrative cost of 14 percent for SSN programs. While a part of this higher overhead cost can be explained by the difficult nature of operating in disaster contexts, lower cost options should still be considered. Bolstering existing SSN programs with elements which can absorb demand in the context of natural disasters could help to reduce the pressure on humanitarian assistance to deliver relief and reduce overall costs. 208. Mainstreaming and harmonization of e-payments is anticipated to reduce program delivery costs and have other benefits. The Government is leading efforts toward a national e- payments model for social protection programs, commencing with a pilot of SCTP and EPWP payments in the Balaka district, building on an existing Irish Aid investment in e-payments. E-payments will reduce administrative costs and, in the context of Malawi, will not only help reduce corruption, error, and fraud, but would also reduce the burden on district staff who currently provide in person delivery of benefits. This gives district staff greater flexibility to be involved in other activities such as oversight and management of programs. Furthermore, e-payments would also facilitate access to the cash assistance by the poorest and most vulnerable people, including those with limited mobility, and support shock- responsive social protection by allowing timely delivery of humanitarian aid in the form of cash. Figure 70: Average Administrative Expenditure as a % of Total Expenditure, 2015 - 2016114 Source: Administrative Data Note: Overall administrative cost is the weighted average of administrative cost for each program (weighted by program expenditure); TEVET cost available for 2016 only; Administrative cost for humanitarian assistance and FFA PWP may be exaggerated as some cost items such as non-food items provided to beneficiaries are included by WFP as administrative cost. Administrative costs for these two programs include adding land transport shipping and handling (LTSH), direct support costs (DSC – such as office costs), other direct operational costs (ODOC, for example contracts to partners and non-food items), and indirect support costs (ISC – the core contribution toward WFP costs globally); MVAC refers to humanitarian assistance from the food security cluster only. Evidence on Impact 209. The overall impact of SSN programs on poverty as captured in the IHS is negligible given their low coverage and benefit levels, but evaluations indicate significantly stronger results for cash transfers than for other programs. The combination of a program’s coverage and benefit level will largely determine its impact on poverty and inequality. Most SSN programs for which the IHS collects monetary information report small coverage (less than 5 percent of total population and of the poor) and low benefit levels (less than 10 and 20 percent of the welfare of the extreme poor and poor respectively). As a result, the overall impact of these programs on Malawi’s poverty headcount and the national poverty gap is negligible at less than a 1 percentage point reduction. However, closer examination of individual programs indicates important evaluative findings, which could be leveraged if interventions were appropriately coordinated, scaled and financed. 210. The SCTP has shown some of the strongest and most consistent positive impacts on consumption, livelihoods, schooling and earning outcomes across the Africa region. Malawi’s SCTP has the highest impacts on total consumption and food consumption out of 27 SSN programs in 14 African countries, with beneficiaries spending an equivalent of 179 percent of the transfer amount, compared to an average of 74 cents per dollar regionally (Ralston et al, 2017). Other studies reported positive, significant outcomes in household productive benefits, including increased investments in productive agricultural assets, time spent on farm activities, and a lower likelihood of engaging in casual agricultural labor outside the home (de Hoop et al 2017; Zezza et al 2010; Covarrubias et al 2017 and Boone et al 2013); improvements in school enrollment, attendance and education-related expenditure for children (de Hoop et al, 2017); and among girls there was a lower prevalence of HIV/AIDS and the 114 FISP administrative cost is for 2015 only and TEVET for 2016 only. FFA PWP administrative cost of 58 percent in 2016 was due to external shipping and handling costs of a $15 million donation from USAID. SCTP: Average administrative expenditures were 14 percent in the Mchinji District pilot (Miller et al 2010). Herpes Simplex Virus type 2 (Baird et al, 2011). Studies have also found that expenditures on temptation goods such as alcohol and tobacco did not increase in beneficiary households, a common but largely unfounded criticism of cash transfer programs (Evans and Popova, 2014; Handa et al., 2017). 211. However, MASAF PWP was characterized by mixed results and a weak accountability framework. An impact evaluation conducted in the 2012-2013 season found no significant improvements in the food security of beneficiaries. It also highlighted several challenges related to the program’s design, such as poor targeting, significant rationing, the relatively small size of the transfers, the infrequency of payments, and insufficient duration of public works projects (Beegle et al, 2015). Substantive design modifications and technical assessments were conducted with a view to improving program effectiveness, including a stronger focus on asset creation, enforcing multiyear employment for beneficiaries, and extending the number of working days offered each year. In 2016, investigations conducted by the Malawi Anti-Corruption Bureau identified small but widespread levels of misappropriation within MASAF PWPs. As a result, a deep analysis of the PWP was undertaken, and supervision efforts by the Government and the World Bank were strengthened, including a stronger fiduciary focus and changes to design and operational procedures. Despite some progress, the MASAF IV PWP concluded in March 2019.115 212. The information and evidence base across non-cash transfer programs is patchy and limited, despite these being among the largest social protection interventions in Malawi. For instance, the SMP provides support to at least one million children. Yet there is a lack of rigorous evidence on the SMP’s outcomes, and on the effect of design elements, such as different feeding modalities (in- school feeding versus take-home rations) and procurement models (centralized procurement, grants, or community production). The most immediate effect of in-school feeding is the short-term alleviation of food insecurity. With expenditures of US$ 59 PPP per child per year, however, the SMP in Malawi was the second most expensive in a comparative study of WFP School Meals Programs in Africa, well above the average of US$ 40 per child per year. Commodities make up 50 percent of the total costs of Malawi’s SMP (MoF, 2016). There is also limited published and robust evidence on FFA, which was supporting approximately 695,000 beneficiaries as of 2019. Better evaluation of substantial SSN programs such as SMP and FFA would help to develop a better understanding of their impacts. 213. Similarly, evidence on the effectiveness of the FISP is mixed . The FISP proved to be exceptionally popular in its first year, during which maize yields doubled and food surpluses reached 53 percent (Chinsinga, 2012; Pauw & Thurlow, 2014). In the ensuing years, however, studies found that the FISP failed to reach the poorest populations, benefiting wealthier households with stronger local networks. Chinsinga (2012) estimated that as much as 61 percent of improved yields were due to favorable climatic conditions, and only 32 percent were due to the FISP. A study by Chibwana and Fisher (2011) in two districts in central and southern Malawi showed that farmers who received the FISP coupon increased the land allocated to maize cropping by 45 percent compared to those who were not FISP beneficiaries, gradually leading to less diverse diets and nutritional deficiencies. Others argue that most of these evaluations measured only the direct effects of the program and ignored indirect effects such 115 The lessons from MASAF IV and public works programs in the region have set the basis for a rethinking of design and control mechanisms for a new public works approach. A pilot is underway in 10 districts covering approximately 10,000 beneficiaries, providing consumption support to poor beneficiary households with labor capacity, who are being provided temporary employment working on restoring catchment areas. The main priority of the reform pilot is to test approaches for integrated watershed management and improved asset outcomes, and on focusing on timing interventions across the seasonal cycle. Improving asset creation by linking projects to watershed catchment management initiatives has proved successful under Ethiopia’s Productive Safety Net Program. A successful pilot demonstration would provide the basis for the commencement of the new enhanced public works approach under the SSRLP, covering up to 35,000 households. as spillovers and macroeconomic effects, and risk factors such as weather and world price shocks. Some have argued that FISP spending crowded out investment on other necessary SSN interventions during this time (World Bank, 2016; Chibwana and Fisher, 2011). Institutional Arrangements 214. The provision of SSNs is characterized by fragmentation at the national level. Several different line ministries are tasked with various aspects of implementing the programs, and the MoF does not have sufficient capacity to ensure that all stakeholders are held accountable or to enforce cross- ministerial coordination. If any links between programs exist, they are not guided by specific policies or through shared administrative systems. Some stakeholders have recently suggested that developing a legal framework for social protection in Malawi may provide a foundation for strengthened Government coordination. 215. The absence of an adequate M&E system for the MNSSP has been a crucial problem . An updated M&E framework has recently been developed. However, it remains to be seen whether the new framework will encounter the same challenges as the previous one, which was not comprehensively implemented. Considerable capacity gaps at all levels led to inefficiencies and hampered harmonization and coordination. Furthermore, individual social protection programs worked largely in isolation with their own systems, lines of reporting and accountability mechanisms (MoF, 2016). A regularly updated monitoring dashboard, ideally public, could help align internal and external stakeholders at all levels through greater accountability and peer pressure. 216. Weak coordination of donor efforts contributes to ownership, sustainability, and fragmentation issues. Fiduciary concerns have impaired donor trust in Government financial systems and resulted in separate and fragmented funding for safety nets. Between 1999 and 2010, Malawi received greater levels of donor support than other Sub-Saharan, low-income and Least Developed Countries (Dionne et al. 2013). This support stopped briefly after the 2013 ‘Cashgate’ scandal relating to the theft of public money. The SMP has hardly any Government involvement, whereas the SCTP and MASAF PWP have made considerable use of Government systems. Donors contribute to this fragmentation by often insisting on separate financing structures from that of the Government, even within the same program, to protect funds, different disbursement time frames, and reporting requirements. (Donors also commonly undertake their own audits of the same program.) Not only do these requests become difficult to coordinate, they require costly reporting and logistical oversight. This raises concerns about the ownership and sustainability of programs that exist almost independently of the Government. It also reduces incentive for the Government to learn, take over and run safety net programs if development partners refuse to share tasks and risks. 