89003 DISCUSSION PAPER NO. 1405 Botswana Social Protection Cornelia Tesliuc, José Silvério Marques, Lillian Mookodi, Jeanine Braithwaite, Siddarth Sharma and Dolly Ntseane Africa Social Safety Net and Social Protection Assessment Series December 2013 BOTSWANA Social Protection Cornelia Tesliuc, José Silvério Marques, Lillian Mookodi, Jeanine Braithwaite, Siddarth Sharma and Dolly Ntseane December 2013 Africa Social Safety Net and Social Protection Assessment Series Until recently, most countries in Africa implemented safety nets and social protection programs only on an ad hoc basis. In the wake of the global economic, food and fuel price crises starting in 2008, however, policymakers in Africa began to increasingly view safety nets as core instruments for reducing poverty, addressing inequality, and helping poor and vulnerable households to manage risk more effectively. During FY2009-2013, to support governments in their quest to understand better how to improve the efficiency and effectiveness of safety nets in their countries, the World Bank’s Africa Region undertook social safety net or social protection assessments in a number of countries in Sub-Saharan Africa. By 2014 assessments have been completed or are under preparation for over 25 countries in sub-Saharan Africa. These assessments analyze the status of social protection programs and safety nets, their strengths and weaknesses and identify areas for improvement, all with the aim of helping governments and donors to strengthen African safety net systems and social protection programs to protect and promote poor and vulnerable people. They were all carried-out with the explicit aim of informing governments’ social protection policies and programs. With the results of analy tical work like these assessments and other types of support, safety nets and social protection programs are rapidly changing across Africa. For a cross-country regional review, please see "Reducing Poverty and Investing in People: The New Role of Safety Nets in Africa," which pulls together the findings and lessons learned from these assessments and other recent studies of safety net programs in Africa. Abstract The report lays out options to use social safety nets to meet the goal set by the Government of Botswana to eradicate extreme poverty by 2016 and in a budget neutral way. Introducing a Family Support Grant is argued to be a more efficient way to address absolute poverty by offering a family benefit to all the families in absolute poverty. Reforms in the Ipelegeng public works program would allow addressing existing shortcomings associated with coverage, targeting, generosity, and wrong labor incentives. The combination of the above will allow for a better weaving of the Social Safety Nets to ensure that the goal of lifting 16% of population from absolute poverty by 2016 is achieved. JEL Classification: I32, I38, J32, H53 Key Words: social protection, systems, safety nets, social assistance, welfare, administration, public policy, public sector reform, developing countries TABLE OF CONTENTS ABBREVIATIONS AND ACRONYMS ............................................................................................... V ACKNOWLEDGEMENTS ............................................................................................................. VII EXECUTIVE SUMMARY ............................................................................................................... IX I. CONTEXT, PURPOSE, AND CONCEPTUAL FRAMEWORK ..........................................................1 A. CONTEXT.............................................................................................................................. 1 B. PURPOSE.............................................................................................................................. 4 C. CONCEPTUAL FRAMEWORK .................................................................................................. 5 II. MAJOR RISKS AND VULNERABILITIES.....................................................................................8 A. VULNERABILITY AND POVERTY ............................................................................................. 8 1. SOURCES OF VULNERABILITY .............................................................................................8 2. POVERTY ........................................................................................................................ 12 B. PRINCIPAL VULNERABLE GROUPS AND RISKS ...................................................................... 16 1. POOR FAMILIES WITH ORPHANS AND VULNERABLE CHILDREN ......................................... 16 2. CHILDREN OUT OF SCHOOL.............................................................................................. 21 3. JOBLESS YOUTHS ............................................................................................................. 24 4. WORKING POOR AND POOR UNEMPLOYED ..................................................................... 26 5. THE DISABLED AND POOR ELDERLY.................................................................................. 27 6. PEOPLE LIVING WITH HIV ................................................................................................ 29 III. MAJOR SOCIAL PROTECTION AND LABOR POLICIES AND PROGRAMS ............................... 31 A. IMPACT OF THE GLOBAL CRISIS AND COUNTRY RESPONSE .................................................. 31 1. GOVERNMENT RESPONSE ............................................................................................... 32 2. HOUSEHOLDS’ RESPONSE ................................................................................................ 34 B. MAJOR SOCIAL PROTECTION AND LABOR PROGRAMS ......................................................... 36 1. SOCIAL ASSISTANCE PROGRAMS...................................................................................... 38 2. ACTIVE LABOR MARKET PROGRAMS ................................................................................ 53 3. SOCIAL INSURANCE ......................................................................................................... 59 C. INFORMAL SOCIAL SAFETY NETS ......................................................................................... 63 1. THE EXTENDED FAMILY AND KIN SUPPORT SYSTEMS ....................................................... 63 i 2. COMMUNITY SUPPORT NETWORKS ................................................................................. 64 3. CIVIL SOCIETY ORGANIZATIONS ....................................................................................... 66 4. MUTUAL AID ASSOCIATIONS ........................................................................................... 67 IV. ASSESSMENT OF THE SOCIAL PROTECTION AND LABOR SYSTEM ...................................... 69 A. SPENDING .......................................................................................................................... 69 B. PROGRAM COVERAGE, OVERLAPS, AND GAPS .................................................................... 75 C. PROGRAM GENEROSITY...................................................................................................... 80 D. TARGETING EFFICIENCY ...................................................................................................... 85 E. COST-EFFECTIVENESS .......................................................................................................... 89 F. INCENTIVE COMPATIBLY AND SUSTAINABILITY .................................................................... 99 G. MONITORING AND EVALUATION ...................................................................................... 104 H. INSTITUTIONAL ARRANGEMENTS ..................................................................................... 111 I. SUMMARY ASSESSMENT ................................................................................................... 114 V. OPTIONS TO STRENGTHEN THE SPL SYSTEM ...................................................................... 116 A. VISION AND STRATEGY ..................................................................................................... 116 1. THE GOVERNMENT’S VISION ......................................................................................... 116 2. BUILDING A COHERENT SOCIAL PROTECTION AND LABOR SYSTEM ................................. 117 3. ELEMENTS OF A STRATEGY ............................................................................................ 119 B. ERADICATING ABSOLUTE POVERTY IN A BUDGET NEUTRAL WAY ....................................... 121 1. INTRODUCING THE FAMILY SUPPORT GRANT PROGRAM............................................... 123 2. STRENGTHENING THE PUBLIC WORKS PROGRAM IPELEGENG ......................................... 131 3. NUTRITION PROGRAMS ................................................................................................ 134 4. SCHOOL FEEDING PROGRAMS ....................................................................................... 135 5. SPONSORSHIPS AND SCHOLARSHIPS .............................................................................. 137 6. ACTIVE LABOR MARKET PROGRAMS .............................................................................. 138 7. CONTRIBUTORY PENSIONS ............................................................................................ 139 C. IMPROVING ADMINISTRATIVE SYSTEMS AND TOOLS......................................................... 140 D. INSTITUTIONAL ARRANGEMENTS ..................................................................................... 153 GLOSSARY ............................................................................................................................... 154 REFERENCES ............................................................................................................................ 159 ii LIST OF BOXES Box 1: Views on cash versus food........................................................................................................ 94 Box 2: International experience with public works ............................................................................. 97 Box 3: Concerns over work disincentives induced by targeted social assistance programs ............. 100 Box 4: Armenia – does the family benefit program create a disincentive to work? ......................... 103 Box 5: Conditional Cash Transfers ..................................................................................................... 130 Box 6: Ethiopia Productive Safety Net Program ................................................................................ 134 Box 7: International Experience with School Feeding Programs....................................................... 137 Box 8: Building the National Registry in Brazil (CADÚNICO) ............................................................. 146 Box 9: Social Cash Transfers and Financial Inclusion: Evidence from four countries........................ 150 LIST OF FIGURES Figure 1: Stunting by wealth quintiles (percent) ................................................................................. 19 Figure 2: Youth and total unemployment rates .................................................................................. 25 Figure 3: GDP and mining growth rates, 2007-2012 (percent) ........................................................... 31 Figure 4: Botswana food and fuel imports .......................................................................................... 32 Figure 5: Prices of selected food items (2008 Pulas) .......................................................................... 35 Figure 6: Social protection and labor spending ................................................................................... 71 Figure 7: Distribution of social assistance spending by type of program, 2012/13 (percent) ............ 72 Figure 8: Spending on SSN programs compared to the food and total consumption gaps of the poor, 2009/2010.................................................................................................................................. 74 Figure 9: Coverage of targeted programs and absolute and total poor ............................................. 77 Figure 10: Concentration curves for the destitute persons program ................................................. 87 Figure 11: Concentration indices for selected social protection programs ........................................ 88 Figure 12: Single registry of beneficiaries and program management information system............. 148 iii LIST OF TABLES Table 1: Spending and coverage of the social protection and labor programs, 2012/13 .................... xi Table 2: Summary of report’s diagnostics and recommendations ....................................................xxii Table 3: Selected social indicators ........................................................................................................ 2 Table 4: Change in poverty rates between 2002/03 and 2009/10 ..................................................... 13 Table 5: Estimates of poverty by age group, extrapolating 2003 results to 2012 .............................. 15 Table 6: Under-nutrition rates ............................................................................................................ 17 Table 7: Pre-primary education enrollment (percent) ........................................................................ 21 Table 8: Secondary education net enrollment rates ........................................................................... 23 Table 9: Major social assistance and active labor market programs: beneficiaries and benefits....... 37 Table 10: BPOPF contributions and expenses, 2010/11-2011/12 (P million) ..................................... 60 Table 11: Botswana: Social protection and labor expenditures ......................................................... 70 Table 12: Public sector spending on social protection (percent of GDP) ............................................ 73 Table 13: Program coverage ................................................................................................................ 75 Table 14: Pension coverage ................................................................................................................. 79 Table 15: Program benefits, 2012/13.................................................................................................. 81 Table 16: Generosity of social assistance programs in developing countries ..................................... 81 Table 17: Wage setting in public works programs .............................................................................. 83 Table 18: BPOPF average monthly retirement pensions .................................................................... 84 Table 19: Concentration and progressivity indices of selected programs .......................................... 89 Table 20: Program policy and guidelines............................................................................................. 90 Table 21: Typical operating costs as a percentage of total program costs around the world ............ 92 Table 22: Cost of the primary school feeding program ....................................................................... 93 Table 23: Operating costs of selected social protection programs ..................................................... 95 Table 24: Ipelegeng expenditures (P millions) .................................................................................... 96 Table 25: Operating expenses of pension funds ................................................................................. 98 Table 26: Evaluations of social assistance programs ........................................................................ 107 Table 27: Illustrative options and cost of the family support grant .................................................. 128 Table 28: Reforming Ipelegeng: Offering a lower wage rate would not only improve targeting, but cover more poor people for longer period of time .................................................................... 