217. Finally, ineffective coordination at the district level is one of the most significant constraints to the development of a more coherent system. Throughout the MNSSP II consultation process, district-level stakeholders stressed that better coordination and integration among safety net programs would result in more efficient use of district officers’ time and energy. Currently, each SSN program requires district officers to participate in a range of meetings, often involving the same stakeholders. Efforts are now underway to create single social protection committees in every district to coordinate programs. However, coordination is only part of the problem, as programs must also contend with inadequate resourcing, infrastructure and staffing, which limit their ability to operate effectively at the district level. Conclusions and Recommendations 218. Despite facing significant challenges, Malawi maintains an ambitious social protection agenda and is equipped with an array of programs that strive to meet international targets to support its poorest households. Coverage and financial trends are improving, but significant gaps remain to meet the needs of the poor and the wider population. Rigorous evaluations have identified the effectiveness of programs such as the SCTP, and the Government’s cutbacks to FISP in support of the SCTP indicate an appetite and political capacity for reform. The recently finalized MNSSP II reflects the Government’s commitment to improve pro-poor interventions and serves as a vital entry point for dialogue around social protection reforms. The recommendations below offer a set of programmatic and financial reforms based on the above analysis. Reform efforts would benefit from involving all stakeholders, from development partners to district and national level implementers, and should take Malawi’s broader political economy challenges into account. The need to link increasingly available evidence on program impacts to future social protection planning and decision-making remains an imperative. i) Programmatic reforms which are crucial to improve the overall effectiveness of the sector 219. Reallocate resources from less effective initiatives toward evidence-backed social protection programs. Only a small portion of social protection spending is for SSNs. With high levels of spending directed to the FISP (and now AIP) and social insurance, opportunities exist to reallocate resources toward more effective and efficient programs. The SCTP is a good example of how rigorous evidence can support the scale-up of effective interventions. Efforts to advance the goal of national coverage for the SCTP should be sustained, and opportunities to explore asset creation and skills development linkages should be explored to further amplify the SCTP’s impact. Additional evidence is needed to influence decisions for other consumption-focused programs, particularly over cost effectiveness. 220. Improved targeting of the poor requires new approaches and systems. Efforts should be sustained to improve the targeting of support to people who are chronically poor and who are transitory poor. This requires continued revision of targeting practices and innovations, including geographic, community and means testing approaches. In supporting these efforts, the expansion of the UBR to register households nationally, and its use across multiple social protection programs, particularly SSN programs and humanitarian interventions, would be an important step. Scale up of the UBR will only be successful if a predictable implementation and financing strategy is put in place, and key roles and responsibilities in the UBR’s implementation are established with sufficient flexibility for adaptation. If the UBR can be implemented nationwide successfully, then it can serve as an example for other African countries taking initial steps toward creating their own social registries. 221. Incorporate human capital-oriented elements in safety net interventions, including through an integrated or multi-sectoral approach. As endorsed in the MNSSP II, a more systematic approach toward SSN design, particularly one that integrates human capital and livelihood opportunities, is critical given Malawi’s evolving demographic patterns, skills requirements and political environment. A key lesson from experiences elsewhere in Africa is the importance of ensuring that interventions are complementary and coordinated. Multi-sector approaches that capitalize on synergies between social protection and other sectors can amplify outcomes (World Bank Social Protection Policy, 2012 - 2022). For example, an integrated approach to social protection, education, health, agriculture and nutrition interventions is likely to be more effective in breaking the inter-generational cycle of poverty than if the sectors operated independently. Two immediate areas where this approach could be taken in Malawi is the promotion of early childhood development through nutrition-sensitive safety net interventions, and the provision of livelihood enhancement programs focusing on skills development and access to capital (for instance COMSIP, VSL) The Government will need to work closely with local communities and institutions to successfully design and implement the human capital-centered ‘graduation pathways’ endorsed in the MNSSP II. An institutional hub for active labor market interventions could be one such intervention to coordinate across complementary initiatives while designing and launching new ones. 222. Improve the shock sensitivity of safety nets through greater linkages between safety net programs and humanitarian assistance, and by establishing scalable financing mechanisms. As the MNSSP II notes, investments in advance-planning, clear protocols and early warning systems— including estimating potential needs—can improve the speed and accuracy of humanitarian assistance to households most at risk. Safety nets and humanitarian assistance operate independently despite having many overlapping functions. Yet predictable extreme weather patterns mean that certain households will be more affected than others, with a portion of the population likely to require repeated seasonal assistance. This creates space for safety net programs to work with humanitarian assistance on strategies to prevent and mitigate weather-related shocks for households at risk. For instance, humanitarian assistance programs could consider moving away from relying on annual targeting exercises to identify vulnerable households, using this method only for households facing exceptional shocks. Sustainable financing mechanisms with multi-year horizons can provide longer-term support for affected households. Payments and beneficiary selection platforms can enable efficient and timely selection of beneficiaries and disbursement of payments to help shift the response burden away from humanitarian assistance during times of emergency, and reap cost-savings. The introduction of urban cash transfers in response to COVID-19 highlights the importance of considering both urban and rural contexts in developing a comprehensive shock-responsive social protection system. Finally, program contingency financing can support the adaptive capacity of safety net programs to scale up as needed. 223. Establish a national productive inclusion strategy that ties together existing programming and provides a vision for the sector. There is now an opportunity to work toward a “safety nets plus” model, which builds on existing interventions to promote stronger linkages to economic and social inclusion. Throughout the MASAF series of projects, livelihoods support was designed and implemented as a largely stand-alone activity, where linkages to the beneficiaries of other programs were not actively sought. During the most recent phase of MASAF IV implementation, efforts were made to formalize these linkages, and the experience was positive. This work could now be advanced while also looking more broadly at experiences across related sectors, and the wider international context. A considerable amount of global literature supports this agenda. 224. Continue investing in core delivery systems. Malawi should continue to support its social registry, e-payments, grievance and transparency initiatives to strengthen the sector. For example, about a third of resources devoted to humanitarian assistance are currently spent on setting up and delivering benefits to beneficiaries. A streamlined beneficiary identification and payments system could significantly reduce this cost, reduce leakages and improve the likelihood that resources are delivered directly to beneficiaries. Additionally, effective implementation of a national M&E framework and information system could enable better tracking of progress and coordination across programs. The World Bank assessment of Malawi’s social registry highlights MoF’s visions toward an integrated delivery system, with e-payments, UBR and M&E forming key components (Lindert et al, 2018). Operationalizing this will require concerted effort from the MoF. The UBR offers a platform that may be helpful in social sector programming beyond just social protection, such as the provision of safety nets in the aftermath of climatic shocks; in education (for example, bursary programs); in health (for example, health insurance or locations of primary healthcare centers); and targeted subsidy programs in agriculture. There is also a need to institutionalize the UBR to give it a permanent home as well as build a sustainable financing strategy for it (Lindert et al, 2018), and strengthen oversight and M&E mechanisms to enhance data quality and maintain its relevance through further rounds of data collection. 225. Support improved analysis around the future needs and sustainability of social insurance. Pension expenditures are likely to be unsustainable in the future as the share of the old age population increases. The Government will face a large fiscal burden if it responds to demographic changes by scaling up pension coverage, through both the CSPS and NPS. Against this backdrop, further analysis is required to consider future needs and liabilities, and potential innovations including scope to integrate common approaches to social insurance across the public and private sectors. Establishing a clearer evidence base is an important precursor to informing discussions around financing. ii) Financing reform necessary to support a fiscally sustainable scale up of effective programs in the sector 226. Reallocate current fiscal spending toward social safety net programs within the social protection program mix. Malawi’s agricultural budget has focused disproportionately on FISP fertilizer and seed subsidies for maize, with studies suggesting that political favor has crowded out spending on more productive interventions within the agricultural sector (World Bank, 2016; Chibwana and Fisher, 2011). Recent analysis estimates that reallocating expenditures from the FISP (and now AIP) to the SCTP would result in less leakage of benefits to households in the highest income deciles (World Bank, 2018b), while the IMF (2015) has advocated for increased cash transfer programming and investments in agricultural research and development to complement the FISP spending. FISP reforms aimed at restricting the focus to agricultural objectives while scaling up consumption support programs through the SCTP should be continued and can provide a starting point for further discussions around spending reallocations. 227. Create a national financing strategy for social protection. With a shift to more effective social protection programs, such as SSNs, it is increasingly important that the Government establish a concrete social protection financing strategy. Relying on donors to finance Malawi’s social protection programs, especially safety net programs, leaves substantial gaps, results in inadequate support and increased vulnerabilities for the poorest segment of the population. The Government must establish and meet spending budgets for safety net interventions, while concomitantly utilizing existing evidence to shape a better funded, more efficient and flexible strategy for safety nets. Evidence from the region indicates that it is possible to achieve three goals in parallel: (i) increase the level and sustainability of financial resources; (ii) identify the most appropriate mix of domestic, foreign, public and non-public funding sources; and (iii) utilize a flexible financing strategy to respond effectively to shocks (Beegle et al 2018). 228. Utilize contingency or reserve financing mechanisms. The Government can better consolidate safety net, humanitarian, and disaster responses through contingency financing mechanisms. Such funds are common in countries such as Colombia, Mexico and Uganda, with Kenya and Mozambique in the process of establishing similar funds. Such initiatives provide households with loans that increase liquidity immediately following a shock. Some multilateral development banks are also familiar with this method, offering lines of credit deployable after shocks. 229. Consolidate existing funding arrangements. Perceived fiduciary risk has resulted in many donors avoiding Government-based financial systems in favor of separate funding channels, resulting in fragmented and inefficient funding streams that generate considerable administrative burdens. Donors’ insistence on different financing channels and implementation modalities make it difficult to coordinate, with separate structures sometimes even within the same program, such as the SCTP. The MoF has expressed the desire to prioritize pooled, harmonized funds (MoF, 2016). Recent initiatives like the Social Support Fund Task Force play a critical role in advancing the discussion around ways to establish such an enabling environment. Donors should also make efforts to acknowledge the additional burden that separate funding channels place on an already burdened implementation system. 230. Engage donors in discussions around funding structure revisions. Humanitarian assistance for food security and nutrition averaged 2.2 percent of GDP between 2011 and 2016. This is significantly higher than other countries in the region. Humanitarian assistance provided outside Government is not fungible, raising questions as to whether existing donors would be willing to make multi-year commitments to financing seasonable and scalable safety net interventions in lieu of short-term humanitarian assistance. Any substantial shift toward greater integration of safety nets and humanitarian assistance would therefore require discussion with donors working in the social protection and humanitarian sectors. Intergovernmental Fiscal Transfers and Local Service Delivery Review Introduction 231. This review was undertaken to support the Government of Malawi to analyze the efficiency and equity of public expenditure in its ongoing efforts to accelerate and deepen decentralization. It is intended to provide a comprehensive analysis of the intergovernmental fiscal transfer system (IGFTS) with the goal of developing recommendations on how to improve efficiency and equity of resources, responsibilities, and personnel vertically (between central Government and districts) and horizontally (across and within districts). The exercise also aims to qualitatively analyze the impacts of the IGFTS on local service delivery and to better understand the “coping mechanisms” that local authorities (LAs) have developed and operate amid an often unreliable system of fiscal transfers. Decentralization in Malawi: Promise vs. Practice 232. Decentralized service delivery was formally enshrined in the 1995 Constitution as part of Malawi’s transition to multi-party democracy. It was intended to diffuse overly centralized power and bring services closer to citizens through elected local councils. Implementation of decentralization in Malawi has been directed by the 1998 National Decentralization Policy (NDP), Local Government Act, and two national decentralization programs. 116 In practice, policy reforms aimed at decentralized governance and service delivery were rolled out in a fractured, uneven and incomplete fashion. To date, decentralized service delivery in Malawi has been characterized by a blurring of intergovernmental accountability lines – both between and within each level of Government. Figure 71: Decentralization Roles and Functions 233. The unique nature of Malawi’s political settlement and economic environment has resulted in power struggles centering around the development capital budget which is highly dependent on aid. Political economy constraints concentrate competition at the center where public sector rents are allocated. This leads to an incomplete and uneven implementation of administrative and 116 NDP I, 2001-2004; NDP II, 2005-2009 (revised to 2008-2013) fiscal decentralization. Even in the most devolved sectors, education and health, central Government continues to manage capital investment and expenditure, procurement and distribution of materials (drugs, textbooks, etc.), and even small works such as boreholes. At the local level, fragmentation provides the incentive that allows the benefits and costs of leadership decisions to be contested. 234. High level political support for decentralization has been, and is likely to remain, ambiguous and dynamic. While decentralization was a key feature of MDGS II, there was high level pressure to overcome resistance to decentralize power and resources from line ministries. Taking these dynamics into account, recent moves in favor of decentralization have presented an opportunity to endorse more workable arrangements between central line ministries and local authorities that could promote incentives for improved services. Malawi’s Intergovernmental Fiscal Transfer System and Service Delivery 235. Malawi’s intergovernmental fiscal transfer system (IGFTS) is the primary mechanism by which decentralized local service delivery responsibilities are financed. However, this system has persistently failed to establish a match between policy commitments and fiscal outlays as Government transfers for health and education have decreased by 25-45 percent for almost all local authorities on a real per capita basis since 2014. Implicit policy appears to be staying as close to the nominal value of the previous year’s budgets and transfers. The 1998 NDP mandates that at least 5 percent of net national revenues (NNR) be transferred to local Governments for development financing, but this has yet to exceed 1.65 percent.117 236. The challenges posed by the disconnect between “form and function” are well documented, particularly for human capital services which receive the lion’s share of outlays in national and local budgets and by donors. Health sector policy acknowledges that “resources are inefficiently and inequitably allocated and managed at all levels” 118 and that resources made available to local levels have become “increasingly less sufficient to enable local Governments to discharge their responsibilities.”119 Education sector policy is similarly disconnected from the need to address declining quality, access and affordability.120 There are glaring disjunctures in outlays for primary and secondary education, and vast unmet needs for capital spending disable the benefits of steady increases in outlays on human resources, posing serious challenges for primary and secondary completion rates, especially for girls. 237. The IGFTS is characterized by a large number of fragmented transfers that are earmarked or conditioned primarily by the center. The IGFTS in Malawi works through cash-based transfers, where the flow of funds passes through several organizations and institutional processes. 121 The system comprises 22 different transfers, which includes 18 sector transfers for non-wage other recurrent transactions (ORT), which tend to be heavily earmarked. The transfers with their basic features are listed in Annex 2, and trends in the composition of transfers are depicted in Figure 72, below. Although personal 117 The definition of NNR remains disputed. For this purpose, development transfers include: DDF; Infrastructure Development Fund (IDF - discretionary resources for the urban councils); Constituency Development Fund (CDF); and ad hoc special purpose transfers for boreholes and community development. 118 Government of Malawi (2018). Health Sector Policy (2017-2020) p. 20. 119 IGFTS Policy Note. 120 The Borgen Project (2018). Facts about Education in Malawi, in https://borgenproject.org/10-facts-education-in-malawi/; UNICEF (2019). Education Budget Brief, 2018/19, Towards Improved Education for all in Malawi. 121 The flow of funds is administered by the Ministry of Finance, involves the Accountant General and the Central Bank, and ultimately executes transfers through manual transactions with commercial banks. Annex 1 displays the different stages in the flow of funds from central to the lowest level. emoluments account for 75 percent to 80 percent of all budgeted funds at the local level, the funds are not part of the transfer formula and are never actually transferred to local council accounts. 122 Figure 72: Composition of Transfers from FY 13-14 to FY 18-19 (by transfer) 35 33.1 B 33.0 B Percent of Total Annual Transfers (Billion Kwacha) 2.40 29.9 B 30 29.0 B 6.34 2.25 1.16 2.21 2.70 4.00 25 2.84 3.24 1.76 22.0 B 4.40 2.33 1.27 1.74 3.47 20 4.10 17.5 B 1.74 2.47 4.91 2.48 1.09 1.78 1.39 0.95 1.02 1.30 1.02 1.14 0.88 15 0.76 0.61 5.86 6.83 7.58 5.37 6.35 10 6.80 5 9.85 9.48 9.06 9.08 8.87 7.28 0 2013-14 2014-15 2015-16 2016-17 2017-18 2018-19 Education Health Agriculture GRF CDF IDF DDF Boreholes COMSEP Other transfers (Key to Figure 72, above) Health and Education have remained as the two largest transfers, but have remained relatively constant in size despite the increased decentralization of the respective sectors Other transfers saw a onetime bump in 2016 following transfers to the 4 city councils for road rehabilitation The DDF was introduced for the first time in 2016, but has declined significantly in size over the past two years following midyear budget slashes… …while the size of total development transfers has continued to grow, but with increasing fragmentation 238. Development transfers have increased as a percentage of the total transfer pool in recent years but have also become highly contested, putting their utility at risk. With the introduction of the District Development Fund (DDF) in FY 2015-16 to complement the Constituent Development Fund (CDF), development transfers increased from 6 percent of the total pool in FY 2013-14 to 38 percent in FY 122 Payment of salaries is a deconcentrated function with districts entering payroll data into the system through one of six payroll servers located throughout the country and most Government employees are paid directly by transfers from the National Treasury Account into their personal bank accounts. 2018-19 (see Exhibit E in Figure 72, above). The ‘vibrancy’ of this decentralization policy has not played in one direction, however, as this pot has grown with increasing fragmentation as shown in Figure 73, including: (i) a disproportionately growing CDF which has increased by 132 percent since 2014; (ii) a dedicated ‘boreholes grant’ which provides funding for four boreholes in every constituency annually; and (iii) in FY18-19, a community-driven fund (COMSEP) 123 which splits MWK 4 billion across all constituencies through Area Development Committees (ADCs). Figure 73: Composition of Development Transfers from FY 15-16 to FY18-19 (by transfer) 12 11.1 B 10 9.3 B 4.0 8 2.2 MWK (Billions) 6.7 B 6 5.2 B 1.7 3.4 4.4 4 2.3 4.1 2 2.8 3.2 2.7 1.3 0 2015-16 2016-17 2017-18 2018-19 Fiscal Year DDF (Actual) CDF (Actual) Boreholes (Actual) COMSEP (Budgeted) 239. Ultimately, the capability of Malawi’s IGFTS must be evaluated in the context of the power contests and accountability relationships which have characterized the country’s decentralization journey to date. The precarious nature of the IGFTS – the mechanism by which the center is meant to realize policy priorities for local service delivery – disables the accountability relationship between local government and citizens. At the same time, it enables poor financial and resource management practices. Council officials point to a slew of constraints that stem from the central government and are beyond their control – insufficient staff, information and communications technology (ICT) problems affecting proper use of the IFMIS, and inadequate policy guidance. Untimely and unpredictable fiscal transfers, particularly the DDF, are a serious impediment to effective investment planning and implementation. Political economy factors surely contribute to patterns of mismanagement. But to a certain extent, deviation from required practices is in effect a coping strategy, bending the rules in order to meet pressing needs, such as paying contractors on time, supplying health facilities, responding to seasonal demands in agriculture and meeting the needs of traditional. 240. The urgency of the COVID-19 response has accentuated the shortcomings of the IGFTS and broader arrangements of decentralized service delivery in Malawi to efficiently deploy funds in a coordinated fashion. At the onset of the pandemic, a blanket transfer of between MWK 30 – 35 million with loose guidelines for implementation were transferred to each Local Authority directly by the Ministry of Health, bypassing the systems of the formal IGFTS through the NLGFC. As resources continued to flow into the country, there were increased concerns around the lack of transparency and coordination 123 This window was the controversial result of an attempt by the Government to split the amount among 86 MPs who voted against the electoral reform bill in December. Following protests, the Government was pressured to allocate the funds across all constituencies. The window is not expected to remain in the budget in FY 20-21. surrounding the funds and equipment earmarked for Local Authorities and the (lack of) mechanisms for strategic allocation of resources. Finally, there have been challenges in bridging the gap between national and local COVID-19 response plans. All 35 Local Authorities have developed budgeted COVID-19 response plans, but initial indications reveal these are largely disconnected from what has been planned and budgeted for at the national level. IGFTS Performance – Parameters and Coping Mechanisms 241. The Intergovernmental Fiscal Transfers System is described and evaluated according to five parameters: adequacy, timeliness, predictability, transparency and equalization. These parameters are analyzed using data from the transfers system by invoking basic statistical measures, trend analysis in the case of adequacy, and drawing upon qualitative information to assess transparency. In many cases, the data quality limits the analysis, and thus the confidence that can be placed in findings. Adequacy 242. Revenue adequacy is a measure of the extent to which transfers made to subnational authorities match the expenditure needed to meet the governance and service delivery obligations assigned to them in policies or laws. An IGFTS meets this criterion when local authorities have sufficient resources to cover their unmet revenue needs after raising own source revenue and when the transferred resources over time grow with expanding needs. 