132 iv GOVERNMENT FISCAL YEAR April 1 - March 31 CURRENCY EQUIVALENTS Currency Unit: Botswana Pula (Average 2012) US$1: P7.6 ABBREVIATIONS AND ACRONYMS AIDS Acquired Immune Deficiency Syndrome ALMP Active Labor Market Programs APP Alternative Packages Program- Poverty Eradication Initiative ARV Anti-retroviral BAIS Botswana AIDS Impact Survey BCWIS Botswana Core Welfare Indicators Survey BGP Backyard Garden Program BIDPA Botswana Institute for Development Policy Analysis BOCONGO Botswana Council of Non-governmental Organizations BOTA Botswana Training Authority BPOPF Botswana Public Officers Pension Fund CHBC Community Home-based Care CSO Central Statistics Office DSP Department of Social Protection DSS Department of Social Services FRSD Food Relief Services Division FSG Family Support Grant GDP Gross Domestic Product GICO Government Implementation Coordination Office GoB Government of Botswana HIES Household Income and Expenditure Survey HIV Human Immunodeficiency Virus ISPAAD Integrated Support Program for Arable Agricultural Development LIMID Livestock Management and Infrastructure Development MESD Ministry of Education and Skills Development v MFDP Ministry of Finance and Development Planning MIS Management Information System MLG Ministry of Local Government MLHA Ministry of Labor and Home Affairs MTTC Madirelo Training and Testing Center MOA Ministry of Agriculture MYSC Ministry of Youth, Sport and Culture NBFIRA Non-Bank Financial Institutions Regulatory Authority NDP National Development Plan NGO Non-governmental Organization OAP Old Age Pension OVC Orphans and Vulnerable Children PMT Proxy Means Test POS Point of Sale PTA Parent Teachers Association RADP Remote Area Development Program SADC Southern Africa Development Community SB Statistics Botswana (ex-CSO) SPL Social Protection and Labor STPA Short-term Plan of Action for the Care of Orphans TB Tuberculosis TEC Tertiary Education Council UISA Unemployment Individual Savings Account UNDP United Nations Development Program UNICEF United Nations Children's Fund VDC/WDC Village/Ward Development Committee VGFP Vulnerable Group Feeding Program WDI World Development Indicators (World Bank) vi ACKNOWLEDGEMENTS This Social Protection Assessment report was prepared by a World Bank and Botswana Institute for Development Policy Analysis team comprising Cornelia Tesliuc (Team Leader, AFTSP), José Silvério Marques (World Bank consultant), Khaufelo Raymond Lekobane (Research Fellow-Statistician, BIDPA), Lillian Mookodi (Associate Researcher-Economist, BIDPA), Anush Bezhanyan (HDNSP), Jeanine Braithwaite (Consultant), Anna Mohan (Consultant), Mimi Otsuka (AFTP1), Elina Scheja (OPCAE), Siddarth Sharma (AFTSP), Dolly Ntseane (Consultant), Sasun Tsirunyan (Consultant), and Molly Schmalzbach (Consultant). Fiona Mackintosh provided valuable editing support. Ana Lukau, Lucie Muchekehu, and Gaamangwe Thato Thipe provided valuable assistance with document processing and logistical support. The team would like to thank the government officials working on social protection programs for their help as well as the partners and other stakeholders that are supporting the strengthening of the social protection system. The team would particularly like to thank Dr. John Mothibi, Permanent Secretary, and Mrs. Chada Koketso, Deputy Director, Poverty Eradication; Mrs. Ruth Radibe, Director, and Mrs. Papadi Evelyn Nguvauva, Acting Commissioner of Social Benefits, Department of Social Protection, Ministry of Local Government and Rural Development; Mrs. Rose N. Sennanyana, Commissioner, Department of Labor and Social Security and Mr. Ikwatlhaeng G. Bagopi, Permanent Secretary, Ministry of Labor and Home Affairs; Ms. Tlhabologo Ndzinge, Director, Department of Youth, Ministry of Youth, Sports, and Culture; Ms. Bontle Boiteto Majaphoko, Director, Department of Basic Education, Ministry of Education and Skill Development; Mrs. N. W. Senegelo, Deputy Secretary, Development and Budget Division, Ministry of Finance and Development Planning; Mr. Willibald O. W. Erulu, Manager vii Pensions, Non-Bank Financial Institutions Regulatory Agency; Dr. Tebogo B. Seleka, Executive Director, BIDPA; Ms. Bagaisi Mabilo, Executive Secretary of BOCONGO; Dr. Collie B. Monkge, Vision 2016 Coordinator; Mr. Joshua Andi Emmanuel, UNICEF; and their staffs. The team benefitted from the guidance of peer reviewers Margaret Grosh, John Elder, and Ruslan Yemtsov. viii EXECUTIVE SUMMARY 1. Botswana has maintained one of the world's highest economic growth rates since independence in 1966. Economic growth between 1966 and 2008 was 8.7 percent. Through fiscal discipline and sound management, Botswana has transformed itself from one of the poorest countries in the world with per capita GDP of US$ 70 at independence to an upper-middle-income country with a per capita GDP of US$ 8,533 in 2011. The growth rate has reduced extreme poverty—the share of population living below one dollar a day—from 23.4 percent in 2003 to 6.4 percent in 2009/2010. Botswana's heavy reliance on a single export (diamonds) was a critical factor in its 8 percent decline in real GDP in 2009. The expected levelling off in diamond mining production within the next two decades may overshadow the country’s long-term economic prospects. 2. Despite its rapid economic growth and high per capita income level, the country is still plagued by many development problems common to low-income countries. Inequality is very high, with a per capita consumption Gini of 0.49 in 2009/10. Despite the relatively high average income per capita, high inequality explains why in 2009/2010, 19.3 percent of the population are poor, and 16 percent of the population have consumption below the food component of the Poverty Datum line (referred in this report as absolute poverty). In monetary terms, the absolute (food) poverty line was P170 per capita per month (about US$20) and the total poverty line was P220 per capita per month (about US$26)1. A large number of children suffer from malnutrition and 31.4 percent of children under the age of five exhibit stunted growth. A large number of children are not in school, or are behind their expected grade level. Formal labor market opportunities are very limited, and a large share of adults work in low productivity—low wage jobs (e.g. in agriculture) or are unemployed. 1 The total poverty line was P880 per month for a household of 4 people. The absolute (or food) poverty line was P680 per month for a household of 4 people. See Statistical Briefs 2013/11 (July 2013) and 2011/15 (November 2011) on Botswana Core Welfare Indicator Survey 2009/10 by Statistics Botswana. ix The unemployment rate is also very high at 17.8 percent in 2009/10, though if discouraged workers are taken into account, the rate would be closer to 30 percent. National HIV/AIDS prevalence among adults was estimated at 23.4 percent in 2011, the second highest rate in the world. 3. To respond to these economic and social challenges, Botswana has put in place a mature and complex social protection system (see Table 1). Botswana is one of the few countries in Africa that fully funds the social protection programs out of own resources, and dedicate a large part of its GDP to this endeavour. During the 2012/13 fiscal year, social protection spending accounted for about 4.4 percent of GDP, or P5,347 million. Social insurance spending accounted for 1.2 percent of GDP and consisted of contributory pensions mainly for public sector employees. To support employment, Botswana also finances a number of active labor market programs whose total spending represent 0.17 percent of GDP in 2012/13. Social assistance spending represents another 1.7 percent of GDP, allocated to a diverse mix of programs that address the most vulnerable groups in the country. The social assistance program mix includes a public works program (Ipelegeng); a social (non-contributory) old age pension; cash and in-kind assistance for destitute (indigent) persons and families who take care of orphans; nutrition programs for infants and pregnant and lactating women; and school feeding for primary and secondary school children. A fourth area of spending focuses on sponsorships and scholarships for students in tertiary education, accounting for 1.4 percent of GDP spent in 2012/13. These policies lie at the border between human capital development policies and the safety net. x Table 1: Spending and coverage of the social protection and labor programs, 2012/13 Share of No. of Share of Spending GDP beneficiaries Population 1. Pensions and Other Insurance 1,435 1.2 Public Officers Pension Fund 1,087 0.9 6,600 0.3 Non-contributory Public Plan 348 0.3 n.a. 2. Active Labor Market Programs 214 0.2 4,928 0.2 National Internship Program 73 0.1 2,891 0.1 Apprentice Program (MTCT) 20 0.0 949 0.0 Youth Development Fund 97 0.1 1,088 0.1 Youth Empowerment Scheme 24 0.0 n.a. 3. Social Safety Net 3,698 1.7 Cash and Near Cash 833 0.7 Orphan Care 301 0.2 40,030 1.9 Destitute Persons 241 0.2 30,518 1.5 Old Age Pension 279 0.2 93,639 4.5 In-kind Food Transfers 651 0.5 Vulnerable Group Feeding 166 0.1 383,392 18.3 Program Primary School Feeding 275 0.2 268,761 12.8 Program Secondary School Feeding 210 0.2 161,929 7.7 Program Public Works / Ipelegeng 409 0.3 55,000 2.6 Poverty Eradication Initiative (APP) 104 0.1 3,586 0.2 Other Transfers a/ 38 0.0 3,424 0.2 4. Sponsorships/Scholarships 1,674 1.4 28,507 1.4 Total Social Protection and Labor 5,347 4.4 Memorandum items GDP (P millions) 122,500 Population 2,099,601 Source: Program managers, MFDP’s Budget Books, and CSO (2012) Notes: a/ Other transfers include: Community Home-based Care, MLG transfers to NGOs, Remote Area Development Program, and Word War Veterans 4. At the same time, the social protection system exhibits a number of features that are suboptimal. While Botswana has many social protection programs, some of them are rather small relative to the target group they try to cover or to the number of poor people, which limits their effectiveness. Contributory pensions cover less than 13 percent of the workforce, reflecting the structure of the economy and its small formal sector. Targeted xi programs for the poor such as the Destitute Persons or Ipelegeng cover less than 3 percent of the population. Furthermore, safety net programs are fragmented. They are implemented by different ministries, diluting scarce administrative capacity. The current social assistance spending is skewed in favor of a single program—scholarships and sponsorship for tertiary students—which absorbs 45 percent of all social assistance spending. It is very likely that these programs benefit mostly the non-poor. The effectiveness of the nutrition program is in doubt, given the high and persistent level of malnutrition. The current system relies too heavily on in-kind distribution of food (e.g. school feeding programs, Vulnerable Groups Feeding Program), which require that a large share of their budget is spent on administration and logistics. In addition, the safety net covers only a small fraction of the absolutely poor because of its reliance on programs targeted at individuals, not families, and lack of a last resort family-focused anti-poverty program. 5. These weaknesses, however, could be corrected over the next few years and by 2016 Botswana could emerge with a modern, effective and efficient social protection system, capable of eliminating absolute poverty. Using the analytical framework of the recently adopted Social Protection Strategy of the World Bank, this report offers punctual suggestions on how to strengthen the social protection system in Botswana. Based on an in- depth assessment of the social protection programs, this report recommends three strategic directions and associated policy measures that would strengthen the social protection system over the next three to seven years. First, the report proposes a number of options to tighten the safety net and contribute to the elimination of absolute poverty in a budget neutral way. These policy measures are focused on filling existing gaps in coverage, eliminating program fragmentation and curtailing over-generous benefits. Second, the report spells out a comprehensive but feasible administrative reform agenda, targeted to develop the administrative tools and systems that underpin an efficient and effective service delivery across all social protection programs. A third set of xii recommendations focus on strengthening institutional arrangements and coordination mechanisms. Eradicating absolute poverty in a budget neutral way 6. Recently, the Government set an ambitious goal to eradicate absolute poverty by 2016, relying on its safety net as one of the key pillars. In 1996, the Government has developed Vision 2016, an overarching strategy that identifies the goals for Botswana’s development and provides a timeline for achieving them by 2016. One of the pillars of the Vision calls for “a more equitable income distribution and the eradication of absolute poverty in Botswana”. To achieve this goal, Vision 2016 and the accompanying Social Development Policy Framework recognize the need for a social safety net for the poor and vulnerable that can withstand unexpected disasters of crises. As of 2009/10, 16 percent of the total population lives in absolute poverty. If the share of population living in absolute poverty has not changed by 2012/13, it would represent about 84,000 families or 336,000 persons. 7. The goal of lifting 84,000 families from absolute poverty by 2016 is achievable through redirecting 0.4-0.6 percent of GDP toward families living in absolute poverty and implementing a targeting system. Tactically, this will require a better weaving of the safety net through the introduction of a last-resort, poverty-targeted program; a reduction in the wage rate of the public works program, Ipelegeng, to improve both the targeting and the coverage of the program; and accompanying cost-cutting measures to ensure budget neutrality. 8. The suggested road map rests of three pillars: • A first pillar will consist of a new family-based last resort anti-poverty program, tentatively called the Family Support Grant (FSG). This program will provide last xiii resort cash assistance to all absolute poor families that are not adequately supported by other social protection instruments. • A second pillar includes the Ipelegeng public works program, which will continue to focus on poor adults without employment opportunities. Currently, the program achieves modest targeting results due to poor regional allocation of its funding and by offering a relatively high wage rate which makes the program attractive for the non-poor. The report recommends a set of parametric reforms to improve the targeting and the effectiveness of this program. • A third pillar consists of programs with good design, coverage and budget that will continue to operate in the same manner and will cover a part of the absolute poor. In this pillar we include the Old Age Pension and the Community Home-based Care, which will continue to cover all the elderly poor and the chronically ill that require special diets and care. 9. While Botswana has numerous safety net programs aimed at vulnerable individuals, it lacks a last-resort anti-poverty program that would protect all individuals against absolute poverty. The focus on programs serving vulnerable individuals is likely to lead to overlapping assistance for some families, while others would not be covered at all. In particular, most families with orphans, destitute persons, elderly over 65 years old, or whose breadwinners can get a place in Ipelegeng, receive enough cash assistance to bring their total income or consumption above the absolute poverty line. The rest of the families in absolute poverty are left out. Economic growth, especially if pro-poor, could bring some of these families out of poverty. But not all families can be lifted out of poverty solely with economic growth. 10. To reach all families living in absolute poverty, the report recommends designing, piloting and rolling out a Family Support Grant (FSG) program by 2016. To cover families in xiv absolute poverty that are not reached by existing programs, many countries have put in place a last-resort, poverty targeted program. The proposed FSG is such a program, targeted using a proxy-means test toward the poorest and offering a benefit of P85 per capita per month which is slightly more than the average consumption gap of the poor. This is equivalent to P340 per month for an average family of four. Chapter 5 sketches three alternative options for FSG program. The first two options suggest replacing two existing programs—Destitute Persons and Orphan Care programs—with the Family Support Grant program. Option 1 would cover the poorest 24 percent of the population to account for inherent leakage to non-poor of about one third of the beneficiaries. This option would be budget neutral. Option 2 would cover 36 percent of the population, assuming 50 percent leakage, and will require an additional 0.2 percent of GDP. Both program options will continue to cover poor and lower-middle-income families that take care of orphans or that have destitute persons. If replacing existing programs with FSG program would be politically difficult, a third option suggests a “complementary FSG”. The program will offer a P85 per capita per month to all families identified by the proxy-means test as the 24 percent poorest, but only to family members who are not covered by other individual, more generous programs. Family members that benefit from the Destitute Persons, Orphans, Old Age Pensions or Ipelegeng programs would continue to get these benefits, instead of the P85 offered by the FSG program. The third option would cost an additional 0.35 percent of GDP. 11. The report suggests an implementation road-map for the FSG program: a gradual approach, where the design of the program is developed during 2013 based on the micro data of the Core Welfare Indicator Survey 2009/10. Piloting would take place during 2014 and the full program will roll out by the end of 2015. Regardless of the FSG option endorsed by the Government, the implementation of FSG would require complementary administrative reforms, such as the development of a targeting system based on a proxy- means test. Once the program matures, the Government could consider additional reforms that would enhance the impact, i.e. by introducing work requirements for individuals xv capable of working (in areas where the public works program does not operate) and by conditioning cash assistance to certain behavior changes related to the use of nutrition, health and education services. 12. To improve the targeting and effectiveness of the public works Ipelegeng program the Government may consider reducing the wage rate while increasing the coverage and duration of the program. Currently, the program suffers from a poor regional allocation of its funds across regions and areas and is not self-targeted because it offers a higher wage rate than the prevailing market rate for low skilled workers. Because the wage rate is high, the program has become attractive for non-poor beneficiaries, who are taking the place of the poor in the program. The high demand for Ipelegeng has forced the administrators of the program to ration the excess demand through a lottery system. Most likely, this type of administrative rationing has reduced the targeting of the program. To improve its targeting, the program needs to offer lower wages for an extended period of time. For example, reducing the wage rate by one third to P390/month (in line with the minimum wage for agricultural work) would allow extending the duration and the number of beneficiaries served by a combined 50 percent under the same budget; and the targeting of the program would improve. To increase its effectiveness, the size of the program and its budget need to be determined based on the total number of poor working-able adults. Its regional distribution needs to be determined based on the regional distribution of absolute poverty (e.g. based on the regional estimates of absolute poverty of the CWIS, or even better through a poverty map). Increasing the efficiency of other social assistance programs 13. To address the high and persistent level of chronic under nutrition, we recommend a thorough review of current Vulnerable Group Feeding Program, which does not seem effective. The review should focus on both the design and implementation of these programs. In particular, it should answer whether the design of the program conforms to xvi good practice principles (e.g. focusing the support on first 1,000 days, using proven interventions), and whether the actual implementation of the program follows these principles. The recommendations of the review should help improve the effectiveness of the program by 2016. 