243. The effects of suboptimal adequacy on local governance and service delivery will manifest not only through underfinanced local mandates, but also through weakened accountability functions and fiduciary practices as staff and councilors scramble to meet competing demands on a daily basis. As a coping mechanism, local authorities resort to ‘internal borrowing’ to bridge cash flow constraints in priority sectors and sustain service delivery. If revenue adequacy is chronically suboptimal, service delivery weakens. This would undermine the legitimacy of local political processes (such as planning, budgeting or reporting on expenditures and results), and results in citizen disaffection with the local authority, corroding the basis of local political accountability. 244. The adequacy of Malawi’s IGFTS is well below par. A vertical balance principle is not clearly recognized in policy and there is no policy indicating that the ‘finance should follow function”. In fact, the implicit policy appears to be staying as close to the nominal value of the previous year’s transfers. From FY 13-14 to FY 17-18, in real per capita terms, transfers increased only marginally from MWK 1,522 to MWK 1,566. But as a percentage of Net National Revenue (NNR),124 while the total of all transfers have fluctuated, they have declined from 6.2 percent in FY 13-14 to 5.4 percent FY 17-18 (see Table 7 below). Table 7: Total transfers as a share of NNR, FY13-14 – FY17-18 (in MWK billion) FY 13-14 FY 14-15 FY 15-16 FY 16-17 FY 17-18 Net National Revenue 284.67 330.96 424.48 552.98 610.47 Total transfers 17.60 22.01 33.05 29.05 33.04 Transfers/ NNR 6.2% 6.7% 7.8% 5.2% 5.4% Development Transfers/ NNR 0.37% 0.55% 1.37% 1.35% 1.66% Source: Government of Malawi. 124 Defined as tax collected less recurrent statutory expenditure i.e. debt service, refunds and retentions, pensions and gratuities 245. Development financing to districts includes funds that are not captured by the general resource fund (GRF) or other recurrent transactions in the sector. 125 While development transfers have consistently grown as a percentage of the overall transfer pool over the past 6 years – the sum still falls well short of this policy target, ranging from 0.37 percent (in FY 13-14) and topping out at 1.65 percent (in FY 17-18) (see Figure 74, below). Since FY 15-16, actual DDF transfers have steadily declined as a share of NNR, from 0.67 percent to 0.44 percent. Figure 74: Development transfers as a share of NNR, FY 2014-2018 (in MWK) 1.8% 1.66% 1.6% Development Transfer as % of NNR 1.37% 1.35% 1.4% 1.2% 1.0% 0.8% 0.6% 0.55% 0.37% 0.4% 0.5% 0.8% 0.5% 0.7% 0.6% 0.6% 0.7%0.4% 0.2% 0.4% 0.01% 0.03% 0.0% 2013/14 2014/15 2015/16 2016/17 2017/18 DDF + IDF CDF Boreholes Total Devt. Transfers Source: Government of Malawi. 246. Figure 75 depicts the trends in various transfers over time in real, per capita terms. There has been a consistent decrease in non-wage recurring expenditure in health and education over the past four years. This is seen in the yellow band where per capita health and education transfers have decreased by between 25 percent to 45 percent for almost all councils from 2014 – 2018. 126 At the same time, CDF transfers per capita have increased by an average of 132 percent across the councils over the same timeframe. This suggests that national policy targets are significantly disconnected from trends in transfers. For instance, commitments made in the Education Act 2013 to ‘decentralize’ primary education’ are belied by the fact that, since 2013, education transfers in per capita real terms decreased materially for all councils. This is also seen with the DDF, which have been static or declining for most districts since its introduction. 125 Development transfers includes DDF, IDF, CDF, Boreholes, and later COMSEP (which only begins in 18-19). These are a subset of the ‘Total Transfers’ in the row above. 126 In absolute amounts, health and education transfers have remained relatively constant over the past 5 years (Health MWK 6.7 billion in FY13-14 vs. MWK 6.4 billion in FY18-19; Education MWK 7.3 billion in FY13-14 vs. MWK 8.8 billion in FY18-19) despite significant inflation, currency devaluation, and population growth. Figure 75: Real, Per Capita District Transfers Changes Over Time (2014 – 2018) Education PC Change 2014 - 2018 Health PC Change 2014 - 2018 DDF PC Change 2016 - 2018 CDF PC Change 2014 - 2018 200% % Change Per Capita Transfer (Years) 150% 100% 50% 0% -50% -100% Ntcheu Chiradzulu Likoma Mzuzu City Nkhatabay Zomba City Neno Mulanje Thyolo Chikhwawa Nkhotakota Nsanje Karonga Mwanza Chitipa Mchinji Rumphi Blantyre City Ntchisi Blantyre District Machinga Mangochi Lilongwe City Lilongwe District Salima Zomba District Dedza Dowa Phalombe Kasungu M'mbelwa Balaka 247. The effects of persistent vertical imbalance are egregious. There is no evidence that efforts have been made to correct vertical imbalances. This is evidenced by the fact that for sectors like health and education, transfers have not grown and have decreased in real, per-capita terms, despite the continued deconcentration of functions. Second, while development transfers have been increasing, they are increasingly fragmented and growing at the expense of the DDF. Meanwhile, the DDF represents the only pool of truly ‘discretionary’ development funds for local authorities to implement their District Development Plans/Annual Investment Plans and respond to local needs. Finally, in the context of the lack of an overall vertical allocation strategy, there is no basis to determine the true shortfalls in transfer amounts. Predictability 248. Predictability is the degree of certainty associated with the amounts to be received in transfers by local government against the amounts denominated at the beginning of the year. Where transfers are chronically unpredictable, common patterns emerge: local governments undervalue their plans and budgets, their support and supervision of service facilities becomes unreliable, they fail to meet contractual obligations with vendors and contractors. The coping mechanism employed here involves local contracting and splitting contracts. Under contract splitting, the project scope is split into more than one contract to allow for local contracting, cost minimization, and better cash flow management. Despite these short-term benefits, local contracting often compromises on quality and is prone to local political influences. 249. Unpredictability also deters local authorities from meeting their obligations to co-produce new civil infrastructure assets and maintain basic services together with communities. Again, the credibility of local councils is undermined, citizens refuse to take council business seriously, politicians are more likely to behave opportunistically, and the accountability of administrators becomes blurred by a myriad of ad hoc arrangements and undocumented decisions. 250. Predictability of the IGFTS in Malawi varies across transfers (see Table 8 below). For the largest transfers of sector ORT in health and education, there has been relative predictability both in annual budgeted figures never decreasing from the previous year and end-of-year receipts always being greater that 90 percent of the budgeted amount. The GRF in FY 16-17 and FY 17-18, received an average of 97.5 percent of budgeted funds but took a significant hit in FY 18-19, only receiving 62.3 percent of what was originally budgeted for the year. Development transfers display a different story, with the politically “protected” CDF growing and recognizing above 99 percent of budgeted totals every year – while the DDF has been extremely volatile both in budgeted figures and year-end receipts. Table 8: Actual End-of-Year Receipts as a % of Budgeted Amounts (by transfer) FY 16-17 FY 17-18 FY 18-19 Average Health 92% 95% 100% 96% Education 100% 100% 100% 100% DDF 98% 59% 40% 66% CDF 100% 99% 100% 100% GRF 100% 95% 62% 86% 251. The DDF has suffered the most unpredictability of all transfers since its introduction in FY 15-16. When it was originally conceived in 2014, the DDF was budgeted as a MWK 5.0 billion pilot for 5 councils but settled at MWK 3.2 billion over the past two years. Actual transfers have regularly been cut, however – averaging a 50 percent reduction vs. budgeted amounts in three of the past four fiscal years (see Figure 76, below). Figure 76: DDF Budgeted vs Actual End-of-Year Receipts (by year) 6.0 5.0B 5.1B 5.0 43% cut 47% cut MWK (Billions) 4.0 3.3B 3.3B 3.2B 3.2B 3.2B 3.0 2.7B 60% cut 2.0 1.3B 1.0 0.0 2015-16 2016-17 2017-18 2018-19 2019-20 Fiscal Year DDF (Budgeted) DDF (Actual) 252. This makes it difficult for local authorities to carry on with the planned expenditures budgeted with the DDF. In other cases, delays in planned project components force local authorities to increase reliance on ‘community contributions’, attracting allegations of mismanagement and souring this important relationship. The quantum of contribution is usually declared and agreed at the outset, but often evolves during project implementation. 253. Despite the headline annual predictability of sector ORT, there are still variances across the districts for monthly transfers. As seen in Figure 76 (above), the actual education transfers equaled the budgeted. However, during the year, there was considerable variance among councils across months. Given that the annual transfers matched the budgeted amounts for each council, the monthly fluctuations were balanced out during the fiscal year. 254. The DDF’s annual unpredictability is also characterized by significant intra-year variability – with relatively significant differences across districts. This has significant consequences for project timelines and unit costs. It has also left many local authorities wary about disbursing funds to a project until they have accumulated the full amount of the contract, resulting in an increasing number of delayed and incomplete projects as a coping mechanism. In FY 15-16 and FY 16-17, the DDF was transferred on a monthly basis to local authorities – with many frequently receiving zero allocation for a given month. 127 In FY 17-18, DDF transfers were switched to a quarterly basis with the expectation that this would help improve cash flow at the district level. Figure 77, however, indicates this was not the case in the face of significant mid-year budget cuts. Through Q2 the benefits were disbursed as planned, with councils receiving 49.0 percent of their budgeted DDF allocations on average (albeit with variance across councils – with a range of 37 to 55 percent). Following mid-year cuts in Q3, however, local authorities only received an additional 2.0 percent of their annual allocation on average – and finished the year receiving only a total of 59 percent of their total budgeted DDF. Figure 77: DDF Budgeted vs Actual End-of-Year Receipts (by year) 100% Cumulative % Budgeted DDF Allocated 90% 80% 70% +8.3% 60% +2.0% 59.2% 50% 49.0% 50.9% +25.1% 40% 30% 23.8% 20% 10% 0% Q1 Q2 Q3 Q4 Quarter Note: Error bars represent minimum, average, and maximum level of district DDF allocation for the quarter vs. budgeted total In FY 15/16, 9 rural districts received "0" DDF for at least one month in a year, with 2 of these receiving "0" DDF for two 127 months; In FY 16/17, all rural districts received "0" DDF for at least one month, with 5 of districts receiving "0" DDF for two months, and another 2 receiving "0" DDF for three months. Timeliness 255. Timeliness refers to the congruence between the administratively stipulated dates of transfers and the actual dates when transfers are received. Timeliness, therefore, is assessed by the lag between the two dates. The timeliness of transfers is crucially important where, as in the case in Malawi, local authorities operate in a tight and precarious fiscal space with numerous expenditure responsibilities on which multiple service facilities depend for continuity. 256. Lack of timeliness affects local governance and service delivery in several ways. It results in delays in payment of salaries of those staff on district payrolls, postponed supply of essential materials to schools, clinics and other services and rescheduling of work on infrastructure projects with potential cost overruns and additional payments potentially needed to redress grievances with vendors and contractors. When the delays become a permanent feature of the IGFTS, the authorities may evolve standard operating procedures in the way local budgets are executed. These ad hoc mechanisms may mitigate the disruptions caused by untimely transfers. However, they can also have perverse effects on the quality of accounting, including the integrity of the IFMIS, and on the reliability of financial reports. 257. The timeliness of transfers could be analysed only partially. This was due to data unavailability, which made it hard to verify claims of chronic delays and unpredictability in transfers . Curiously, the administrative practice is to record the date of receipt of transfers – regardless of how delayed they may be – in the date when they were due. However, dates when the NLGFC notified local authorities of transfers were available for the first 9 months of the current fiscal year. Whereas the administrative calendar stipulates that transfers occur on the 10th day of each month, Figure 78 shows that the transfers were received on average 22 days later than that date. Figure 78: Average Transfer Delay (Planned vs. Actual Transfer, 2018 – 2019) July 24 August 21 September 26 October 19 November 16 Avg transfer delay: 21.9 days December 19 January 28 February 29 March 22 April 15 0 5 10 15 20 25 30 35 Number of days transfer delayed 258. Sub-optimal timeliness of transfers is mitigated by coping mechanisms deployed by district administrations, but severely disrupts health and education facility operations. (See Box 6 below). The problem of timeliness cannot however be readily or consistently accommodated at the level of service facility operations. This varies by the district as some schools in some districts experienced longer delays in transfers. Similarly, drug supplies against allocated shares can be inordinately delayed for health centers, with local managers seeking out essential supplies from donor vertical programs where they are available. Such coping mechanisms can result in decisions that starve service facilities for certain periods. 259. Facilities reported delays much more than districts and this may be due to cash flow management at the district level as well. At present, the issue of lack of timeliness is magnified by the inadequacy of transfers. In both the health and education sectors, district authorities distribute resources, funds or supplies, according to ad hoc rules that they have developed. The timeliness of allocation of resources, therefore, almost solely depends on the discretion of the District Health Officer and ‘leveraging of relationships’ is required as a coping strategy. Box 6: Managing at the Front Lines of Health Service Delivery Malawi’s primary and community level health facilities (e.g. health centers, community hospitals, and maternity clinics) serve as the first point of contact for healthcare for most of the citizens. They provide a range of preventative, diagnostic and curative services – most notably malaria treatment, HIV/AIDS, and maternity services. Under the IGFTS, primary and community level health facilities do not receive any funds for local expenditures. They do receive drugs from the Central Medical Stores Trust (CMST) and supplies from the local authority. Field work displayed how health centers work with gaps in infrastructure and systems with little means to address them. Primary and community-level facilities complain that requests to the district health office take many months to be fulfilled. This contrasts with the district hospital whose needs are immediately addressed, according to some sub-district health facility respondents. Coping mechanisms observed by health centers to bridge the gap in an increasingly resource constrained environment include: Staff that are available are reassigned to technical roles, even if not trained; Drugs are rationed or doses recombined; Donor programs used to get drugs otherwise not on their lists; Community contributions raised where possible; Reorganize service delivery within the possible domain (i.e. prescribe drugs without confirmed diagnosis through tests); Relations with community managed at the levels of service delivery that allow for periodic failures. Transparency 260. Transparency of the IGFTS is dependent on three factors, namely, (i) The formula for allocating transfers, (ii) adherence to formula and (iii) publication of district allocation shares via budget and actual receipts and reporting. Complex formulae fail as instruments to communicate national policy preferences to local governments. This impedes ‘intergovernmental transparency’ and adds to policy incoherence between the center and localities. Furthermore, when publication of shares is ad hoc it automatically renders the IGFTS opaque, vulnerable to political negotiations, and disincentivizes good fiscal behavior. Lastly, formula-based shares must be published in forms that are accessible and clear to local government officials and the wider populace. 261. Without transparency local officials and their constituents are not able to ascertain how their share of a particular transfer was determined. Simpler but explicit formulas would facilitate understanding of why different jurisdictions receive different amounts and allow local authorities to assess whether they are being treated fairly. Adherence to formulas reduces the possibilities for capricious political manipulation of transfer allocations. This weakens accountability of civil servants and politicians to their constituents. This also opens the space for political settlements and perverse incentives. The lack of information also hinders performance-based governance, adversely affecting service delivery. The manifestation of this complexity and lack of transparency could be clearly seen in the application of the School Improvement Grant, as outlined in the box below. Box 7: Navigating the School Improvement Grant School Improvement Grants (SIGs) are allocated based on a formula that was devised by the Ministry of Education. The formula sets the minimum amount to be allocated at MWK 600,000 per school. The study found that schools do get less than the minimum and that the formula is not strictly followed. To add to this, none of the district officials were able to produce the formula and it was not applied consistently. It also remains uncertain whether allocation of these funds was done at district level or central level. Lastly, the study showed that the schools’ needs and the formula applied are mostly misaligned. Based on past patterns, schools anticipate that funds will be delayed and therefore adopt coping mechanisms to deal with cash flow problems. They delay expenditures, raise community contributions or sometimes may mis-post expenditures (to different expenditure classifications) to deal with the situation. School managements report varied levels of dissatisfaction with the conditionality of SIG funds. Some school heads find this undermines their priorities, while others find it to be a good guide for their expenditures. The latter group finds it a mechanism to deal with the multifarious demands on the funds. The earmarking provides a mechanism to deal with and deflect some of the demands on SIG funds or general complaints about school conditions. 262. Malawi’s IGFTS scores low on transparency. The complexity of the formula for the horizontal sharing of funds diminishes transparency. As such, by the time the funds reach facilities, it is unclear what basis was used for the calculations. For example, in the Education Sector, the transfer is made through 5 different sub-formulas consisting of 12 different variables (Figure 79, below). The NLGFC publishes technical reports on transfers but these are difficult to interpret by the general public. The complexity of the formula and the absence of published regular analysis of transfers does not make the system very transparent. Figure 79: Composition of Education Transfer Formula Composition of Education transfer formula 100% 5 90% 10 20 80% 5 10 50 70% 60 60% 20 50% 100 5 40% 10 80 30 30% 20% 35 40 10 10% 10 0% 1. Leave Grants 2. Operations 3. School 4. Teaching & 5. Special Needs Maintanance Learning Materials Education No. of primary school Teachers (2010-2013 DEP) Primary school Enrollment (2010-13 DEP) No of schools No. of government schools Pupil to Teacher Ratio Land Area Education in Urban setting (utility bills) Topogrpahy Index Proportion of Children under 14 years Student /school density No. of Special needs Centres Special Need Enrolment (edu 2007-08) 263. The IGFTS is largely opaque. Recipients of the transfers do not clearly comprehend the formula, its variables, their weights or how the shares are determined. The local authority shares are also not strictly in accordance with the formula-based calculation. There are no reports publishing the shares of the transfers or how these are computed. 264. The opacity of the application of the IGFTS formulas can be seen through a deeper analysis of the DDF. The established formula of the DDF is meant to be applied as 30 percent allocated on equal shares principles and 70 percent allocated on four factors carrying equal weight: population, land area, illiteracy and infant mortality. In a review of the actual budget allocation of the DDF in recent years it is apparent that there is a very significant equal share or possible ad-hoc allocations that determine the grant allocations. Per capita allocations for 2018 for the DDF ranged from MWK 51 to MWK 474. 265. This lack of transparency affects some sectors more than others as the formulas for some sectors are more complex than others as seen in Figure 80, below. For example, health and education are far more complex than housing and trade department transfers which are mainly based on just population and an equal share. Absence of ex ante and ex post reporting brings little information into the public domain, preventing informed engagement by stakeholders. Figure 80: Composition of Sector Transfer Formulas 100% 5 5 5 10 90% 5 15 5 25 25 30 80% 5 30 45 50 50 50 40 70% 60 50 60% 50 40 40 50% 100 40% 25 80 70 30% 65 15 55 50 50 50 20% 40 40 35 35 30 10% 15 0% 5 Population Population 15-64 Population of Farming Families % population not yet served Pop acces ext svcs-Nr Seedling Planted Equal share Land area Poverty Head Count Ultra-Poor Persons Iliteracy Rate Crude Death Rate Infant mortality No. of Business Establishments Arable Land Currently in Use Potential Area for Irrigation Area Under Irrigation Nr LA plantation rehab / created Forest reserves & village forest areas Capture Fisheries Acquaculture Fisheries (%) Ornamental Trade (%) OPD Utilisation Rate Number of water points Stunting % below -3SD Bed Capacity Brown Environment (Nr boreholes) Piped inside Dwelling Uinit 266. There is also a lack of transparency in the allocation of donor funded programs. The Health Services Joint Fund (HSJF) is supposed to contribute 80 percent of the total utility bills incurred by each district health office every month. In practice, the contributions of HSJF vary from district to district. The fund also has opaque access conditionalities. In principle, the district health office is supposed to pay 20 percent of the total utility bills monthly as a sign of commitment and submit the payment slips. Districts are not formally oriented as to what the exact conditions for accessing funds are and how to fulfil them. Each district has an HSJF desk officer but the district officials did not have uniform clarity on access and compliance conditions. Districts cite forging personal relationships in the HSJF secretariat as helpful in getting access to funds. Equalization 267. As a principle of IGFTS, equalization can be defined and applied in ‘basic’ or ‘complex’ ways. The basic precept is that transfers should be pro-poor. A complex treatment would mean that in addition to being in favor of the poor the transfers should aim to provide equitable access to services regardless of place of residence, incorporating relative revenue base endowment and thus fiscal capacity, as well as differentials in cost of delivery. It addresses differences in cost of service delivery and tax-based endowments. The objective of equalization is to reduce differences in the ability of local authorities to provide public goods and services, if not eliminate such differences altogether. 268. Weak equalization will adversely affect local service delivery. It generates inefficiency in addition to inequity. Taking local fiscal capacity into consideration would not only render the system equitable but also enhance local fiscal effort. If this does not happen, mobilization of own source revenue will remain low. In the absence of robust equalization, there will be wide variations across jurisdiction in the mix and level of service delivery, and this can sometimes induce migration across jurisdictions creating suboptimal allocation factors. The coping mechanism employed in this case involves promoting ‘intra district service delivery equity’. 