14. Over the next three years, we recommend the Government to develop a strategy and action plan to gradually shift assistance from in-kind or cash with restricted use to electronic payments using financially inclusive mechanisms. The Strategy should be informed by rigorous impact and implementation evaluations carried on for the largest programs. The current system relies too heavily on in-kind distribution of food (e.g. Vulnerable Groups Feeding Program, Destitute Persons), which are known for spending a larger share of their budget on administration and logistics or reduce the choices of beneficiaries. Moving away from in-kind programs toward cash assistance would reduce administrative costs and give beneficiaries more choices and offer more social protection from the same budget. International trends in payments for safety net programmes have seen a move away from in kind benefits or coupons which can only be used towards purchase of certain food items towards outsourced electronic payments and financially inclusive mechanisms. A financially inclusive payment mechanism provides recipients with access to mainstream financial services most frequently with a savings mechanism such as an account. 15. In addition, linking cash transfers to behaviour changes could enhance their impact. Conditional cash transfers have emerged in many countries as a strategy for breaking the intergenerational transmission of poverty and reducing inequality. They provide money to poor families contingent on them making investments in human capital, such as keeping their children in school or taking them to health centres on a regular basis for growth monitoring and immunization. xvii 16. Some of the measures recommended would require additional budget. To finance the additional budget for the FSG program, savings could be obtained over the short term by reducing the generosity of costly programs (e.g. scholarships). The current social assistance spending is highly skewed in favor of a single program: scholarships and sponsorship for tertiary students. This program costs the budget P1.674 million in 2012/13 or 1.4 percent of GDP and absorbed 45 percent of the total social assistance spending. The high volume of spending for these programs shows the commitment of the Government of Botswana to ensure that every child gets access to education, including tertiary. The intended goal of these programs is to break the cycle of inter-generational poverty, by ensuring that children attain the highest level of education, so that they can compete in the labor market. To achieve the Vision 2016 goal, the country could reallocate a small share of these resources away from this program and channel them through programs with good targeting accuracy. The government could reduce the generosity of scholarships for tertiary education by about 25 percent starting from 2015 for high income students. These programs are likely to be regressive, benefiting mostly rich and upper-middle-income students. Their use results in a private benefit, with fewer positive externalities than in primary or secondary education. While these programs have multiple objectives to support skills development, from a public policy perspective, there is a weak rationale for an across-the-board state intervention in this market: students from high income families could cover a share of the cost of their tertiary education. For poor students the report recommends maintaining merit-needs based scholarships covering full cost. Increasing the efficiency of labour market and pensions 17. Botswana operates a number of active labour market programs (ALMP) that enhance youth employability and workers skills, but have low coverage and high benefits (e.g. for internship programs), are disconnected from SSN interventions, and have not been evaluated. Under its Human Resource Development Strategy, which is currently being xviii implemented, the GoB has already decided to streamline the existing institutional arrangements, eliminate existing fragmentation, and assign clear roles and responsibilities to the sectoral institutions. The Human Resources Council, which should be established shortly, plans to begin operating an effective web-based labour market exchange (already developed), to keep track of labour market developments through the Labor Market Observatory, and to rationalize youth programs such as the Internship Program while continuing to support programs that have reportedly shown promising results such as the Apprentice Program. To get more value for money from these programs, the Government could link the ALMPs and SSN beneficiaries, recalibrate the benefits of the internship program, and combine skills training with internships in the private sector in youth employment programs. 18. The contributory pension schemes protect the covered population against the drop in income during old age, but have low coverage and are excessively fragmented. The National Development Plan 10 calls for a review of the social insurance system to increase social security coverage through comprehensive social security policy and legislation. At the request of the GoB, the ILO is currently evaluating the possibility of establishing a mandatory pension plan in Botswana. Currently only public employees are required to contribute to the BPOPF. The limited amount of information that we have on social insurance indicates that Botswana faces three challenges in this area: (i) low coverage, as social insurance covers less than one-half of the formal labour force and about 13 of the labour force; (ii) large-scale fragmentation of the private system (with over 97 private pension plans in existence); and (iii) lack of unemployment insurance. A revised social insurance system in Botswana should: (i) eliminate the existing fragmentation of the private pension system; (ii) continue to allow for voluntary savings; (iii) facilitate the enrolment of the nearly 60 percent of the formal labour force with no insurance; and (iv) facilitate the participation of informal sector workers. Eventually, the existing universal OAP could link to the pension plan and be transformed into a minimum pension guarantee for all those who were not able to contribute to their retirement or contributed only very little. xix Improving administration and institutional arrangements 19. To obtain more value for money from its current investment in social protection programs, Botswana could consider improving the administrative systems and tools that underpin service delivery across key programs, building on the already significant progress made in this area. These tools include targeting methods, beneficiary registration and management information systems (MIS), payment mechanisms, grievance systems, monitoring and evaluation (M&E). In addition, the Government could benefit from a greater use of administrative and survey data in assessing current programs, simulating new programs (e.g. the FSG program), or modifying the parameters of current programs (e.g. the parametric reforms to Ipelegeng). 20. Botswana could develop a PMT-based targeting scheme to target the poor more accurately in the next two years. The calibration of the PMT formula could be done based on the relatively recent CWIS 2009/10. The simulations presented in Annex 1 based on 2002/03 HIES show that developing such a scheme is feasible and cost-effective. However, the information from the HIES 2002/03 is old and does not reflect the current welfare distribution in Botswana. To jumpstart the process, the Government could place the CWIS data under open access and contract a consultant to develop the PMT formula, as well as the operational manual for administering eligibility under the PMT scheme. 21. While the PMT is a precondition for the development of the FSG program, Botswana could use this targeting instrument over time across a larger set of programs. The beneficiaries of the Poverty Eradication Initiatives could be selected among those identified with the PMT as poor. Also, the PMT system could be used to determine which students would be entitled to a fully paid sponsorship or scholarship (e.g. if their family income ranks them in the bottom half of the income distribution), or else have to pay a co- payment. The PMT could be used to identify disadvantaged students who currently are not enrolled but who could benefit from the scholarship and sponsorship programs. The higher- xx income students will continue to be subsidized, but at a lower rate. This targeting mechanism is compatible with the provisions of the Childrens’ Act (2009), which establishes the right to education for every child in Botswana. 22. The Government should facilitate the increased frequency and utilization of administrative and survey data for evidence-based policy making, by promoting an open access policy. Policy analysis and planning in Botswana is hindered because of the lack of up-to-date survey information. Labor force and Botswana Core Welfare Indicators Survey (BCWIS) surveys are done only every seven to nine years, which is too infrequent to allow for timely analysis and policy corrections. The detailed results of the 2009/10 BCWIS are still not available, and some of the analysis in this assessment had to rely on data from the 2002/03 survey. In particular, the Government could benefit from a greater use of administrative and survey data in assessing current programs, simulating new programs (e.g. the FSG program), or modifying the parameters of current programs (e.g. the parametric reforms to Ipelegeng). 23. Institutional Arrangements. The Ministry of Local Government, as key implementing agency, will need to further strengthen its capacity through developing operational systems for improved targeting, efficient delivery of benefits, better case management and monitoring. As the MLG’s front-line staff will be increasingly required to focus on social work/case management, they will need additional training and support. Finally, local authorities should continue to play a key role in the implementation. They will need to be strengthened to perform these functions in a more effective manner. 24. A summary of the diagnostic and recommendations of the report is presented in Table 2. xxi Table 2: Summary of report’s diagnostics and recommendations Social Protection and Function (Primary, Strengths/Issues Recommendations Labor Programs Secondary) Insure against drop in income during Increase social security old age BUT: Low coverage, excessive coverage, eliminate Pensions (Contributory) Resilience fragmentation of private voluntary fragmentation of private schemes schemes Promote linkage between ALM programs and SSN Enhance youth employability and beneficiaries, recalibrate workers skills, BUT: Low coverage, Active Labor Market benefits of internship Opportunity not linked to SSN, duplications, Programs* program, combine skills generous benefits of internship training with internship in programs, no impact assessment private sector in youth employment programs Social Safety Net Protection of elderly poor, benefit Maintain the program within Cash (Old Age Pension) Equity adequate to ensure minimum food current parameters security Replace with FSG program Protects against destitution, BUT: (option 1); or complement limited coverage, does not consider the safety net with FSG Near Cash (Orphans, family needs, coupon limits choices, program (option 2); over Equity Destitute, Needy Children) lacks horizontal equity, overgenerous medium term, link FSG with benefit, does not build human capital human capital investments of children in health and education; protect the disabled Protects poor individuals from Focus on first 1,000 days irreversible losses of human capital and use effective nutrition In-kind Food Transfers (nutrition, health) BUT: Not focused Equity interventions, cost-recovery (Feeding programs**) on the first 1000 days, limited impact in school feeding for those on under nutrition, relatively high who can afford to pay cost per student in school feeding Recalibrate benefit level, Insure against drop in income, BUT: extend duration & number Public Works / Ipelegeng Equity/Resilience Benefit level too high relative to of beneficiaries, distribute minimum wage, poor targeting places at regional level using poverty criteria Target beneficiaries of FSG, Promote access to productive work Poverty Eradication Opportunity/Equit simplify interventions, BUT: limited coverage, difficult to Initiative (APP) y assess impact before scaling monitor up Consider introducing co- Promote human capital Opportunity/ payment for students from Sponsorships/Scholarships accumulation, BUT: likely regressive, Equity upper middle and high fiscally unsustainable income households xxii I. CONTEXT, PURPOSE, AND CONCEPTUAL FRAMEWORK A. Context 25. Botswana is a small, landlocked, upper-middle-income country with a GDP of US$17.3 billion in 2011. Botswana’s 2 million inhabitants are spread over an area of about 600,000 sq. km, approximately the size of France or Kenya. Its environment is mostly arid with the Kalahari Desert covering 70 percent of the country, and the population is mainly concentrated in the little arable land that exists in the eastern parts of the country. 26. Botswana has maintained one of the world's highest economic growth rates since independence in 1966. Botswana’s average economic growth between 1966 and 2008 was 8.7 percent.2 Through fiscal discipline and sound management, Botswana transformed itself from one of the poorest countries in the world with per capita GDP of US$70 at independence to an upper-middle-income country with a per capita GDP of US$8,533 in 2011. However, Botswana's heavy reliance on a single export (diamonds) was a critical factor in the sharp economic contraction of 2009 when real GDP fell by 7.8 percent. Economic growth resumed strongly in 2010 and 2011 at 8.6 and 6.1 percent respectively, but growth weakened in 2012 to 3.7 percent as mining output drop by 8 percent.3 The expected levelling off in diamond mining production within the next two decades may impede the country’s long-term economic prospects. 27. Botswana’s social indicators compare well with its upper -middle-income neighbours (South Africa and Namibia) and other upper-middle-income countries elsewhere in Africa (Gabon) and in Latin America with similar population sizes (Table 3). 2 Ministry of Finance and Development Planning (2010), page 35. 3 Statistics Botswana (2013a), Table A 1. 1 One area where Botswana (and South Africa) falls short compared to its Latin American peers is in terms of life expectancy, a consequence of the devastating HIV epidemic. Table 3: Selected social indicators Botswana South Africa Namibia Gabon Jamaica Panama Costa Rica Population 2.0 50.0 2.3 1.5 2.7 3.5 4.7 (million) GDP per capita 8,533 8,070 5,293 11,114 5,330 7,498 8,647 (US$, 2011) Life expectancy 52 53 63 64 76 79 82 (years) Maternal mortality 160 300 200 230 110 92 40 (per 100,000 live births) Under 5 mortality rate 25.9 46.7 41.5 65.6 18.3 19.5 10.1 (per 1,000 live births) Adult literacy (15 years and 84.5 N/A 88.8 88.4 86.6 94.1 96.2 above) Source: UNDP and World Development Indicators (http://databank.worldbank.org/data/databases.aspx) Notes: Most recent available information mostly for 2010/2011. a/ Out of 186 countries 28. Despite high levels of economic growth over the past two decades, poverty and inequality persist in Botswana and unemployment is high. According to Statistics Botswana (SB),4 the proportion of the population living below the poverty datum line (PDL) declined from 30.6 percent in 2002/03 to 19.3 percent in 2009/10. The latest available estimate of the Gini coefficient of per capita consumption is 0.49 in 2009/10, one of the highest in the world. The unemployment rate is also very high (17.8 percent in 2009/10), although if discouraged workers were taken into account, the rate would be closer to 30 percent.5 4 Statistics Botswana (2011a) 5 See, for example, International Labour Organization (2011), page 12. 2 29. Botswana is one of a group of countries in Southern Africa that have the highest HIV/AIDS prevalence rate in the world. According to UNAIDS, the national prevalence rate of HIV/AIDS (among people aged 15 to 49 years old) in Botswana was estimated to be 23.4 percent in 2011, and this appears to have stabilized in the past few years. There are 300,000 people living with HIV, 160,000 of whom are women and 15,000 are children under 15 years old.6 The number of children orphaned by AIDS is estimated to be over 137,000 or 16 percent of all children.7 30. The Government of Botswana (GoB) has developed and approved key strategic documents that articulate its long-term development objectives, including those in the social protection sector. Vision 2016 is an overarching strategy that identifies the goals for Botswana’s development and provides a timeline for achieving them by 2016. The Social Development Policy Framework, which was prepared by the Ministry of Local Government (MLG) and its Department of Social Services (DSS), represents a comprehensive strategy for making Botswana “a compassionate, just, and caring nation” and for fulfilling the third pillar of Vision 2016, which calls for a more equitable income distribution and the eradication of poverty in Botswana. To achieve this, Vision 2016 and the Social Development Policy Framework recognize the need for a social safety net for the poor and vulnerable that can withstand unexpected disasters or crises. Although progress has been made in alleviating poverty and several social protection programs already exist in Botswana, concerns remain about the medium-term fiscal sustainability of the current safety net as revenues from diamonds are expected to decline. To fulfill the goals of Vision 2016, social protection programs must reach the most vulnerable segments of the population in a cost-effective manner. 