269. Equalization in Malawi is weak. Data limitations only allowed for an analysis of per capita equity across districts and not for analysis of relative poverty/endowment and the cost of service delivery in terms of distance, density, or economies of scale. The analysis showed that there are wide differences in per capita transfers to different local authorities. For example, the recipient of the largest education transfers, Mulanje District Council, received MWK 722 in 2018 compared to Mwanza District Council, which received MWK 28 per capita. Similarly, in health, Likoma District Council received MWK 3,151 per capita, 21 times the MWK 151 per capita received by Mwanza District Council. A further analysis would show that the equalization effect of various transfers moves in opposite directions. For example, while education transfers have an equalizing effect on top of own sourced revenue transfers for agriculture, DDF, CDF, boreholes and roads are counter equalizing (see Figure 81, below). Figure 81: Equalization Effect of Transfers 1.1 Cumulative coefficient of variation 1 0.9 Not equalizing 0.8 Equalizing 0.7 0.6 0.5 OSR +Education +Health +Boreholes +CDF +Agric +DDF Additional Sectoral Transfer Conclusions and Recommendations 270. Malawi’s IGFTS is in need of reform. The present system reduces local governments to de facto spending units of the national Government and, as such, it does not generate any of the putative benefits of decentralized governance in terms of allocative efficiency, or strengthened accountability, be this between citizens and the state, between local government sector managers and staff at the facility level responsible for delivering services, or between elected leaders and administrators. Put differently, the current system provides little basis for local discretion and as a result, local resource allocation and service delivery remain suboptimal. It is simply not possible for services to be ‘expressed, defined, and delivered locally,’ in line with the subsidiarity principle that underpins successful devolutions. Local governments in Malawi operate as extensions of delegated or deconcentrated mandates of central ministries. 271. In considering the scope and key points of orientation for such an exercise, two points should be highlighted: The presence of multiple transfers that are for specific purposes, leaving little room for local budgetary decisions to address local need. Second, Malawi has weak revenue decentralization with own source revenue for rural councils comprising around 7 percent of total revenue and 18 percent for city councils. In view of this, the absence of a general-purpose grant precludes fiscal space for local political priorities that are highlighted through local mechanisms. 272. With reference to the five parameters of the IGFTS, the areas that should be prioritized for reform are recounted here along with an indication of their relative importance for governance and service delivery: (a) Reduce the fragmentation of development transfers. High Importance: While the vertical imbalance of the IGFTS remains stark, a key first step to correcting this would be a reduction in the fragmentation of development transfers to local authorities and a commitment to the DDF as a consolidated fund for discretionary development expenditures. Efforts to reduce this fragmentation and increase the quantity of discretionary development funds would be an essential part of any credible signaling by the Government that it is committed to reforming inter-governmental relations. (b) Improve the predictability of annual and monthly transfers. Medium importance: A commitment should be made to improved cash flow management. That in turn should generate sufficient predictability for local authorities to plan and execute budgets optimally. In addition to year-end predictability, month to month predictability should be the focus. Floors and ceilings for variations from the budgeted shares should be specified, creating an accountability mechanism for the Ministry of Finance as well. NLGFC’s mandate to report on predictability should be reiterated. (c) Reduce the uncertainty in timeliness of transfers. Medium Importance: The flow of funds system should be made transparent and NLGFC should be given a role in monitoring and reporting on timeliness. However, it is crucial to ensure that districts are more accountable for the consequences of their cash management decisions on the operations of schools and clinics. (d) Increase the transparency of the IGFTS. Medium importance: Support should be agreed on the NLGFC’s capabilities and incentives offered for performance for annual collation, analysis, reporting, and a public forum for the implementation of the Fiscal Decentralization strategy. Improvements in transparency should be feasible through simplification of the transfer formula scheme as well as instituting a role for periodic NLGFC reports on various parameters of the IGFTS. Application of formula for DDF and IDF should be ensured as an immediate priority. A good data regime, including completion of work on IFMIS, should be prioritized. (e) Strengthen the equalization effects of the transfer system. High importance: Equalization should be strengthened through the revision of formulas and inclusion of rational expenditure needs and fiscal capacity measures in the formulas. This should be achievable over time, starting with an explicit policy on equalization with a strategy for implementation. Annexes Education Annexes Education Annex 1. Additional tables and figures Figure A1: Trend in out-of-school rate for school-age cohort children (%) by gender, area and household wealth quintile 25 20 20 20 18 17 16 17 15 15 15 16 13 13 14 14 14 14 15 12 13 12 13 11 10 10 7 8 6 4 4 5 3 - 2010 2017 2010 2017 2010 2017 2010 2017 2010 2017 2010 2017 2010 2017 2010 2017 2010 2017 2010 2017 2010 2017 2010 2017 2010 2017 2010 2017 Male Female Total Urban Rural Poorest Richest Richest Male Female Total Urban Rural Poorest Gender Area Quintile Gender Area Quintile Primary school age Lower secondary school age Source: Authors’ calculations based on IHS 2010 and 2016/17 Figure A2: Share of Standard 4 who can perform basic Mathematics and English tasks 7-4 7+7+7+7+7+7 12 X 6 953-164 70/5 100+20 201 X 8 match simple noun to picture identify missing letter in sequence reading comprehension complete single noun from picture choose correct format of "to be" plural form of word "mango" 0% 20% 40% 60% 80% Source: Author’s calculations based on MLSS baseline (2016). Note: initial analysis suggests similar levels for MLSS midline (2018/19) Figure A3. Enrollment growth and trends by secondary school type in Malawi Average Enrollment Growth Rates Trends in Enrollment by School Type (2013- (2013-2018) 2018) 200,000 25% 20% 150,000 15% 100,000 11% 50,000 5% 1% - 2013 2014 2015 2016 2017 2018 -5% -1% 0% CDSS Approved CDSS Not Approved Govt Boarding Govt Day -15% -13% ODSS Private Source: EMIS Figure A4: Students Performance at the Malawi Secondary Certificate of Education, 2017, overall (left) and math (right) MSCE Pass Rates - 2017 MSCE Pass Rates for Mathematics - 2017 90 80 80 70 70 60 60 50 50 40 40 30 30 20 20 10 10 0 0 Male Female Total Male Female Total Source: MANEB Figure A5: Rate of returns on education by education level, gender, employment sector, and employer type 300% 262% 250% 199% 203% 201% 192% 200% 172% 161% 162% 166% 144% 150% 125% 114% 116% 105% 101% 100% 83% 76% 66% 70% 67% 57% 48% 40% 42% 46% 43% 50% 33% 25% 27% 29% 14% 17% 0% National Male Female Priavte Public Agriculture Industry Services Some primary Some lower secondary Some upper secondary Post secondary Source: Authors’ calculations based on IHS 2016/17 Figure A6: Comparison of public expenditure on education as share of GDP and total public spending for select countries (percent) 7 30 6.4 6.1 5.8 5.6 5.6 6 25 5.4 5.1 5.0 5.0 4.8 4.8 4.7 4.6 4.6 5 % of total spending 20 4.2 3.7 3.6 4 3.5 %of GDP 3.2 3.1 15 2.8 2.7 2.7 2.6 3 2.4 2.1 10 2 5 1 0 0 Mauritania Togo Gabon SSA Average Zimbabwe Senegal Sierra Leone Burundi Côte d'Ivoire Ghana Benin Gambia Liberia Niger Zambia South Africa Burkina Faso Namibia Ethiopia Malawi Mozambique Guinea Uganda Rwanda Kenya Djibouti Education share in GDP Education share in total public spending Source: Authors’ calculation based on the Ministry of Finance (MoF) Budget (2017/18) and United Nations Educational, Scientific and Cultural Organization Institute of Statistics (UIS) for comparison countries, 2017 or the latest available data. Figure A7A: Share of public expenditure by level of education (percent of public education expenditure) Primary Secondary Tertiary Source: Based on UIS 2017 or the latest available data. Note: Some countries do not add up to 100 percent due to other forms of education other than the three presented above such as non-formal education or pre-primary school not under primary. B: Per student public spending, an international comparison (% of GDP per capita) Primary and Secondary Tertiary South Africa 18 20 Malawi 820 Burkina Faso 16 19 Burkina Faso 351 Zimbabwe 14 22 Burundi 309 Niger 13 16 Ethiopia 266 Côte d'Ivoire 13 20 Burundi 13 31 Niger 260 Mozambique 12 43 SSA average 205 Senegal 12 11 Zimbabwe 161 SSA average 11 22 Côte d'Ivoire 150 Mauritania 10 14 Senegal 143 Malawi 10 41 Mozambique 140 Benin 9 11 Rwanda 108 Ethiopia 8 17 Guinea 7 8 Mauritania 96 Ghana 6 19 Guinea 90 Rwanda 5 37 Benin 73 Ghana 55 South Africa 47 Source: Authors’ calculation based on the Ministry of Finance, Economic Planning, and Development (MoFEPD) Budget 2017/18, EMIS 2018, The World Bank Group for Malawi and UIS for comparison countries, 2017 or latest. Figure A8: A: Breakdown of household education payment by level of education, and school type Tuition/Fees Books Uniform Boarding Transport Extra tutoring Others 6% 10% 5% 8% 9% 2% 8% 0% 1% 10% 7% 1% 27% 1% 6% 10% 9% 23% 11% 12% 21% 5% 18% 0% 17% 0% 11% 26% 30% 39% 9% 43% 73% 72% 62% 59% 39% 39% 30% 24% TERT IARY SECOND ARY SECOND ARY TOTAL PRIMARY PRIMARY PRE-PRY PRE-PRY PUBLIC PRIVATE B: Breakdown of household education payment by household wealth quintile and school type Tuition/Fees Books Uniform Boarding Transport Extra tutoring Others 8% 7% 8% 8% 9% 8% 8% 6% 8% 6% 11% 11% 0% 0% 0% 1% 1% 0% 1% 2% 1% 1% 2% 13% 2% 16% 21% 17% 11% 13% 24% 25% 22% 15% 16% 26% 13% 22% 40% 24% 30% 32% 49% 45% 41% 36% 30% 41% 65% 49% 52% 38% 40% 38% 26% 25% 27% 29% 31% 27% Q1 Q2 Q3 Q4 Q5 TOTAL Q1 Q2 Q3 Q4 Q5 TOTAL PUBLIC PRIVATE Source: Authors’ estimations based on IHS 2016/17 Figure A9: Factors affecting children being out of school 30 Not interested/Lazy Costs Married/Pregnant Illness/Disability 25 Parents/Family Support Others 20 15 16 7 6 3 17 5 9 2 6 7 10 18 18 25 16 23 19 27 5 56 0 41 39 39 Urban Malawi Rural Male Female Q1 Q2 Q3 Q4 Q5 MALE FEMALE MALE FEMAL E PRIMARY SECOND ARY Source: Authors’ estimations based on IHS 2016/17 Table A1: Functional allocation of public education budget and execution rates Budget allocation by function (%) 2015 2016 2017 2018 2019 Compensation for Employees 48% 53% 54% 49% 49% Goods and Services 11% 8% 6% 6% 6% Grants 29% 28% 21% 21% 19% Equipment 1% 1% 1% 1% 1% Development 11% 10% 17% 24% 25% Execution rate (%) Compensation for Employees 140% 102% 98% 97% 102% Goods and Services 46% 47% 42% 55% 49% Grants 116% 73% 104% 94% 112% Equipment 36% 45% 55% 49% 39% Development 40% 46% 46% 48% Total 110% 84% 81% 82% 87% Source: Authors’ calculations based on Ministry of Finance (MoF) Budget (2014/15- 2018/19) Box A1. Education financing in Malawi Source: Created by the team based on the information collected during the study Education expenditure in Malawi is largely divided between the Ministry of Education, Science and Technology (MoEST) and Local Government Authorities (LGAs), with primary financed largely at LGA level. The budget of MoEST includes salaries for all secondary school teachers, operational budgets for headquarters and agencies, all secondary schools, the six education divisions, Teacher Training Colleges (TTCs), and development projects at primary, secondary, and higher education levels and teacher education. LGA budgets include operational costs for all primary schools and, since 2016/17, salaries for all primary school teachers. Education expenditure is primarily classified into recurrent expenditure, comprised of personal emoluments (PE), and ORT; and development expenditure, spread across development budget Part I which is donor financed, and development budget Part II which is financed by the Government. Funding can also accrue to local level education through the CDF. For this study, data was primarily available on Vote 250 and other national subventions, but limited on use of LGA-level finance (see Section III). Source: Joint Sector Review (JSR) 2018/19 (Government of Malawi) Education Annex 2. Fiscal simulation methodology The authors, in consultation with and inputs from MoEST, developed a simulation tool to estimate the fiscal impacts of various policy parameters with relevance to primary education. The tool presents estimates of enrollment growth to 2030, employing UN population projections to estimate the rate of growth in the population entering school age; specifically, we use population projections from the United Nations Department of Economic and Social Affairs, adjusted to reflect school-age population as reported in the UNICEF Multiple Indicator Cluster Survey. The tool then uses recent trends in student intake into Standard 1, and promotion and retention into grades, to model students’ progress through the system. The tool then estimates the annual cost of primary education, based on baseline levels of inputs (e.g. pupil-classroom ratio, pupil-teacher ratio (PTR)) drawn from the most recent Annual School Census. Unit costs are drawn from official MoEST sources as well as independent sources including MLSS. Fiscal projections can be based on maintenance of current input standards, or improvements specified by the user (e.g. lowering average PTRs to the policy standard of 60). The tool captures the main recurrent and development costs pertaining to primary education , including teacher salaries and construction of classrooms and latrines; but does not include school grant finance; education administration and management costs such as inspections, examinations, and district and sub-district officials; MoEST operating costs; or teacher training. Table A2-1. Primary simulation model parameters and assumptions Indicator Baseline (2018/19) Target for 2025* Inputs parameters Pupil-teacher ratio 80 60 Share of qualified teachers 98% No change Pupil-classroom ratio 150 60 Pupil-latrine ratio 91 30 Enrollment Assumptions Gross Intake Rate 155% No change Share of private intake 2% Promotion rate Std.1 61% Std.2 74% Std.3 68% Std.4 73% No change Std.5 68% Std.6 69% Std.7 64% Repetition rate Std.1 28% Std.2 22% Std.3 22% Std.4 19% No change Std.5 21% Std.6 19% Std.7 19% Std.8 20% Dropout rates (for reference) Std.1 11% Std.2 4% Std.3 10% Std.4 8% No change Std.5 12% Std.6 12% Std.7 18% *Model assumes steady trend from 2018/19 to 2025 target. Social Protection Annexes Social Protection Annex 1: Social protection programs of focus Coverage Cost Program Start Area Program Name Program Description Target Population Targeting Approach (% Pop, (% GDP, Type Year 2019) 2019) Self-targeting + Communities and Provision of high-quality school meals for categorical / Nutrition and learners and pupils at primary schools in selected automatic (school access to primary Food 2012 preschoolers in 2 1.3% 0.04% districts to support regular participation children receive education (NAPE) (BMZ128) plus 5 districts in schools. meals while they're (EU) in school) Aims to reduce dropout rate, particularly among girls and orphans in standards 5– School Meal Children in food Self-targeting + 8; promote regular attendance; improve Food 2007 12.6% 0.1% Programme (SMP) insecure districts categorical children’s capacity to concentrate and learn; and increase enrollment rates. Malawi Social Ultra-poor but not Geographic + Social Action Fund Public The program runs in two 24-day cycles Public labor constrained 1996 community-based + 0.1% 0.02% Assistance Works Programme per year works households, active in all self-targeting+ PMT (MASAF – PWP) 28 districts Geographic (mostly) + Aims to reduce immediate need for food Food For Asset Public Food insecure community-based security while building assets to improve 2008 3.7% 0.2% (FFA – PWP) works individuals (define local long-term food security and resilience development need & participation) Social Cash Started in 2006 in Mchinji. Targets ultra- Ultra-poor and labor Transfer poor and labor constrained households constrained Community-based Programme - (about 10% of household population). Cash 2006 6.7% 0.4% households (10 % of + PMT Mtukula Pakhomo Beneficiaries receive monthly cash grant the total population) (SCTP) averaging MWK 7,000/HH/month. 128 The Federal Ministry of Economic Cooperation and Development (German: Bundesministerium für wirtschaftliche Zusammenarbeit und Entwicklung), abbreviated BMZ, is a cabinet-level ministry of the Federal Republic of Germany. Coverage Cost Program Start Targeting Area Program Name Program Description Target Population (% Pop, (% GDP, Type Year Approach 2019) 2019) STEP consists of four components, three major ones and one auxiliary. First, via UNESCO, the program analyses the technical and vocational Youth, with a Self-targeting education and training (TVET) sector, for Skills and focus on (people sign up for example by reviewing curricula for TET Technical In kind: vulnerable a skills building / Labor programs. Second, via the National Authorizing 0.07% Education Skills 2016 groups, for technical NA Market Officer Support Unit, the program aims to (2016) Programme - training instance women education refurbish 28 technical colleges spread across 20 STEP (TEVET) or school program, firms districts. Third, the program provides grants to dropouts apply for grants) private TVET providers. Fourth, the program retains some of its budget to commission studies on relevant issues. National The NPS is a mandatory, defined contribution 0.2% 1.2% Pension System Pensions 2011 Elderly Categorical pension fund for the private sector. (2016) (2016) Social (NPS) Insurance Civil Service CSPS is a pay-as-you-go defined benefit pension Pension Scheme Pensions fund for public sector employees. (CSPS) Decent and Affordable The program provides subsidized cement, In kind: Low income 0.1% 0.1% Housing Subsidy corrugated iron sheets and other building 2014 Community based Housing households (2018) (2018) Programme materials for low income households General (DAHSP) Subsidies Farm Input Objective: To achieve food self-sufficiency and Community based Subsidy In kind: increased income in poor households through 2005 - no categorical 21.7% 0.6% Programme Agriculture increased maize and legume production and no PMT (FISP) Coverage Cost Program Start Area Program Name Program Description Target Population Targeting Approach (% Pop, (US$, Type Year 2016) 2016) Humanitarian assistance in Food insecure households, with a Community based / top-down - The Food insecurity response to MVAC preference for vulnerable groups MVAC assessment determines the response based assessment of food (headed by females, children, or number of people in need of food on assessment insecurity, providing elderly; households with assistance down to the Traditional by Malawi emergency food Food 2013 malnourished children, or people Authority level. Within Traditional 32.8 5.8 Vulnerability relief, either in cash with HIV/AIDS; households caring for Authorities, the District Executive Assessment or as in-kind orphans; households with crop Committee selects which villages to Committee transfers if local failure for two or more successive target. Within those, communities (MVAC) markets are deemed years are asked to target households. Humanitarian Assistance too thin. Social Protection Annex 2: Expenditure by Program (in US$ million) and Source of Financing Start 2011- 2012- 2013- 2014- 2015- 2016- 2017- 2018- 2019- Programs Funding Source Year 12 13 14 15 16 17 18 19 20 Skills and Technical Education Programme - European Union 2016 - - - - - 3.8 NA NA NA STEP (TEVET) National Pension System (NPS) Government of Malawi 2011 78.4 54.5 54.0 47.4 80.7 63.9 NA NA NA Nutrition and access to primary education - European Union; GIZ 2012 - 1.3 1.3 1.3 1.1 1.7 3.7 2.7 3.5 NAPE (SMP) World Bank + EU (90%) MASAF Public Works Programme (PWP) 1996 4.9 2.7 21.9 16.3 4.8 22.3 10.9 15.1 1.3 Government. of Malawi (10%) World Bank + EU + KfW + UNICEF Social Cash Transfer Programme - Mtukula + Irish Aid (90%) 2006 3.4 0.4 3.4 7.2 5.6 14.6 18.8 30.2 34.5 Pakhomo (SCTP) Government of Malawi (10%) Decent and Affordable Housing Subsidy Government of Malawi 2014 - - - - 4.6 6.3 8.2 6.9 NA Programme (DAHSP) Farm Input Subsidy Programme (FISP) Government of Malawi 2005 153.2 211.6 164.1 119.7 129.5 69.7 46.9 56.5 48.9 Humanitarian assistance in response to DFID, Germany, Norway, Irish Malawi Vulnerability Assessment Committee 2013 - 28.0 55.0 10.3 113.9 314.6 112.9 40.1 19.9 Aid, WFP (MVAC) assessment Food for Assets (FFA) public works WFP 2008 NA NA NA 1.0 1.8 17.5 21.5 23.2 15.0 School Meals Programme (SMP) WFP 2007 15.0 13.0 14.2 16.5 20.4 17.2 18.3 11.9 7.4 Social Protection Annex 3: ASPIRE Program classifications SPL Area Program classification Program Name Contributory pensions Pensions Social Insurance Other social insurance NA Passive labor market programs NA Labor Market Active labor market programs NA Unconditional cash transfers, allowances, last resort programs NA Conditional cash transfers NA Non-contributory social pensions NA Free maize Free food (other than maize) Food and in-kind programs Likuni Phala to Children and Mothers (Targeted Nutrition Program) Social Assistance Supplemental feeding for malnourished children School feeding School feeding program MASAF, PWP Public works and food for work Inputs for work program Fee waivers and targeted subsidies NA Scholarship for secondary education Other Social Assistance Other Intergovernmental Fiscal Transfers Annexes Intergovernmental Fiscal Transfers Annex 1: Flow of Funds Intergovernmental Fiscal Transfers Annex 2: Summary of Transfers in the IGFTS Start Which ministry Sector Objects of expenditure date prescribes policy Teaching and Education methods learning materials advisory services for primary schools Education Science (education standards), are currently Education 2005 and Technology and management and procured directly by administration of schools primary schools and education services. (since 2016 2017 financial year) Over the years funds Primary health care, for drugs were sent district hospitals, rural to the districts but Health and hospitals, health center Health 2005 this was Population Services facility, clinics, public recentralized in health inspection, and 2014/15 financial health technical support. year Handing over all Rural Development Project (RDP), Extension Planning Area (EPA) and section Part of the district level extension related council’s budget is functions to district used for the Farm assemblies, farmer group Agriculture and Food Input Subsidy Agriculture 2005 mobilization and training, Security Program (FISP) in-service staff training, activities such as message development, verification of district-level inspection beneficiaries. and monitoring functions, and district-level coordination of agricultural services. Local Government Provide resources for DDF 2015 and Rural implementation of Development development plans These were not All Members of devolved from any Local Government Parliament to implement ministry and there is CDF 2005 and Rural micro projects at no devolution plan Development community level associated with these functions. These were not Enable urban councils to devolved from any Local Government implement infrastructure ministry and there is IDF 2005 and Rural improvement in their no devolution plan Development respective councils associated with these functions. Start Which ministry Sector Objects of expenditure date prescribes policy These were not devolved from any Local Government Enable district councils to ministry and there is Boreholes 2016 and Rural drill 4 boreholes per no devolution plan Development constituency associated with these functions. These were not For councils to manage devolved from any Local Government their daily operations, pay ministry and there is GRF 2005 and Rural councilors honoraria no devolution plan Development (20%), conduct council associated with meetings these functions. Management of water resources, provision of technical services including: siting boreholes, supervision of drilling, tenders; rehabilitation of small Irrigation and water dams; protection of Water 2013 development catchment areas, water supply and sanitation, maintenance of rural piped water systems, boreholes and community-based management of water resources Provision of office and housing accommodation, provision of rural housing, provision of Housing 2006 Land and housing housing plots, enforcing building standards and regulations, and provision of valuation services. Business licensing, Commerce trade Trade 2006 registration of garages and industry and related businesses Youth economic empowerment, youth Youth 2013 Labour participation and leadership, youth health Economic empowerment Gender Community of vulnerable groups, Gender 2006 Development and provision of social social welfare welfare services Start Which ministry Sector Objects of expenditure date prescribes policy Community was previously under Community Information and Provision of community gender and since 2006 Development Civic Education development services 2018/19 has different cost centers Employment and labour relations, health and Labour 2013 Labour occupational safety and workers compensation Talent identification, Sports 2013 Sports mass participation in sports Energy extension services, forestry extension services, Climate change and mining extension Environment 2013 environment services, fisheries extension services, and environmental extension services. Agriculture and Food Management of fisheries Fisheries 2013 Security at district level Provision of technical Irrigation and water Irrigation 2013 support in small scale development irrigation Home Affairs and Vetting of passport Immigration 2013 Internal Security applications Office of the Management of vital NRB 2013 President and statistics, of births and Cabinet certificates Notes: * Councils are required to spend all the resources for a particular function within the sector; no inter borrowing. These resources are predominantly for operations. References Abdoulayi, S et al. “Malawi Social Cash Transfer Program: Midline Impact Evaluation Report.” Working Paper. Abramsky, T., K. Devries, L. Kiss, J. Nakuti, N. Kyegombe, E. Starmann, B. Cundill, L. Francisco, D. Kaye, T. Musuya, L. Michau, and C. Watts (2014) “Findings from the SASA! Study: A Cluster Randomized Controlled Trial to Assess the Impact of a Community Mobilization Intervention to Prevent Violence against Women and Reduce HIV Risk in Kampala, Uganda.” BMC Medicine 12:122. Arndt, C., K. Pauw, and J. Thurlow. 