6 http://www.unaids.org/en/regionscountries/countries/botswana/. Page accessed April 16, 2013. 7 CSO (2009b), pages 55 and 109. 3 31. Policy analysis and planning is hindered because of the lack of up-to-date survey information. Labor force and Botswana Core Welfare Indicators Survey (BCWIS) surveys are done only every seven to nine years, which is too infrequent to allow for timely analysis and policy corrections. The detailed results of the 2009/10 BCWIS are still not available, and some of the analysis in this assessment had to rely on data from the 2002/03 survey. B. Purpose 32. The purpose of this assessment is to inform Botswana’s future social protection and labor strategy and policies and help to achieve the goals of Vision 2016. The assessment concentrates on social assistance programs as major reforms are already taking place or being considered for ALMP and social insurance programs. As social assistance programs are a significant draw on the GoB’s budget and revenues from mining are projected to decline over the medium term, it will inevitably be necessary to increase the cost-effectiveness of existing programs. In addition, there is an increasing concern in Botswana with the fact that transfer programs may be creating adverse incentives to work and may be promoting “dependency” and with the related need to “graduate” beneficiaries from the programs. 33. To meet its objectives, the assessment is organized as follows. After a brief discussion of the conceptual framework underpinning the assessment in this first chapter, Chapter II discusses the principal vulnerable groups in Botswana and the risks that they face. Chapter III reviews the principal policies and programs that are already in place to address the risks that have been identified as well as the existing informal social protection arrangements. Chapter IV assesses the adequacy of the social protection system by analyzing spending; program coverage, overlaps, and gaps; program generosity; targeting efficiency; cost-effectiveness; incentive compatibility and sustainability; monitoring and evaluation; and institutional arrangements. Chapter V contains recommendations. 4 C. Conceptual Framework 34. This assessment of Botswana’s social protection system is based on the conceptual framework of social risk management developed by the World Bank in the early 2000s. This framework underpinned the World Bank’s Social Protection Strategy during the 2000s 8 as well as the World Bank’s recently published Social Protection and Labor (SPL) Strategy for 2012-2022. This latest World Bank strategy defines social protection systems as “portfolios of coherent programs that can communicate with each other, often share administrative sub-systems, and work together to deliver resilience, equity, and opportunity.”9 As such, a social protection system “has three intertwined goals: • Resilience for the vulnerable through insuring against the impact of drops in well- being from a range of shocks. Key sources of resilience are social insurance programs that minimize the negative impact of economic shocks on individuals insurance, old- age pensions, and scalable public works programs. • Equity for the poor through protecting against destitution and promoting equality of opportunity. Social assistance programs (also known as safety net programs— including cash transfers and in-kind transfers, such as school feeding and targeted food assistance) address chronic poverty. They also protect poor individuals and families from irreversible and catastrophic losses of human capital (nutrition, health, and education), thereby contributing to equality of opportunity. They also lay the foundation for equality of opportunity, notably by helping families feel secure enough to invest in their future and their children. 8 Holzmann and Jorgensen (2000). The World Bank’s first Social Protection Strategy -From Safety Net to Springboard was published in 2001. 9 World Bank (2012a), page xiv. 5 • Opportunity for all through promoting human capital in children and adults and “connecting” men and women to more productive employment. Institutions that promote opportunity are often integrated with those supporting resilience and equity. Cash transfers incentivize investments in human capital by promoting demand for education and health and by helping address gender inequalities. Public works programs provide cash payments to the poor, while increasing physical capital investments. And labor market programs provide unemployment benefits, build skills, and enhance workers’ productivity and employability.”10 35. This assessment also draws on the World Bank’s Africa Social Protection Strategy for 2012-2022 which, in addition to the issues addressed by the overall World Bank social protection strategy, stresses the need to focus on the targeted delivery of social services and on issues of good governance, efficient program design and delivery, and country- driven system development.11 36. In reviewing Botswana’s social protection system, it is important to bear in mind the attributes of a “good” social protection system. A recent background paper for the World Bank’s Social Protection and Labor (SP&L) Strategy 2012-2022 has identified the following principles that should guide the design of an effective social protection and labor system:12 • Equity: A good SP&L system can ensure that fiscal resources are equitably distributed, achieve horizontal equity in the ratio of contributions to benefits among workers of similar levels of income, establish and enforce common rights and obligations, and help to ensure greater equality of opportunities. 10 World Bank (2012a), page 1. 11 World Bank (2012b). 12 Robalino et al (2012). 6 • Inclusion: A good SP&L system identifies coverage gaps, works to ensure the inclusion of the most vulnerable in the SP&L system, and seeks to eliminate any overlaps and redundancies. • Fiscal sustainability: A good system will be fiscally sustainable. This in turn requires an analysis and acknowledgement of the fiscal costs of financing social insurance and social assistance and a projection of program costs under different demographic scenarios. • Incentive compatibility: A good system should be based on program rules that create incentives for individuals to work, save, and participate in risk-pooling arrangements; incentives for employers to register their workers in the SP&L system and to collect and pay the required contributions; and incentives for SP&L insurers and service providers to enroll all eligible beneficiaries and provide good quality services efficiently. • Results focus: A good system will have clearly established goals that are linked to well-articulated programs and will use monitoring and evaluation to track progress toward meeting those goals, monitor implementation, and fill any knowledge gaps about program performance. Good monitoring and evaluation serves as the basis for results-based management, informs policymaking, and supports accountability. • Ability to respond to risks and shocks: A good system will be able to respond effectively to help households to manage idiosyncratic risks during normal times as well as counter-cyclically to respond appropriately to large covariate shocks. 7 II. MAJOR RISKS AND VULNERABILITIES 37. In this chapter, we identify the major sources of vulnerability for Botswana households, the most vulnerable groups, and the principal risks that they face. It focuses on the major risks facing individuals and families that, if not addressed, can perpetuate the intergenerational transmission of poverty. The risks were identified with the input of key stakeholders and on the basis of the analysis of available information in the form of reports and surveys.13 A. Vulnerability and Poverty 38. In this section the main sources of vulnerability for Botswana households are analyzed briefly as well as the trends and determinants of poverty. 1. Sources of Vulnerability 39. The poor and vulnerable in Botswana face a series of important risks. These risks result from: (i) economic shocks and consequent variations in employment, income, and consumption; (ii) the HIV epidemic; (iii) natural disasters, mostly droughts and floods; and (iv) household conditions that expose the poorest families to a series of adverse situations and make them vulnerable.14 The role of the social protection system is to minimize the 13 The Household Income and Expenditure Survey (HIES) 2002/03, the Botswana AIDS Impact Survey III (2008), and the Core Welfare Indicators Questionnaire Survey (BCWIQ) 2009/10. The detailed results and the database of the BCWIQ 2009/10 have still not been released by Statistics Botswana, though it published two reports with the poverty results in November 2011 and July 2013 (Statistics Botswana, 2011a and Statistics Botswana 2013c). 14 The terms risk and vulnerability are sometimes used interchangeably because people or households that are at risk are usually considered to be vulnerable and vice versa. Risk is the probability of a loss or injury. Its potential negative impact may arise from some present process or future event. In the context of poverty analysis, vulnerability is defined by the World Bank as the probability or risk of being poor today or of falling into deeper poverty in the future. It is a key dimension of welfare since the risk of large changes in income 8 impact of these risks using a “portfolios of coherent programs” that work tog ether to deliver resilience, equity, and opportunity.15 40. The Botswana economy is highly dependent for its growth on mining (diamonds) which accounted for about one-third of GDP in 2011.16 Mining also provides the country with the largest share of its export earnings (76 percent of total merchandise exports) and fiscal revenues (33.7 percent of total central revenues).17 The lack of diversification of the economy makes it very susceptible to external economic shocks. Consequently, the diversification of the economic base is a key objective of Vision 2016 and of the GoB’s development plans.18 Meanwhile, in order to diminish the risk originating from external economic shocks and cyclical variations in economic activity, an integral approach to social protection is needed.19 For example, policymakers could choose to respond to deteriorating terms of trade by resorting to market insurance mechanisms such as futures contracts in international markets or by diversifying the country’s external markets. To counter the volatility of private capital, policymakers might use contingent lines of credit. On the other hand, given the fluctuations in global economic activity, it is important to contain fiscal expenditures in times of economic expansion to increase expenditures in times of recession. Maintaining a solid financial system is crucial for ensuring that the depositors’ funds are preserved and to ensure that credit and other services are still available to firms and families, particularly during difficult economic periods. Finally, maintaining prudent fiscal may constrain households from investing in productive assets and their human capital. Vulnerability can influence household behavior and coping strategies and is thus an important aspect of poverty reduction policy. For example, the fear of bad weather conditions or the fear of being expelled from the land that they cultivate can deter vulnerable households from investing in more risky but higher productivity crops and affect their capacity to generate income (Coudouel et al, undated). 15 World Bank (2012a), page xiv. 16 Statistics Botswana:http://www.cso.gov.bw/index.php?option=com_content1&parent_id=370&id=433&nid=370. 17 IMF (2012a), Table 21 and Table 3. 18 Republic of Botswana (2011a), page 8. 19 deFerranti et al (2000), page 7. 9 and monetary policies will help to contain inflationary pressures, as high inflation is one of the most damaging taxes that can be imposed on the poor population, particularly those living on a fixed income and/or subsistence wages. As discussed in Chapter III, the GoB has pursued many of these policies to minimize the impact of external shocks on the population, including during the recent food, fuel and financial crisis when it drew down the reserves that had been accumulated during good times. 41. The first diagnosed case of HIV in Botswana came in 1985 and by the following year HIV had spread already to 1 percent of the adult population. The HIV epidemic quickly grew until it stabilized at approximately 17 percent of the population in the late 2000s. It has affected all ages and socioeconomic groups. The prevalence is significantly higher for the adult population, particularly for those in the 30 to 45 age cohort. Women are more likely to be HIV-positive than men (20.4 percent versus 14 percent). Disease prevalence is higher in urban areas than in rural areas. HIV prevalence is estimated at 19.1 percent for cities, 22.1 percent for towns, and 17.1 percent for rural areas.20 The number of AIDS orphans has been increasing, and in 2008 orphans reached 137,805 or about 16 percent of children in Botswana.21 The social protection system should contribute to addressing the needs of those suffering from the disease as well as those of the thousands of orphans that it has created. 42. Botswana is vulnerable to natural disasters, particularly floods and drought, as the country is located in the sub-tropical high pressure belt of the southern hemisphere in the interior of Southern Africa and away from oceanic influences. Temperatures are high, rainfall is low and variable, and drought is a recurring element of Botswana’s climate. According to the last National Development Plan (NDP 10) “drought adversely affects the already fragile food and agricultural situation in the country and seriously impairs the rural 20 World Bank (2012c), page 7. 21 Republic of Botswana (2008). 10 economy and socio-cultural structures.”22 During the last 25 years, the GoB has declared a major drought five times.23 The last drought that hit the country was in 2011/12. According to a Crop Assessment study by the Ministry of Finance and Development Planning (MFDP), the adverse weather resulted in crop failure and diminished livestock grazing areas across most parts of the country with the Bobirwa and the North West districts being the hardest hit. Children were affected. Underweight prevalence in children under the age of 5 increased from 3.6 percent to 4.7 percent because of a “shortage of supplementary foods and the poor green harvest that was experienced in the previous year.”24 43. In order to diminish the risks associated with natural disasters, countries should have social protection programs in place that can not only mitigate the effects of a crisis after it has hit but that are also capable of playing a preventive role in advance of a crisis. Botswana’s National Disaster Relief Fund (NDRF), created in 1996 and administered by the National Disaster Management Office in the Office of the President, is in charge of disaster prevention and financing the cost of disaster relief. Every year P5 billion is allocated to the fund. The fund is built up in good years and run down when disaster strikes. At the end of 2009/10, the fund had a balance of P8.8 billion. 25 44. Household vulnerabilities consist of the risks that poor households face related to the conditions in which they live.26 These risks are usually associated with a lack of skills, assets, employment, income-generating opportunities, and income security in old age and a lack of access to basic services. They are also related to specific conditions that can affect groups such as the elderly; orphaned, abandoned, or abused children; the physically 22 Ministry of Finance and Development Planning (2010), page 6. 23 Ministry of Local Government (2010), page 154. 24 http://www.mmegi.bw/index.php?sid=1&aid=156&dir=2012/august/wednesday1 25 Ministry of Local Government (2010b), page 54. 26 Risks can be classified as idiosyncratic when they only affect one individual or household (such as the death of the breadwinner) or covariate when they affect several households or even the whole country (such as the recent oil price shock). 11 disabled; and the chronically ill, including those living with HIV/AIDS. Existing vulnerabilities can magnify the impact of economic disturbances or natural disasters. For example, if a youth has no skills, he or she may find it difficult to secure employment even in good economic times, but in bad economic times, there will be even more and better-qualified job seekers competing for the same few jobs. On the other hand, the coping strategies that individuals and families tend to use can make them less vulnerable to risks associated with changes in economic activity or natural disasters. For example, a member of a farming family may get a job in a non-agricultural sector (diversification of production) or participate in a training course (investment in human capital). The family may open a savings account (diversification of assets) or make contributions to a pension (insurance). These are self-insurance, self-protection, and market insurance strategies at the individual level that help households to confront the kind of risks that can affect entire communities or even the entire country. 