2016. “The Economy-wide Impacts and Risks of Malawi's Farm Input Subsidy Program.” American Journal of Agricultural Economics, Volume 98, Issue 3, April 2016, pages 962– 980. Baird, S., C. McIntosh, and B. Özler. 2011. "Cash or Condition? Evidence from a Cash Transfer Experiment." Quarterly Journal of Economics, 126: 1709–53. Baird, S., F.H.G. Ferreira, B. Özler, M. Woolcock. 2013. “Relative Effectiveness of Conditional and Unconditional Cash Transfers for Schooling Outcomes in Developing Countries: A Systematic Review.” Campbell Systematic Reviews 2013:8. Banerjee A., E. Duflo, N. Goldberg, D. Karlan, R. Osei, W. Parienté, J. Shapiro, B. Thuysbaert, C. Udry. 2015. “A multifaceted programme causes lasting progress for the very poor: Evidence from six countries.” Science 348, 1260799. Beegle, K., E. Galasso, and J. Goldberg. 2015. "Direct and Indirect Effects of Malawi’s Public Works Programme on Food Security." Policy Research Working Paper Series 7505. Washington, DC: The World Bank. Beegle, K., A. Coudouel, and E. Monsalve. 2018. “Realizing the Full Potential of Social Safety Nets in Africa.” Africa Development Forum; Washington, DC: World Bank. Beierl B. 2017. NSSP Discussion Note on Public Works. Ministry of Finance, Economic Planning and Development Poverty Reduction and Social Protection Division P.O. Box 30136, Lilongwe 3, Malawi Boone, R., K. Covarrubias, B. Davis, and P. Winters. 2013. "Cash Transfer Programmes and Agricultural Production: The Case of Malawi." Agricultural Economics. 44: 365–78. Botea, I., S. Chakravarty, and S. Haddock, and Q. Wodon (2017) Interventions Improving Sexual and Reproductive Health Outcomes and Delaying Child Marriage and Childbearing for Adolescent Girls, Ending Child Marriage Notes Series. Washington, DC: The World Bank. Bridges K. and M. Woolcock. 2017. “How (Not) to Fix Problems That Matter. Assessing and Responding to Malawi’s History of Institutional Reform.” World Bank. Policy Research Working Paper 8289 Canning, D., S. Raja, and A. S. Yazbeck (2015) Africa’s Demographic Transition: Dividend or Disaster? Washington, DC: The World Bank. Chibwana, C. and M. Fisher. 2011. “The impacts of agricultural input subsidies in Malawi.” IFPRI Policy Note. Chinsinga, B. 2012. “The Political Economy of Agricultural Policy Processes in Malawi: A Case Study of the Fertilizer Subsidy Program.” Future Agricultures Working Paper, 39. Chinsinga, B. 2009. “Political Economy of Cash Transfers In Malawi.” ODI Report. Chirwa, E., Matita, M., & Dorward, A. 2011. “Factors influencing access to agricultural input subsidy coupons in Malawi”. Future Agricultures Consortium Working Paper No. 027 Coady, D., V. Flamini, and L. Sears. 2015. “The Unequal Benefits of Fuel Subsidies Revisited: Evidence for Developing Countries.” IMF Working Paper, WP/15/250. International Monetary Fund, Washington DC. Covarrubias, K., B. Davis & P. Winters. 2012. “From protection to production: productive impacts of the Malawi Social Cash Transfer scheme”. Journal of Development Effectiveness 4:1, 50-77. Dang, Hai-Anh H., and Andrew L. Dabalen. 2017. “Is Poverty in Africa Mostly Chronic or Transient? Evidence from Synthetic Panel Data.” Policy Research Working Paper 8033, World Bank, Washington, DC. Davis B., S. Handa, N. Hypher, N.W. Rossi, P. Winters, and J. Yablonski. 2016. “From Evidence to Action The Story of Cash Transfers and Impact Evaluation in Sub-Saharan Africa.” The Food and Agriculture Organization of the United Nations (FAO), The United Nations Children’s Fund (UNICEF), Oxford University Press. de Hoop, J. Valeria Groppo, & Sudhanshu Handa. 2017. “Household Micro-entrepreneurial Activity and Child Work: Evidence from Two African Unconditional Cash Transfer Programs.” Dionne, Kim Yi, Eric Kramon, & Tyson Roberts. 2013. “Aid Effectiveness and Allocation: Evidence from Malawi.” Evans, D.K.; A. Popova. 2014. “Cash transfers and temptation goods: a review of global evidence (English) .” Policy Research working paper; no. WPS 6886; Impact Evaluation series; no. IE 127. Washington, DC: The World Bank. Field, E. and A. Ambrus (2008) “Early Marriage, Age of Menarche, and Female Schooling Attainment in Bangladesh.” Journal of Political Economy 116(5): 881-930. [FINDEX] Global Financial Inclusion Index Database 2014. Malawi. [FAO] Food and Agriculture Organization. 2017. “Impacts of modifying Malawi’s farm input subsidy programme targeting.” Agricultural Development Economics Working Paper 17-05. Rome, FAO. Government of Malawi. 2016a. “2016/2017 Food Insecurity Response Plan.” Malawi. Government of Malawi. 2016b. “Review of the Malawi National Social Support Programme.” Malawi. Government of Malawi. 2019a. Malawi 2019 Floods Post Disaster Needs Assessment Report . Malawi. Government of Malawi. 2019b. “2018 Malawi Population and Housing Census - Main Report.” National Statistical Office. Malawi. Hallegate, S., L. Bonzanigo, T. Kane, J. Rozenberg, D. Treguer and A. Vogt-Schlib. 2016. “Shock Waves: Managing the Impacts of Climate Change and Poverty”. The World Bank, Washington DC. Hamer, S. and J. Seekings. 2017. “Social protection, electoral competition, and political branding in Malawi”. United Nations University WIDER Working Paper 99 Holmes, R. and Costella, C. with Bailey, M., Kruczkiewicz, A., Poulter, R. Sharp, K. and Scott, L. 2017. “Towards A Shock Sensitive Social Protection System for Malawi”. London: ODI and The Hague: Red Cross Climate Centre. [ILO] International Labor Organization. 2016c. “Extracts from the Full Report: Assessment of Social Protection Programmes and Costing of Policy Options—Programme Specific Report: Village Savings and Loans.” [ILO] International Labor Organization. 2016d. “Extracts from the Full Report: Assessment of Social Protection Programmes and Costing of Policy Options—Programme Specific Report: Public Works Programs.” [ILO] International Labor Organization. 2016a. “Social protection in Malawi. Assessment based national dialogue brief.” [ILO] International Labor Organization. 2016b. “Review of the Malawi National Social Support Programme.” [ILO] International Labor Organization. 2018. “Women and Men in the Informal Economy: A Statistical Picture”. [IMF] International Monetary Fund. 2015. Selected Issues Paper: Reforming the Farm Input Subsidy Program. IMF Country Report No. 15/346. Kalebe-Nyamongo, C and H. Marquette, 2014. “Elite Attitudes Towards Cash Transfers and the Poor in Malawi.” DLP Research Paper 30. Birmingham: University of Birmingham. Kilic, T., E. Whitney, P. Winters, 2013. “Decentralized Beneficiary Targeting in Large-Scale Development Programs: Insights from the Malawi Farm Input Subsidy Program.” World Bank Policy Research Working Paper No. 6713. Koch, C. 2015. “Political Economic Analysis of Malawi.” GIZ. Lea, N. and L. Hanmer. 2009. “Constraints to Growth in Malawi.” World Bank Policy Research Working Paper No. 5097. Male, C., and Q. Wodon (2018) “Girls’ Education and Child Marriage in West and Central Africa: Trends, Impacts, Costs, and Solutions.” Forum for Social Economics, 47(2): 262-74. Margolies, A., N. Aberman and A. Gelli, 2017. “Traditional Leadership and Social Support in Southern Malawi.” IFPRI Strategy Support Team, Malawi. McCarthy, N.; T. Kilic, A. de la Fuente, J. Brubaker. 2017. “Shelter from the Storm?: Household-Level Impacts of, and Responses to, the 2015 Floods in Malawi.” Policy Research Working Paper; No. 8189. The World Bank, Washington, DC. McConnell, C., and M. Mupuwaliywa (2016) Keeping Girls in School: Situation Analysis for Malawi. Country Policy Brief, Washington, DC: The World Bank. Miller, C., M. Tsoka, and K. Reichart. 2011. “The impact of the Social Cash Transfer Scheme on food security in Malawi.” Food Policy, 36. 230-238. Miller, C., M. Tsoka, and K. Reichart. 2010. “Targeting Cash to Malawi’s Ultra-Poor: A Mixed Methods Evaluation.” Development Policy Review, 2010, 28 (4): 481-502. [MoFEPD] Ministry of Finance, Economic Planning and Development. 2016. “A Stakeholder-Driven Review of the Design and Implementation of the Malawi National Social Support Programme (2012-2016).” Ministry of Finance, Economic Planning and Development. 2016b. Harnessing the Demographic Dividend to Accelerate Socio-economic Transformation and Economic Development in Malawi . Noemi Pace, Silvio Daidone, Benjamin Davis, Sudhanshu Handa, Marco Knowles & Robert Pickmans (2017): One Plus One can be Greater than Two: Evaluating Synergies of Development Programmes in Malawi, The Journal of Development Studies, DOI: 10.1080/00220388.2017.1380794 Lindert, K., C. Andrews, C. Msowoya, B. Paul, E. Chirwa, A. Mittal. 2018. “Rapid Assessment of Malawi’s Unified Beneficiary Registry”. Social Protection Discussion Series. World Bank, Washington DC. Nguyen, M. C., and Q. Wodon (2014) Impact of Child Marriage on Literacy and Educational Attainment in Africa, Background Paper for Fixing the Broken Promise of Education for All. Paris and New York: UNESCO Institute of Statistics and UNICEF. Ralston, L., C. Andrews and A. Hsiao. 2017. “The Impacts of Safety Nets in Africa : What Are We Learning?” Policy Research Working Paper; No. 8255. Washington, DC: The World Bank. Resnick D. (2012). Two Steps Forward, One Step Back The Limits of Foreign Aid on Malawi’s Democratic Consolidation. UN WIDER, Working Paper No. 2012/28 Tambulasi, Richard IC. 2010a. ‘Critical Perspectives on Public Sector Reforms: An Introduction’. In Reforming the Malawian Public Sector, 69. Taylor, J.E., J. Kagin, M. Filipski, and K. Thome. 2013. "Evaluating General Equilibrium Impacts of Kenya’s Cash Transfer Programme for Orphans and Vulnerable Children (CT-OVC)." Working Paper. Rome: United Nations Food and Agriculture Organization. Taylor, J.E., K. Thome, and M. Filipski. 2014. "Evaluating local general equilibrium impacts of Lesotho’s child grants program." Working Paper. Rome: United Nations Food and Agriculture Organization. Thome, K., J.E. Taylor, B. Davis, S. Handa, D. Seidenfeld, and G. Tembo. 2014. " Local Economy-wide Impact Evaluation (LEWIE) of Zambia’s Child Grant Program." Working Paper. Rome: United Nations Food and Agriculture Organization. Thome, K., J.E. Taylor, J. Kagin, B. Davis, R.D. Osei, and I. Osei-Akoto. 2014. "Local Economy-wide Impact Evaluation (LEWIE) of Ghana’s Livelihood Empowerment Against Poverty (LEAP) Program." Working Paper. Rome: United Nations Food and Agriculture Organization. [UNDAC] United Nations Office for the Coordination of Humanitarian Affairs. 2015. “Malawi 2015 Floods UNDAC Assessment Results.” [UNDESA] United Nations Department of Economic and Social Affairs, Population Division. 2017a. Population Estimates and Projections (1950-2100). [UNDESA] United Nations, Department of Economic and Social Affairs, Population Division. 2017b. “ World Population Ageing 2017.” New York. Wodon, Q., C. Montenegro, H. Nguyen, and A. Onagoruwa (2018) “Missed Opportunities: The High Cost of Not Educating Girls.” The Cost of Not Educating Girls Notes Series, Washington, DC: The World Bank. Wodon, Q., C. Nguyen and C. Tsimpo (2016) Child Marriage, Education, and Agency in Uganda, Feminist Economist, 22(1): 54-79. Wodon, Q., C. Male, and A. Onagoruwa. 2020. “A Simple Approach to Measuring the Share of Early Childbirths Likely Due to Child Marriage in Developing Countries.” Forum for Social Economics 49(2):166- 179. [WBG] World Bank Group. 2012. “World Bank Social Protection and Labor Strategy 2012-2022.” Washington, DC: World Bank Group. [WBG] World Bank Group. 2013. Malawi Public Expenditure Review. Poverty Reduction and Economic Management 4 Africa Region Report No. 79865 – MW [WBG] World Bank Group. 2013. Malawi Economic Monitor. [WBG] World Bank Group. 2014. Feature Story. “Reducing the Impact of Natural Disasters in Malawi: Empowering Citizens and Taking Charge.” January 29, 2014. Accessed March 28, 2018. [WBG] World Bank Group. 2016. Agricultural Support and Fiscal Management Development Policy Financing. Project Document. [WBG] World Bank Group. 2016. “Malawi drought 2015-2016: post-disaster needs assessment (PDNA).” Washington, DC: The World Bank. [WBG] World Bank Group. 2016. “Malawi Poverty Assessment”. Washington DC. [WBG] World Bank Group. 2016. Malawi Economic Monitor. [WBG] World Bank Group. 2018a. “The State of Safety Nets 2018”. Washington DC. [WBG] World Bank Group. 2018b. Malawi: Systematic Country Diagnostic: Breaking the Cycle of Low Growth and Slow Poverty Reduction. Washington, DC: World Bank Group. [WBG] World Bank Group. 2018c. “Malawi Economic Monitor. Realizing Safety Nets' Potential." [WBG] World Bank Group. 2019. “World Development Report: The Changing Nature of Work.” Washington, DC: World Bank Group. [WFP] World Food Programme. Protracted Relief and Recovery Operations documents. [WFP] World Food Programme. 2015. “World Food Programme Ready To Respond To Food Crisis In Malawi.” Press release 25 September 2015. Accessible at https://www.wfp.org/news/world-food-programme-ready-respond-food-crisis-malawi Zezza, A. B. de la Brière, and B. Davis. 2010. “The impact of social cash transfers on household economic decision making and development in Eastern and Southern Africa.” Presented at the African Economic Conference, 27-29 October, Tunis. Working Paper.