45. This assessment of Botswana’s social protection system focuses primarily on these household vulnerabilities that affect poor families, but it also discusses the country’s readiness to deal with external shocks such as the recent food, fuel, and financial crises. 2. Poverty 46. Botswana’s poverty headcount rate was 19.3 percent in 2009/10.27 The NDP 10 identified the country’s main causes of poverty as structural, including a lack of education and a lack of employment opportunities. Poverty is higher in the rural areas than in the 27 Poverty statistics based on Statistics Botswana (2013c) and CSO (2011). Statistics Botswana estimates poverty based on a poverty datum line (PDL), which measures the “level of income/consumption that is sufficient to afford the minimum necessities of life” and reflects both individual characteristics and household composition (CSO, 2008, p. 2). The PDLs account for differences in age, gender, region, and household composition (CSO, 2008, p. 3). The PDLs differ in the seven major regions ‒ South East, North East, North West, South West, Gaborone, and Francistown (CSO, 2008, p. 4) as well as by family composition. The average PDLs per month for a household of 4 people was P880 (Statistics Botswana 2013c); for food was P680 per month (Statistics Botswana 2011a, Table 2). 12 cities/towns or urban villages, particularly in the remote areas because of their low resource endowments and relative isolation from the country’s main economic activities.28 Poverty is higher for families raising children than for other families.29 Poverty is also higher for female-headed households compared to male-headed households.30 Table 4 indicates that between 2002/03 and 2009/10 poverty in rural areas declined the most (about 19 percentage points). Poverty incidence varies regionally with the highest headcount rates (of over 40 percent) in the Ngwaketse West and Ngami land West districts. The lowest rate was registered in Sowa, Chobe, Gaborone, Jwaneng and Orapa districts. Table 4: Change in poverty rates between 2002/03 and 2009/10 Region Headcount ratio (%) Household (%) Number of people 2009/10 BCWIS Cities/towns 8.0 5.2 31,401 Urban villages 9.9 10.4 134,467 Rural areas 24.3 14.5 196,249 National 19.3 10.8 362,116 2002/03 HIES Cities/towns 10.3 8.8 39,113 Urban villages 25.4 17.4 138,547 Rural areas 44.8 33.4 321,808 National 30.6 21.7 499,467 Change between 2002/03 and 2009/10 Cities/towns -2.3 -3.6 -7,712 Urban villages -15.5 -7.0 -4,080 Rural areas -20.5 -18.9 -125,559 National -11.3 -10.9 -137,351 Source: Statistics Botswana (2011a, 2013c), Table 1 Note: HIES = Household Income and Expenditure Survey. BCWIS = Botswana Core Welfare Indicators Survey 28 Republic of Botswana and United Nations (2010), page 20. 29 Ministry of Local Government (2010b), page 15. 30 Ministry of Finance and Development Planning (2010), page 45. 13 47. Overall, there were 362,116 poor people in Botswana in 2010. Among these, the most vulnerable are those that could not even afford the basic food necessities, whose consumption is below the food component of the Poverty Datum Line. We refer to this segment of the population as being in absolute poverty. The absolute poverty (or food poverty) rate declined from 23.8 percent in 2002/03 to 16 percent in 2009/10. This implies that over three-quarters of the poor or 288,610 people could not afford the basic basket of food in 2010. Indeed, 21 percent of the rural households were food-insecure as they declared they had “gone a whole day without food.” This percentage drops to 12 and 7 for urban villages and cities and towns respectively, with a national average of 15 percent. To eradicate absolute poverty by 2016, NDP 10 recommended increasing employment opportunities, particularly in rural areas; strengthening human resources to increase the employability of the poor; and providing social protection to the poorest and most vulnerable portions of the population.31 48. Botswana also tracks the evolution of extreme poverty, defined as those households whose consumption falls below the international poverty line of US$1 per capita per month in purchasing power parity. Thanks to the impressive growth rate registered during the past decade, Botswana has reduced extreme poverty—the share of population living below US$1 a day—from 23.4 percent in 2003 to 6.4 percent in 2009/2010. 49. Data on the distribution of poverty by age group is vital for identifying the major groups at risk. Unfortunately, the information is still not available from the 2009/10 Botswana Core Welfare Indicators Survey (BCWIS). Therefore, we are forced to use the information from the 2002/03 Household Income and Expenditure Survey (HIES) to estimate the distribution of the poor population. As can be observed in Table 5, poverty incidence is higher for children aged between 0 and 5 years old followed by the elderly aged 31 Ministry of Finance and Development Planning (2010), p 279. 14 65 and over. Indeed, a study on child Poverty in Botswana carried out for UNICEF concluded that poverty particularly affected children living in households with a high dependency ratio or in households with both children and elderly persons, children living in rural areas, children living with adults who never attended school, or children living in households where no parent is working.32 Finally, it is important to indicate that the incidence of poverty among orphans and other children is similar.33 Table 5: Estimates of poverty by age group, extrapolating 2003 results to 2012 Poverty Poor Contribution to Poverty Rate Age Groups Incidence Population 2012 Population Poverty (BCWIS) Rates (HIES) 2012 0-5 40.5 19.0 262,271a/ 80,595 a/ 6-11 35.3 18.6 239,462 78,898 b/ 12-20 30.4 20.9 404,650 88,655 c/ 21-64 24.9 35.0 1,050,648 148,465 b/ 65 and 36.8 6.5 92,172 27,572 above Total 30.6 100.0 2,049,203d/ 19.3 424,185 Sources: Poverty rates from Seleka et al (2007, Table 10), and Statistics Botswana (2013c, Annex 10) a/ Population for 0-11 from Statistics Botswana (2013b, Table 2.5) based on Population Projection for Botswana 2001-2031. b/ Population for 12-20 and 65+ (shares in total population)from UN World Population Prospects the 2010 Revision (http://esa.un.org/unpd/ppp/index.htm). c/ Population for 21-64 calculated as a residual. d/ Total Population in 2012 from 2011 Population and Housing Census (2,024,904) assuming a 1.2 percent annual growth rate. 50. To sum up, despite efforts to diversify the economy and increase resilience to shocks, Botswana continues to be quite vulnerable to economic shocks, HIV epidemic, natural disaster, and households risks. The children and elderly are the groups with the highest incidence of poverty. 32 BIPDA (2009) and http://www.unicef.org/botswana/7579.html 33 BIPDA (2009), Table 3.3.4, page 74. 15 B. Principal Vulnerable Groups and Risks 51. The principal vulnerable groups in Botswana and the risks that they face are discussed in this section. If these risks are not addressed, they can perpetuate the intergenerational transmission of poverty. Some risks such as the HIV epidemic cut across all groups. 1. Poor Families with Orphans and Vulnerable Children 52. Poverty in Botswana is highest among families raising children.34 The HIV epidemic has taken a heavy toll on families, particularly poor families, with the number of orphans increasing far beyond the ability of extended families to absorb them.35 Data from the Botswana AIDS Impact Survey (III) indicates that about 40 percent of the orphans are looked after by the grandparents, 29 percent by the mother or father, 21 percent by a relative, and 1.3 percent are heads of household.36 Also, about 90 percent of female-headed households contained an orphan or vulnerable child.37 53. An orphan is defined in the 1999-2011 Short-term Plan of Action (STPA) for the Care of Orphans as a child under 18 years old who has lost both parents if they were married or one parent in the case of single-parent families. The STPA also defines a vulnerable child as “a child under 18 years who lives in an abusive environment, in a poverty-stricken family without access to basic services, in a child-headed household, with sick parents or guardians, or is infected with HIV/AIDS.”38 According to the Botswana 34 Ministry of Local Government (2010a), page 1. 35 Republic of Botswana (2011a), page 42. 36 Ministry of Local Government (undated), page 15. 37 Ministry of Local Government (undated), page 16. 38 “A vulnerable child is a child under 18 years who lives in an abusive environmen t, in a poverty-stricken family without access to basic services, in a child-headed household, with sick parents or guardians, or is infected with HIV/AIDS” (White and Devereux, 2011, p. 8) . 16 Institute for Development Policy Analysis (BIDPA), in 2002/2003, there were 39,571 registered orphans and by 2009/10 the number had increased to 48,119.39 However, the official definition of orphan is in contrast with the international definition, which defines as orphan a child who has lost one or both parents. The difference is quite large. Using the international definition, the number of orphans was 137,805.40 54. A large number of children in Botswana suffer from malnutrition. The chronic under-nutrition (height for age) rate is 31.4 percent for children under 5, higher than in Botswana’s higher-middle-income neighbours and about six times the level of its Latin American peers of Costa Rica and Jamaica (Table 6). This indicates that 72,000 children under the age of 5 are stunted. Table 6: Under-nutrition rates South Costa Botswana Namibia Gabon Jamaica Panama Africa Rica Height for age 31.4 23.9 29.6 25.0 5.7 19.1 5.5 (chronic) Weight for 7.2 4.7 7.5 4.0 2.7 1.2 1.0 height (acute) Weight for age 11.2 8.7 17.5 8.0 1.9 3.9 1.1 (underweight) Source: UNICEF (2013) and World Development Indicators (http://databank.worldbank.org/data/databases.aspx). Notes:The data are for 2007-11. Height-for-age (chronic malnutrition and stunting) reflects children’s nutritional history. Weight-for-height (acute malnutrition and wasting) reflects their present nutrition status. Weight-for-age (underweight) reflects both the past and present nutritional status of children (and takes into account both chronic and acute malnutrition). 55. This poor outcome may be related in part to the burden put on the health system by the HIV epidemic. Indeed, the NDP 10 indicated that “the increasing demand for health care services, coupled with complex disease management, has over-stretched the existing health care delivery system. This resulted in long waiting times and congestion in health 39 BIDPA (2010). 40 CSO (2009b), pages 55 and 109. 17 facilities, as well as costly external referrals.”41 Also, high rates of under-nutrition may be a result of the drop in exclusive breastfeeding from 33.7 percent in 2000 to 20.3 percent in 2007, which the Ministry of Health also attributes to the HIV epidemic.42 On the other hand, part of the problem may be a lack of nutritional education and poor hygiene and child feeding practices, as pointed out by some clinic staff and reported in the NDP 10: “… there were no reports of shortages of supplies except cooking oil [at the clinics]. Even in the remote settlements, supplies were delivered in a timely manner. However, concern was raised by some nurses that some children benefitting from this program still show signs of malnutrition and stunted growth. They said that this may be an indication that supplements do not go directly to beneficiaries or that parents are neglecting their responsibilities to ensure that children are adequately fed.”43 56. Chronic malnutrition is highest for children living in poor families, the incidence among families in the bottom 20 percent of the income distribution being almost double the level among children in families in the top 20 percent of the distribution (Figure 1). 41 Ministry of Finance and Development Planning (2010), page 294. 42 Republic of Botswana and United Nations (2010), page 22. 43 Ministry of Finance and Development Planning (2010), page 118. 18 Figure 1: Stunting by wealth quintiles (percent) 57. A recent UNICEF report indicated that stunting reduces a “child’s chance of survival, while also hindering optimal health and growth. Stunting is associated with suboptimal brain development, which is likely to have long-lasting harmful consequences for cognitive ability, school performance, and future earnings.”44 This adversely affects the country’s development. Under-nutrition can be reduced by proven interventions such as improving “women’s nutrition, especially before, during and after pregnancy; early and exclusive breastfeeding; timely, safe, appropriate, and high-quality complementary food; and appropriate micronutrient interventions.”45 This must also be accompanied by appropriate infant feeding practices for HIV-exposed infants and by antiretroviral ARV therapy.46 58. The same UNICEF report also revealed that Botswana is among the 10 developing countries with highest rate of overweight children. It indicated that 11 percent of children under 5 years old (or 26,000) in Botswana are overweight or obese.47 Therefore, in addition 44 UNICEF (2013), page iv. 45 UNICEF (2013), page iv. 46 UNICEF (2013), page 18. 47 UNICEF (2013), Figure 15. 19 to the interventions indicated above, the reduction in malnutrition in Botswana requires an intense nutritional education effort to improve feeding practices and to increase food security. 59. There is solid international evidence that early childhood development (ECD) programs benefit poor children and families. Nutritional supplementation and stimulation interventions administered together, especially to poor children under 2 years old, can help malnourished children to catch up with other children.48 Early stimulation, enrolment in early childhood education and in pre-primary school, and participation in other learning and school readiness activities help children to develop and to perform well when they enter primary school. 60. In Botswana, ECD and pre-primary education is mostly provided by the private sector and service organizations (about 97 percent of the enrolment).49 It is not free. According to the Ministry of Education and Skill Development (MESD), the fees charged for these programs vary from a minimum of P50 to P9,800 per year, equivalent to US$ 7 to 1,200. The supply of ECD is limited in rural areas. There are three ECD levels available in Botswana: baby care services for the 0 to 2 year olds, day care services for 3 to 4 year olds, and pre-primary services for 4 to 6 year olds. The first two levels are the responsibility of the Ministry of Health (MOH) and the Ministry of Local Government (MLG), while pre-primary is the responsibility of the MESD. The MESD supervises the pre- primary program while the MLG supervises the nursery and baby care programs. The MESD provides grants to some NGOs and community-owned centers to enable them to purchase materials.50 48 World Bank (2005), page 136. 49 UNESCO’s UIS database (http://www.uis.unesco.org/Pages/default.aspx?SPSLanguage=EN). 50 http://www.moe.gov.bw/index.php?id=39. 20 61. There is no information on how many children attend ECD. Gross pre-primary enrolment was estimated at 18.9 percent in 2010/11 (Table 7). This means that about one in five children do not attend pre-primary. This compares to rates of 65 percent in South Africa and even higher in Botswana’s Latin American income peers. Indeed, the recent review of progress toward the implementation of Vision 2016 concluded that “early child education [is] not seen as a priority.”51 Table 7: Pre-primary education enrollment (percent) Botswana South Africa Jamaica Panama Costa Rica Gross pre-primary 18.9 65.0 113.0 66.1 72.9 enrollment, total Gross pre-primary 18.7 65.0 113.2 65.6 72.5 enrollment, male Gross pre-primary 19.2 65.1 112.7 66.6 73.2 enrollment, female Source: World Development Indicators (http://databank.worldbank.org/data/databases.aspx) Note: Most recent available information mostly for 2010/2011 2. Children Out of School 62. While Botswana has achieved very high rates of primary school enrollments, a significant number of children still do not attend school. Indeed, to address this problem, the MESD has a Children Out-of-School Unit within the newly created Department of Basic Education. The official age range for the seven years of primary school is from 6 to 12 years old, and the range for the three years of junior secondary is from 13 to 15 years old. The 10 years of primary and junior secondary schooling constitute the basic education cycle. 52 Senior secondary education is two years long for teens aged 16 to 17 years old. Over 94 51 See Status at Glance across all Pillars by Key Result Areas at http://www.vision2016.co.bw/vision- content.php?vid=30 (accessed April 17, 2013). 52 Basic education is not mandatory. The GoB’s goal is to achieve universal enrollment at this level. 21 percent of primary and second education students attend public schools.53 All primary school are run by the district councils. Secondary schools are managed by the MESD. 63. According to Statistics Botswana, the net enrollment rate at primary education in 2012 was 88.6 percent.54 This indicates that 11.4 percent of children aged 6 to12 years old who were expected to be enrolled in primary school were not.55 This corresponds to 32,380 children.56 Some of these children might have been in pre-primary and will enter primary school late.57 Children with “special needs, children in remote areas, and children with other social needs” may not be reached by the system.58 Still other children may be from families that cannot afford to send their children to school. Indeed, one MESD official indicated to the team that child labor is still a problem in some areas. A representative of the NGO Marang Child Care Network indicated that street children are a visible occurrence in some cities such as Ghansi. These children are in the streets during the day but return home during the night. 64. At the secondary level, enrollment drops rapidly, with female enrollment surpassing male.59 Table 8 indicates that Botswana lags behind its Latin American peers in this respect. No information is available for its regional neighbours. Most students leave school before senior secondary as the progression from junior secondary is not automatic 53 UNESCO’s UIS database (http://www.uis.unesco.org/Pages/default.aspx?SPSLan guage=EN). Statistics Botswana (2013) indicates that 92.7 percent of primary schools were government-owned and 7.3 percent private-owned. 54 The Status Report 2010 on Botswana’s Millennium Development Goals reported a net primary enrollment rate of 90 percent in 2003, which dropped to 86.9 percent in 2007 (Republic of Botswana and United Nations, 2010, Table 1). 55 Statistics Botswana (2013b), page 6. 56 The World Bank’s World Development Indicators (http://databank.worldbank.org/data/databases.aspx) show 38,192 children out of primary school in 2009, 18,042 girls and 20,150 boys. 57 The difference between the adjusted net primary enrollment and net primary enrollment rates during the last several years has been about 0.2 percentage points (World Development Indicators, http://databank.worldbank.org/data/databases.aspx). The difference between the two concepts accounts for children of primary age who are already in secondary school. 58 Ministry of Finance and Development Planning (2010), page 100. 59 The transition from primary to junior secondary is high since all students are automatically accepted consistent with the GoB’s policy of promoting universal coverage of 10 years of basic education. 22 as from primary to junior secondary. A number of these children go on to vocational training centers. Botswana has a very extensive and to some extent under-used system of vocational training,60 which in 2007 enrolled 10,094 students (6,305 boys and 3,789 girls).61 Additionally there were 5,098 students in eight technological colleges (specializing in, for example, teacher training and health care) managed by the MESD in 2010. Table 8: Secondary education net enrollment rates Botswana Jamaica Panama Total 61.1 83.6 68.7 Female 65.4 86.8 71.6 Male 56.8 80.4 66.0 Source: World Development Indicators, (http://databank.worldbank.org/data/databases.aspx) Note: Most recent available information, mostly for 2010/2011 65. Public primary school is free, while in junior and senior secondary school there has been since 2006 a “cost sharing” arrangement between the MESD and parents. At the secondary level, parents pay P300 for junior secondary per year and P450 for senior secondary. According to NDP 10, school fees represent 5 percent of the total costs of operating secondary schools. The school fees were re-introduced in 2006 for children whose parents could afford such fees. Those households earning less than P1,200 per month or who had fewer than 50 cattle, 100 calves, or 250 goats or sheep were exempted from paying school fees. Government policy is not to exclude children who have not paid school fees as the fees can be waived on the basis of need.62 However, parents do incur other costs to send their children to school such as the cost of school materials, transport, uniforms, pot fees, and PTA fees, as well as the opportunity cost of forgoing the time that the children would otherwise be able to spend working at home, in the shop, or in the field. 60 World Bank (2012d), page 19. 61 Statistics Botswana (2011b). 62 Ministry of Finance and Development Planning (2010), page 99. 23 66. In this context, it should be mentioned that the Labor Force Survey of 2005/06 found that 38,000 children between the ages of 7 and 17 years were engaged in economic activities.63 Also, the 2009/10 BCWIS found 5,152 children aged 12 to 14 and 35,448 teens aged 15 to 19 to be economically active. Moreover, a recent World Bank study64 reported that in Botswana, in one-half of the families surveyed, children were left home alone on a regular or occasional basis. Fifty-two percent of families leaving children home alone relied on other children to help with child care. As Vision 2016 points out “in many cases, parents are exercising a rational choice over whether or not to send children to school. The introduction of universal schooling must go hand in hand with the improvement of socioeconomic conditions to the point where children are no longer viewed as an essential source of labor or income for poorer families, or girl children used to care for younger siblings when their mother works. Young mothers must also be entitled to re-enter the mainstream of education and re-join the schools that they left during pregnancy.”65 3. Jobless Youths 67. The principal risk facing young people is never having attended school or leaving school without the skills that can get them a good and stable job. Most youths without adequate skills remain jobless. The youth rate of unemployment in 2010 was about 40 percent for the 15 to 19 cohort and 34 percent for the 20 to 24 cohort compared with a national average of 17.8 percent (Figure 2). 66 Botswana has no monopoly on high rates of youth unemployment as these are prevalent in most countries in Africa and elsewhere. According to the GoB, the “lack of opportunities for youths poses a serious challenge to the 63 Ministry of Finance and Development Planning (2010), page 204. ILO (2011, page 14) reports that child labor affects 82,400 children (20 percent). According to the ILO, Botswana has ratified its Convention No. 182: On the Elimination of Worst Forms of Child Labor and has in place an action plan to eliminate it by 2015. Other sources report that in 2011 there were 25,000 children under the age of 15 working on farms and cattle posts, and another 1,500 were working in domestic service (United States Department of Labor, 2011). 64 Lundberg and Wuermli (2012) quoting a study conducted by Ruiz-Casares and Heymann (2009). 65 Republic of Botswana (2011a), page 25. 66 Statistics Botswana (2011a), page 5. 24 nation as it increases unemployment and dependence on government assistance programs.”67 A mismatch between the opportunities available in the labor market and the skills possessed by graduates contribute significantly to these high rates of youth unemployment. Many of the young jobless are graduates of secondary and higher education institutions. According to NDP 10, tackling the existing mismatch will require the “reform of the education system such that it provides enterprises with workers with relevant skills and the capacity to adapt to changing circumstances.”68 Figure 2: Youth and total unemployment rates 68. The high rates of unemployment, particularly among recent graduates, may be explained in part by public sector employment, public wages, and education policies. The public sector is the main employer of formal workers in Botswana as it employs near one- half of the formal labor force, and it offers very attractive wages, job security, a good pension plan, and other benefits. Indeed, average cash earnings in the central government are 42 percent higher than in the private sector, about P6,091 compared to P4,294 per month in March 2011.69 According to the Ministry of Labor and Home Affairs officials, this 67 Ministry of Finance and Development Planning (2010), page 93. 68 Ministry of Finance and Development Planning (2010), page 71. 69 Statistics Botswana (2012a), Table 5. 25 average earnings rate in the private sector is similar to the minimum salary established for an entry-level university graduate into public service. On the other hand, according to a recent IMF study, education policies at the secondary and higher levels have been biased towards preparing students in areas that serve the public sector rather than those in demand in the modern private sector. Indeed, a recent World Bank assessment of the vocation training system concluded that unemployment in Botswana “appears to be the consequence of inadequate training or training in the fields of specialization that are not in demand on the labor market.”70 This may result in part from the large scholarship program discussed in Chapter III. Therefore, recent graduates, with very few skills to offer private firms and high reservation wages, are willing to wait for the opportunity to get a public sector job.71 Moreover this wait in many cases is financed by the GoB through various programs such as the Internship Program discussed in Chapter III. 4. Working Poor and Poor Unemployed 69. Botswana’s economy depends heavily on mining which is a capital-intensive industry and thus creates few jobs directly. The unemployment rate was 17.8 percent in 2010.72 This estimate excludes discouraged job seekers, but if they were included, this would nearly double the unemployment rate.73 Unemployment is higher among women than men (21.4 percent versus 14.5 percent) even though women have lower participation rates than men (71.6 percent versus 81.5 percent).74 The “formal labor force” numbered 378,832 workers in 2010. Of these, 192,598 (51 percent) worked in the public sector (central and local government and parastatals). The total labor force in 2011 was 1.1 70 World Bank (2012d), page 32. 71 For a full discussion, see IMF (2012b). 72 Statistics Botswana (2011a), page 5. 73 See, for example, ILO (2011), page 12. 74 Statistics Botswana (2011a), page 5, World Development Indicators (http://databank.worldbank.org/data/databases.aspx). 26 million.75 Therefore, 18 percent of the labor force and one-half of the “formal” labor force work for the government. Those outside the formal labor force, about 677,857 workers, are in the informal sector, many in low-productivity activities. Indeed, about 70 percent of rural households still depend in part on agriculture for their livelihoods, which are based on low and erratic rainfall, poor soil, limited inputs, and rain-fed systems of low productivity.76 The 677,857 workers in the informal sector are likely to include the 148,470 estimated poor between the ages of 21 to 64 years old (see Table 3 above).77 5. The Disabled and Poor Elderly 70. The 2001 population census counted 58,976 people living with disabilities, which was 3.5 percent of the national population. Only 12 percent of the adults were economically active. However, according to a 2011 study from the World Bank and the World Health Organization (WHO), the disability prevalence rate in the adult population in 56 lower-income and higher-income countries averaged 15.6 percent.78 If this rate were also to apply to Botswana, then there would be about 300,000 people with disabilities in the country. 71. People with disabilities have been supported by the Rehabilitation Services Division of the Ministry of Health (MOH) and by several dozen NGO facilities around the country.79 The GoB supports the work of 30 civil society organizations in the sector. The Sir Seretse Khama Memorial Fund has been one source of wheelchairs and appliances for disabled people. In 2010, the GoB established a position in the Office of the President, 75 The World Bank’s World Development Indicators report a labor force of 1,056,689 people in 2011. 76 Ministry of Finance and Development Planning (2010), pages 182 and 183. 77 Statistics Botswana defines the economic active population as comprising those 12 years old and above, though the legal minimum working age is 14 years old. 78 WHO and World Bank (2011), page 28, Table 2.1. The difference in estimates derives in part from different definitions of disability. 79 Ministry of Local Government (2010a), page 15. 27 which is responsible for the coordination of this sector. The Council for the Disabled (BCD) is an umbrella organization that coordinates the activities of NGOs that provide rehabilitation services to people with disabilities. Several associations of disabled people have come together to form the Botswana Federation for the Disabled. These three entities must join together in an effort to help to “overcome the poverty and stigma still often associated with life as a disabled Batswana.”80 The BCD emphasizes the need for enhanced educational approaches and training facilities, especially to help young people to find work. It also urges better care for people with mental disabilities. 72. Many elderly people in Botswana retire without a pension. The number of people aged 60 years and over is estimated to be 135,187 (as of 2012),81 of whom about 40,500 (30 percent) are estimated to be poor.82 The number of BPOPF pensioners is 6,619 and the number receiving pensions from private funds is 2,205.83 We have no information on the number of retirees of the GoB’s defined benefit pension scheme, but they may number about 2,000.84 This would imply that out of all people aged 60 years and over, only 8 percent have a pension. Many of the 40,500 poor elderly must be among the 124,000 Batswana who are 60 years and over and who do not have a pension. Many of these face major financial difficulties in their old age. The HIV epidemic has placed significant stress on the elderly as they can no longer count on the financial support of their own children. At the same time, they have had “to stretch their own merger resources to care for their grandchildren and relatives who have been orphaned or left vulnerable due to HIV/AIDS.” 85 As discussed in Chapter III, the government has established a universal Old Age Pension for all elderly people aged 65 and over. 80 Ministry of Local Government (2010a), page 16. 81 UN Population Division. 82 This is assumed to be the same poverty incidence of the 65 years and over group in Table 3. These are ballpark calculations because of the lack of updated information on poverty by age group. 83 Data received from the BPOPF. 84 Estimated on the basis of the average pension paid by BPOPF. 85 World Bank (2012c), page 18. 28 6. People Living with HIV 73. HIV/AIDS has reversed many of the gains that Botswana has made in health and development over the past 40 years. Botswana lost almost 20 years of life expectancy, reversing all gains since the 1950s. Since the launch of anti-retroviral (ARV) therapies in the mid-2000s, health and mortality indicators have stabilized and recovered some of the ground they had lost. HIV/AIDS is also a significant drag on the country’s budget as health care costs have increased dramatically. Annual HIV/AIDS-related spending is equal to 2.6 percent of GDP. About one-third of that comes from foreign aid, but the remaining two- thirds come from domestic government spending. This high percentage, which is more than in most other African countries, amounted to roughly US$348 million in 2008. A large portion of that funding has gone to ARV treatment. Anti-retroviral drugs (ARVs) have been remarkably successful in helping people live with HIV, and the government has been effective in getting people on ARVs. In its first year, 2002, the government provided no-cost ARV treatment to 10,000 people. By 2007, that number was up to 120,000 and is expected to increase to 190,000 by 2016. This will have major cost implications.86 74. Botswana has had some major successes in fighting HIV epidemic. However, HIV continues to be a “major public health and development problem in Botswana” as the prevalence rate has stabilized at a high level (17 percent).87 In addition, a large proportion of Batswana do not know their HIV status, and many have limited knowledge of the disease.88 75. A recent Vision 2016 progress report acknowledged that Botswana has one of the most progressive and significantly funded set of programs to combat HIV/AIDS and that 86 World Bank (2012c), page 7-8. 87 World Bank (2012c) page 7. 88 Ministry of Finance and Development Planning (2010), page 299. 29 some progress has been made for some demographic groups such as pregnant women and people aged 15 to 19 years old. However, the report also founds some weaknesses: • Little or no progress has been made for other demographic groups. • Incidence is still extremely high. • HIV/AIDS creates major social and economic problems for Botswana including the proliferation of orphans, the breakdown of the family structure, little or no population growth, lost human resources and skills, absenteeism, and lost income. • There has been too little exposure and discussion of some of the significant drivers of HIV/AIDS including cross-generational sex, gender power relations, and the implicit acceptance of multiple partners, regardless of marital status. 76. Vision 2016 concludes that the goal of “no new HIV infections by 2016” is probably not realistic.89 77. To sum up, despite its relatively good social infrastructure and education and health indicators, Botswana still confronts several challenges, including reducing under- nutrition rates for young children, ensuring that all children attend school, creating jobs for youths and other groups with high unemployment rates, increasing access to social insurance, and continuing the efforts to eliminate new cases of HIV and to treat those already suffering from the disease. 89 Vision 2016 (2013a). Status at Glance across all Pillars by Key Result Areas at http://www.vision2016.co.bw/vision-content.php?vid=30. 30 III. MAJOR SOCIAL PROTECTION AND LABOR POLICIES AND PROGRAMS 78. This chapter begins by analyzing the policies and programs implemented by the Government of Botswana to minimize the impact of the recent global crisis, particularly on the most vulnerable groups, and how households themselves dealt with the crisis. The chapter then goes on to briefly discuss the public programs that are addressing the major risks and vulnerabilities faced by the Batswana that were identified in Chapter II. The chapter concludes with a brief review of the informal social protection arrangements in Botswana. Figure 3: GDP and mining growth rates, 2007-2012 (percent) A. Impact of the Global Crisis and Country Response 79. Botswana’s economic conditions were seriously affected by the food, fuel, and financial crisis. In 2009 the production and exports of minerals, mostly diamonds, dropped sharply, leading to a real GDP decline of 8 percent (Figure 3). At least 2,000 workers lost their jobs in the mining industry.90 Government revenues dropped by 2 percent, mostly as a result of a 32 percent decline in income taxes from minerals in 2009/10.91 90 ILO (2011), page 14. 91 Republic of Botswana (2013), Table 1.c 31 Figure 4: Botswana food and fuel imports 80. Botswana is a net food importer. Cereal imports (mostly sorghum and maize) covered 81 percent of domestic requirements between 2003/04 and 2008/09.92 The food import bill increased by nearly 50 percent between 2005 and 2008, while the value of fuel imports jumped by 65 percent during the same period (Figure 4).93 1. Government Response 81. The GoB has maintained a very prudent fiscal policy. The windfalls from mineral exploration have been invested in the country’s economic and social infrastructure, and most of the remaining has been saved. The country’s strong foreign reserve position and accumulated fiscal savings have enabled the GoB to pursue a countercyclical fiscal policy in times of natural disaster or other type of shocks. With the recent global crisis, the GoB 92 Ministry of Finance and Development Planning (2010), Table 9.10. See also BIDPA Briefing (2008), page 2. According to BIDPA, higher cereal prices will not have a positive impact on farm households because most of them are net food buyers. 93 In quantity terms, there was a sharp increase in maize imports in 2004-06 but a drop in 2007-08. Droughts were declared in 2001-2005 (whole country) and 2007-08 (whole country) (CSO, 2009c). 32 sharply increased its expenditures. The development budget in 2008/09 increased by about 40 percent over its original level of P8.5 billion.94 This included:95 • P905 million to the MLG for the maintenance of infrastructure and labor-intensive public works to provide income support to the highly vulnerable groups of the society. Converted the labor-intensive public works program, Ipelegeng, from an intermittent to a permanent program and shifted its focus from drought relief to poverty alleviation. The number of people employed doubled from 9,069 in 2005/06 to 18,085 in 2008/09.96 • A 15 percent increase in the benefits of the three major cash transfers programs: Old Age Pension, the Veterans Program, and the Destitute Persons cash allowance. • P191 million to the Ministry of Agriculture (MOA) for the Integrated Support Program for Rainfed Arable Agriculture Development (ISPAAD) to enhance food production locally. The program provides free seeds, fertilizers, and plowing and planting subsidies.97 In 2008, the ministry also established the Livestock Management and Infrastructure Development program (LIMID) which provides small animals and cattle to poor farmers (see below). • P224 million to implement measures aimed at mitigating the looming food crisis arising from the increase in food prices. 94 “While part of the increase in spending may be related to a fiscal stimulus at times of crisis, the increase in spending began prior to the crisis …” (World Bank, 2010a,Chapter 1, paragraph 35). 95 Republic of Botswana (2009), page 26. During this period, the GoB also increased its expenditures from the National Disaster Relief Fund from P14.5 billion in 2007/08 to P12.1 billion in 2008/09, and P7.3 billion in 2009/10 (Ministry of Local Government,2010b, page 54). 96 Ministry of Local Government (2010a), Table 14. 97 BIDPA Briefing (2008), page 6. 33 82. In 2009/10, extraordinary resources of P1.6 billion (above the original budget allocations) were allocated to the MLG to continue to help finance the increase in coverage and benefits of social programs and face the economic crisis.98 As a result of increased spending and revenue shortfalls, the overall balance of the central government budget, which for many years had been in surplus, turned negative in 2008/09, reaching - 6.3 percent of GDP, -12.3 percent of GDP in 2009/10, -6.6 percent of GDP in 2010/11, to - 0.2 of GDP in 2011/12. However, in 2012/13 a surplus is expected. 83. During the most recent drought in 2012, the GoB took a series of measures to mitigate its adverse impact on the poor. These measures included the provision of a second meal for primary schools students in the affected areas, which include the Kgalagadi District, the Kweneng West Sub-District, the North West District, the Gantsi District, the Mabutsane Sub-District, and the Bobirwa Sub-District, until the end of the 2012/13 financial year. The MLG also began supporting vulnerable groups in affected areas through its Destitute Persons Program and increased the enrollment of workers in its labor-intensive public works program (Ipelegeng).99 2. Households’ Response 84. The price of a basket of basic cereals increased in 2007 alone by between 20 and 28 percent in real terms depending on the region (Figure 5). In Botswana, food represents about two-thirds of poor households’ consumption.100 The sharp increase in the price of basic foodstuffs directly affected many households and may have driven them below the poverty line or deeper into poverty. 98 Republic of Botswana (2010), page 28. 99 http://www.mmegi.bw/index.php?sid=1&aid=156&dir=2012/august/wednesday1 100 CSO (2004b), Table 91. 34 Figure 5: Prices of selected food items (2008 Pulas) 85. In 2010, the Government of Botswana in conjunction with UNICEF began a series of surveys on the effects of the global crisis on vulnerable groups and households in selected localities. The surveys indicated that the impact of the crisis has varied significantly by locality. In some remote and arid areas with few economic opportunities, the economic crisis made things worse: “it appears that the residents of Kacgae have diminished ongoing access to such [health and education] services.”101 On the other hand, other communities have been able to continue to access health and education services despite the downturn. With regard to nutrition for children and the quality of food, the impact was also greater in Kacgae than in other communities.102 86. Based on the survey results, UNICEF has reported that, not unlike other countries, the “coping strategies most frequently mentioned by households are to reduce the quality and quantity of food by buying less nutritious substitutes and reducing the number of meals and/or reducing portion sizes. These coping strategies are independently corroborated by 101 UNICEF (2011), page 1. 102 UNICEF (2011), page 1. 35 the Botswana Family Health Survey of 2007 and the Botswana Millennium Development Goals Report 2010.”103 UNICEF concluded that “clearly, the feeding programs that are in place are not succeeding in addressing the issue of childhood nutrition for the poor in Botswana.”104 87. To sum up, the GoB’s strong fiscal position made it possible to respond to the crisis in a timely manner by increasing the benefits of existing social assistance programs, and making permanent and expanding the reach of its labor-intensive public works program. This may have prevented an increase in poverty rates. However, some needy families excluded from these programs may have been forced to adopt counterproductive coping strategies. B. Major Social Protection and Labor Programs 88. The main public programs that exist to address the risks and vulnerabilities identified in Chapter 2 are discussed briefly in this section. Most of the programs covered here are being implemented by public agencies, in some cases with the cooperation of development partners and civil society organizations, though virtually all programs are financed by the government. We have classified these social protection programs into three major categories: social assistance programs or safety nets, active labor market programs, and social insurance. In this section, we describe briefly the main characteristics of these programs, and Table 9 presents a summary of their beneficiaries and benefits. 103 UNICEF (2011), page 1. 104 UNICEF (2011), page 1. 36 Table 9: Major social assistance and active labor market programs: beneficiaries and benefits 2009/10 2010/11 2011/12 2012/13 Safety Net Programs Orphan Care Program No. of Beneficiaries 46,457 45,816 42,069 40,030 Monthly Food Basket/coupon ranging from P500 to P850; school Benefits uniform for children and other school needs Destitute Persons Program No. of Beneficiaries 40,295 31,969 30,915 30,518 Monthly Food Basket/coupon ranging from P500-700 + cash allowance Benefits P90; school uniform for children and other school needsin 2012/13 Veterans Program No. of Beneficiaries 2,758 2,469 2,166 2,079 Benefits, Pula per 359 359 359 390 month Old Age Pension No. of Beneficiaries 96,118 94,252 91,385 93,639 Benefits, Pula per 220 220 220 250 month VGFP No. of Beneficiaries 268,176 268176 268176 383,392 Benefits Food supplements valued at between P25-P76 per month Primary School Feeding No. of Beneficiaries NA NA N/A 268,761 1 meal per Benefits a/ 1 meal per day 1 meal per day 1 meal per day day Secondary School Feeding No. of Beneficiaries 165,097 N/A 152,410 161,929 a/ 1 meal per Benefits 1 meal per day 1 meal per day 1 meal per day day Sponsorships/Scholarshi ps No. of Beneficiaries NA 22302 28507 NA Benefits Tuition, books, living expenses, warm clothing, medical, and other costs Ipelegeng No. of Beneficiaries 56,458 91,759 127,756 55,000 Benefits, Pula per Cash P480+Food 360 360 400 month P100 CHBC No. of Beneficiaries 2,563 1,517 1,343 1,345 Benefits Monthly food basket ranging from P500-P1200 Remote Area Development No. of Beneficiaries 43,070 N/A N/A N/A Benefits All assistance programs plus small basic infrastructure LIMID No. of Beneficiaries 4,664 399 1,340 3,586 37 Small stock, guinea fowls and Tswana chickens for resource-poor Benefits farmers ALMP National Internship No. of Beneficiaries 1,518 2,075 2,594 2,891 Benefits P2000 per month to be reduced to P1400 since March 2013 Apprentice Program No. of Beneficiaries b/ 1,111 887 1,111 949 Benefits Salary negotiated with sponsor enterprise Youth Development Fund No. of Beneficiaries 252 606 1,134 1,088 Maximum amount of loan/grant for individuals is P100,000; groups Benefits P450,000 Source: Program managers and program guidelines. Notes: a/ Two meals per day in remote and drought-affected areas; three meals for children in boarding schools. b/ Data for calendar year. 1. Social Assistance Programs 89. Social assistance programs can be further classified into those that provide cash or near cash, conditional cash transfers, in-kind food transfers, fee waivers and scholarships, public works, and other transfers. Botswana does not have a conditional cash transfer though some transfer programs are in theory conditional on beneficiaries participating in “rehabilitation” activities (the Destitute Persons Program) or making visits to medical centers (the Home-based Care Program). Cash and Near Cash 90. Orphan Care Program. This program was launched in 1999 and in terms of expenditures is currently one of the largest assistance programs in Botswana. It is managed by the MLG’s Department of Social Protection (DSP), Division of Child Services. It is designed to respond to the needs of orphans, including those for food, clothing, shelter, education, protection, and care. An orphan is defined in the 1999-2011 Short-term Plan of Action 38 (STPA) as a child under 18 years old who has lost both parents if they were married or one parent in the case of single-parent families. As discussed, this official definition is in contrast to the international definition in which an orphan is defined as a child who has lost one or both parents. The difference is quite large. Using the official definition, in 2008 Botswana had about 51,806 orphans whereas if the international definition were used, the number would increase to 137,805 orphans.105 Botswana opted for a narrow definition of orphans, which is often used to determine eligibility for social assistance programs in many countries. 91. Orphans receive a food basket and other items (such as a school uniform, clothing and a transportation allowance) worth about P625 per month (US$82), equivalent to 92 percent of the (2009/10 BCWIS) food poverty line. Orphans are also provided with psychosocial support from MLG’s social workers. The beneficiaries are not means-tested, as the program is open to all families who apply. In 2012/13, the GoB invested P301 million to assist 40,030 orphans.106 92. Concerns have been raised about the tendency for relatives to place siblings in different homes so that the orphan benefits can be “shared” by multiple households. There was no rigorous evaluation of the program to measure its impact on health, nutrition and education outcomes of children. It is not known how this tendency might affect children well-being. The Government might consider requiring families caring for orphans to comply with some basic co-responsibilities such as periodic health check-ups and regular school attendance. 93. Starting in 2009, the MLG replaced the delivery of the food basket with a smartcard (called a coupon) that gives the beneficiary the possibility of acquiring the specified foodstuffs at participating shops. The system uses a network that comprises point 105 Republic of Botswana (2008). 106 Information received from Ministry of Local Government’s Department of Social Services. 39 of sale (POS) devices in conjunction with a smartcard to load “food value” onto the smart card, which then enables the beneficiary to spend directly off the card at participating retailers. The smartcard uses fingerprint verification (biometrics) to authorize each transaction. It gives the beneficiaries a choice of when and, to some extent, where they purchase their food. 94. Destitute Persons Program. This program was established in 1980 as a response to the gradual erosion of the traditional safety net.107 It is managed by the MLG’s DSP, Division of Destitute Persons and Old Age Pension. The program was intended to serve the few who have absolutely no other sources of support but also covers other people in need. The program classifies the destitute persons as either permanent or temporary. The permanently destitute are those whose age or physical or mental conditions render them completely dependent. They are therefore eligible for benefits for life but require an annual recertification by social workers. The temporarily destitute are those who are incapacitated by ill health or natural disasters and are thus unable to support themselves in the short term. In theory the program is means-tested. The eligibility criteria established are: “He or she owns not more than four livestock units or has an income of less than P120/month (US$16) without dependents or less than P150/month (US$20) with dependents.” 108 The program income thresholds have not been updated in years and today represent only one- third of the minimum wage in agriculture and less than 20 percent of the food poverty line. This may be contributing to the low coverage of the program and its high exclusion errors (more on this in Chapter IV). In practice beneficiaries are selected by social workers with input from the Village and Ward Development Committee and the local authorities. The old age pension (OAP) received by the elderly is not taken into consideration so they can also qualify for the Destitute Persons Program. As of March 2013, there were 28,001 permanent 107 Revised National Policy on Destitute Persons, July 2002 (Ministry of Local Government, 2002). 108 Dependants are defined as “those under 18 who depend on an adult for more than half of their subsisten ce and are not being supported by another program such as that for orphans or the RADP.” Ministry of Local Government (2010), page 126. 40 (19,516 women and 8,485 men) and 2,034 temporary (1,340 women and 694 men) destitute people registered with the program. 95. The benefits provided by the Destitute Persons Program are a coupon (a smartcard similar to the one used for the Orphan Care Program) to buy food and a cash allowance for nonfood needs (P90 since October 2012), access to social services, and various subsidies (for example, for housing).109 The food coupon and the cash are provided monthly. The food allowance is intended to provide recipients with 1,750 Kcal per day.110 The coupon has a value of between P450 (US$59) and P650 (US$86) depending on the locality, with the highest transfer being provided in remote areas where food prices tend to be higher. The amount of the food coupon depends on family size or the number of dependents under the age of 18. For families of five or more members, the amount of the food coupon is doubled. It should be noted that in the “remote areas” the MLG still distributes food baskets because of a lack of infrastructure to operate the smartcards. The cash allowance is paid through the Post Office using a checkbook-like leaflet. In 2012/13, the program cost P241 million including the food coupon, the cash allowance, and the support to needy students and needy children. 96. The Destitute Persons Program has become the focus of concerns about dependency since at least 2008 when the Department of Social Services (DSS) released a report detailing the increasing numbers of destitute adults in Botswana. This prompted the government in 2009 to adopt a policy that “able-bodied” destitutes who were receiving support from the Destitute Persons Program should be transferred to the Ipelegeng public works program. The assessment of the Destitute Persons Program concluded that 8,765 of its beneficiaries were able-bodied and could be moved to the Ipelegeng program. By 2010, about two-thirds of these workers had been moved to the Ipelegeng program. 109 World Bank (2010a). 110 BIDPA (2010), p. 26. 41 97. The Veterans Program. This pension was introduced in 1998 for veterans of World War II and their survivors (spouses and children up to the age of 21). The program is managed by the MLG’s DSP, Division of Destitute Persons and Old Age Pension and is implemented by the offices of the District Commissioner in various districts. The benefit provided by the program is a monthly cash transfer of P390 (US$51), which is equivalent to 57 percent of the food poverty line. The payments are made through the Post Office using a checkbook-like leaflet. The beneficiaries are not means-tested. In 2012/13, the program invested P11 million on 2,079 beneficiaries. Disbursements in recent years appear to have been consistently “in excess of entitlements based on claimant numbers.”111 The Program offer some protection to survivors in that, when the direct beneficiary dies, the transfer is given to the surviving wife or, if there is more than one, the transfer is divided among the surviving wives. 98. Old Age Pension (OAP). The OAP was established in 1996. It is a universal transfer that benefits all those aged 65 and older. It is managed by the MLG’s DSP, Division of Destitute Persons and Old Age Pension. In 2012/13, the program covered 93,639 beneficiaries. This is higher than the figure of 92,172 for the population aged 65 years and over in 2012 derived from the UN’s population projections (see Table 3 above). Obviously these population projections are based on the 2001 census and may err on the low side. Nonetheless, it is clear that the program take-up rate is high. The value of the monthly pension since October 2012 has been P250 (US$33), which is equivalent to 37 percent of the food poverty line. As in the Veterans Program, payments are made through the Post Office. In 2012/13 the program cost P279 million. 111 White and Devereux (2011), p. 10. 42 In-kind Food Transfers 99. Vulnerable Groups Feeding Program (VGFP). Since independence, Botswana has had feeding programs aimed at distributing meals and nutritional supplements to people who are vulnerable to under-nutrition. The VGFP was introduced in 1988 to provide take- home rations through health clinics to vulnerable children aged 6 to 60 months, to pregnant and lactating women, and to TB and leprosy patients from poor households. The program is managed by the MLG, Department of Local Government Finance and Procurement, Division of Food Relief Services (DFRS) and is implemented by the Ministry of Health. The contents of the take-home rations vary depending on the beneficiary. The monthly rations consist of the complementary food Tsababa for children aged 6 to 36 months and of a fortified pre- cooked sorghum soya meal (Malutu) for those aged 37 to 60 months. Sugar beans and vegetable oil are provided to children aged 37 to 60 months and to medically selected pregnant and lactating women and to TB and leprosy outpatients. The health clinic staff is expected to screen the patients and provide the rations only to those that have symptoms of under-nutrition. According to the DFRS, the MOH implemented screening guidelines in 2011. All pregnant or nursing teenagers (up to 18 years old) receive the ration.112 In remote areas, all households receive an extra ration. Drought-affected areas other than in remote areas also receive an extra ration. The DFRS procures the goods and delivers them to the clinics on the basis of requests from the clinic staff and the previous months’ allocatio ns. There is very little control by the DFRS over how the rations are used. In 2012/2013, Botswana provided take-home rations to 383,392 beneficiaries. Information on the number of beneficiaries by type is not available.113 The (monthly) cost of the take-home ration in 2012/13 for children aged 6 to 36 months was P34.70 (US$4.6), and for children aged 37 to 60 months and other beneficiaries, it was P76.10 (US$10). In 2012/13 the MLG spent P166.3 million on the program. 112 Ministry of Health (2011), pages 3 and 4. 113 In previous years, about one-half of the beneficiaries were children under 6 months and the other half were children between 6 and 59 months and other beneficiaries. 43 100. Primary School Feeding Program. Botswana provides free school lunches to all students who attend public primary schools. About 94 percent of all students attend public primary schools. Primary school children receive a meal equivalent to one-third of the child’s daily nutritional needs. The meals consist of sorghum porridge, maize, stewed beef, beans, bread and UHT milk. They may include locally procured seasonal fruits and vegetables.114 In the remote areas, students receive two meals a day. The program is managed by the MLG, Department of Local Government Finance and Procurement, Division of Food Relief Services (DFRS). All dry (non-perishable) foodstuffs are procured centrally by the DFRS. In the case of sorghum, raw sunflower, and canned meat, the purchase is made exclusively from two parastatals: the Botswana Meat Commission and the Botswana Agricultural Marketing Board. The food is transported to district food depots by the parastatal companies and from the district food depots to the schools by the DFRS transport fleet. Since 2003, the DFRS has transferred funds to the local councils for them to buy fresh bread and bread spread from private suppliers. The DFRS pays for the hand stampers (traditional grain grinding), and the local councils pay for the cook. As a result of the 2008 Presidential Directive, funds are also transferred from local councils to school procurement committees, which include school heads, MOA and MOH officials, and parents, to procure fresh local produce (initially watermelon).115 Parents’ contributions to the program are currently limited to paying pot fees to purchase utensils and items such as salt and detergent for cleaning.116 In 2012/13, there were 268,761 beneficiaries, and the program cost P274.8 million, including P233.7 million for the purchase of foodstuffs and P41.1 million that was transferred to the local councils. The cost per student per year is two to three times higher in Botswana than in school feeding programs in other countries (more on this in Chapter IV). 114 World Bank (2010a). 115 BIDPA (2012), pages 26-28. Other products to be sold to the school are maize cobs, sweet reeds, and beans. 116 BIDPA (2012), page 39. 44 101. Secondary School Feeding Program. The feeding program also covers secondary students in public schools and students who board in special educational institutions (boarders receive three meals per day). The content of the meals is similar to the primary feeding program. The secondary feeding program is managed by the MESD and financed from its budget. Each school procures its own food with transfers received from the MESD.117 In 2012/13, there were 161,929 beneficiaries of the program at a cost of P209.9 million. Fee Waivers 102. Needy Children and Needy Students. These two programs are managed by the MLG’s Department of Social Services and implemented by local governments. Under the Destitute Persons, the CHBC, and the Remote Area Development Programs, beneficiaries with “needy students” receive uniforms, the payment of room and board in a private house, support for transportation to and from school when the classes begin and end, tuition exemptions, and help with other fees (such as PTA and pot fees). Transfers of clothing to “needy students” are procured and delivered by local authorities.118 As of March 2013, there were 29,407 needy students enrolled in the program (14,776 girls and 14,631 boys). The Needy Children Program is for needy youths who are out of school, including orphans that are 18 years and over and no longer qualify for the Orphan Care Program but still are considered to have needs. They are provided with support depending on their needs, which can be extended until they are 29 years old. As of March 2013, there were 1,020 children in the program (511 girls and 509 boys). 117 Ministry of Local Government (2010b), page 33. 118 Ministry of Local Government (2010b), page 42. 45 Sponsorship and Scholarships 103. Sponsorships and Scholarships. The objectives of sponsorships and scholarships programs are multifaceted, with a strong focus on human capital accumulation and skills development. This report includes them under social assistance for cross-country comparability purpose and because this may be an area of cost savings. A full evaluation of the cost effectiveness of the scholarship program and therefore its potential for cost savings is beyond the scope of this study. 104. Sponsorship is the support provided by the GoB to high school graduates for studies in national institutions; scholarship is supported provided by the GoB for studies abroad. The MESD’s Department of Tertiary Education Financing (TEF) is responsible for the program.119 All students who complete high school with a passing grade can apply and obtain a sponsorship to study in the country. Scholarships to study abroad are also granted, mostly for specialties not available in the country or for graduate work. Students are expected to return home and work for the GoB for a couple of years. In April 1995, the government introduced a reorganized bursaries scheme in which students are awarded grants and loans to pursue their tertiary training. The beneficiaries are expected to pay back the loan portion of their sponsorship as soon as they become employed.120 However, according to the TEF, the recovery rate is less than 5 percent. A few students also do not return home after completing their studies abroad. The sponsorships pay for tuition fees, books, equipment, living allowance, warm clothing, medical fees, insurance, and other costs. The tuition and other fees are paid directly to the institution by the program, while the allowance is given in cash to the student. The average annual cost per student studying in local institutions varies between P36,000 (Arts) and P63,000 (computing). For overseas 119 The criterion is to have at least a C average in six subjects. This is equivalent to a 36 score. Recently this minimum score was lowered to 34. The criterion is also lower (-5) for OVC, remote area dwellers, and for people with disabilities. Currently their score is 29. 120 http://www.moe.gov.bw/ 46 studying, the annual cost varies between P479,000 and P547,000. The living allowance for those studying within the country is P1,420 per month (US$ 187) for 12 months, which is equivalent to twice the food poverty line or two minimum wages in manufacturing. In 2011/12, 11,252 new sponsorships and 31 scholarships were awarded. In addition, there were other beneficiaries—187 OVC, 370 students in families benefiting from the Remote Areas Development Program (which supports the most isolated communities in the arid parts of western Botswana, see below), 48 from special education, 9,118 who were progressing in their studies, 7,109 who had returned to school to obtain a higher degree, 95 top achievers, and 297 other beneficiaries. This came to a total of 28,507 students with sponsorships and scholarships. In 2012/13 the program spent P1,674 million, including P30 million on overhead expenditures, making it by far the most expensive social assistance program in Botswana.121 105. In the past there has been no clear policy on what type of studies to support. In recent years, the TEF has supported many IT students, which has led to a large overinvestment by private university and colleges in this area, an investment which indirectly was paid for by the TEF. Currently the TEF is reducing its support for IT and public administration courses. It is also seeking advice from the Labor Market Observatory and the Human Resource Council as to what courses to support. The TEF has 137 permanent staff and 15 interns from the Internship Program. Public Works 106. Ipelegeng. The Ipelegeng (self-reliance) labor-intensive public works program was made permanent in 2008 as an instrument of poverty alleviation both in urban and rural areas. It replaced a long series of drought relief “food-for-work” programs dating back to 121 Information received from the Department of Tertiary Education Finance (TEF). 47 Independence. Wage payments were introduced during the severe drought in 1982.122 The program is managed by the MLG’s Department of Local Government Development Planning, Community Development. The overall number of working places is decided by the Cabinet for each monthly working cycle. The program then allocates the monthly number of working places to constituencies based on their share of the total Botswana population. The program manager has indicated to the team that the incidence of poverty and the availability of basic infrastructure in each constituency are also factored in when deciding the allocation of places and that the districts have some flexibility in reallocating places within their jurisdictions. All those over 18 years old may apply for work. They must have their omang (the national identity card) and be willing to work for the program’s rate of pay. The program was originally designed to be self-targeted. However, because of excess demand and according to the program guidelines, recruitment is done by the Village or Ward Development Committee (VDC/WDC). Workers need to reapply every month and if there are places they are engaged. If there are not enough places, then those who did not work the previous month have priority. Since there is excess demand in most areas, a lottery is used to select the beneficiaries. Only in Gaborone is there often an excess supply of places owing to the rigid allocation formula. In May 2013 there were 56,274 people working in the program. These consisted of 39,998 women (71 percent) and 16,276 men.123 Of the total, 18,034 (32 percent) were working in urban areas while the rest were in rural areas. 107. The program’s pay is currently P480 for six hours of work over a period of 20 or 22 days depending on the number of days in the month. Supervisors receive P560 per month. In addition, starting in 2012/13, a meal is now supplied to the workers at a cost of P5 per day. The meal adds another P100 to the monthly pay for a total of P580 per month 122 BIDPA (2010). 123 Information received from program managers. 48 (US$76.3), which is 30 percent above the monthly agriculture minimum wage of P445.124 It should be mentioned that the meals are supplied by people from the community and therefore the money stays in the local economy and provides additional jobs. In 2012/13 the program cost P408.6 million, including P38.6 million in overheads. Of total expenditure during the last five years, 68 percent was spent on labor, 21 percent on materials, and 11 percent on overheads, which included personnel emoluments, motor vehicle fuel and maintenance, and small tools. 108. The projects executed by the program are decided based on the established project prioritization mechanism. The VDC/WDC initiates the process by identifying the community’s priorities; these are then reviewed by the local authorities and approved by the full local council. The projects are then reviewed by the program and/or specialized government agencies depending on the type of project. The quota of working places for the particular constituency helps to determine what projects can be initiated. All projects financed by the program must be completed. 109. It should also be mentioned that Ipelegeng is supporting many public institutions. For instance, in May 2013, of the 56,274 people employed by the program, 6,590 were crime prevention volunteers and 504 special constables (police assistants). The program has also helped the Ministry of Environment, Wildlife, and Tourism with the so-called Green Scorpions (environment voluntary police) who are hired for one year and the wildlife volunteers who are hired for three months. An undetermined number of Ipelegeng workers help local councils in various tasks. 124 Effective since April 1, 2012. 49 Other Transfers 110. Community Home-based Care (CHBC). This program was established in 1995 in response to the HIV/AIDS pandemic. The program is implemented through a partnership between the Ministry of Health and the MLG’s DSP, Division of Destitute Persons and Old Age Pension. The CHBC was designed to provide quality care at home for terminal patients as health facilities were becoming overwhelmed. The program now also covers people with other chronic illnesses such as diabetes that require special diets. Patients enrolled in the program benefit from clinical medical assistance as well as a food basket that is aimed at meeting their nutritional needs, including feeding tubes if necessary. Eligibility for assistance under this program is in theory “means-tested” and similar to the Destitute Persons Program.125 The number of supported CHBC patients declined over the last few years, mostly because of the introduction of free ART treatment, with numbers dropping from a highpoint of 12,988 beneficiaries in 2003/2004 to 1,208 as of May 2013.126 Some patients receive a food coupon or smartcard similar to the Destitute Persons Program, while those requiring special feeding support receive specialized foods (tube feeding). The cost of the food basket ranges from P500 to P1,200 (US$65.8 to US$179). The program cost P18.1 million in 2012/13. 111. Transfers to NGOs and other CSOs. The MLG transfers funds to various organizations that cater to children, youth, and other groups at risk. These include organizations that manage children’s homes, homes for abused women, rehabilitation homes for drug addicts, and homes for the elderly. The MLG also funds organizations that provide counseling and help to reintegrate children in conflict with the law back into society. Among these organizations are: SOS Botswana, Marang Child Care Network Trust, Childline Botswana (foster parenting), Ark and Mark, the Masiela Trust Fund, Mpule 125 Ministry of Local Government (2010a), page 93. 126 Republic of Botswana (2013), page 6. 50 Kwelagobe, the Kagisanyo Society Women’s Shelter, the Ikago School of Industries , Hajee Goolam Mustapha, and Lorato House/Love Botswana. The transfers to these organizations in 2012/13 amounted to P7.8 million. 112. Remote Area Development Program (RADP). This program was established in 1974 as the Bushmen Training and Settlement Project. Beneficiaries come from marginalized communities in remote and arid parts of western Botswana where there is very little economic activity.127 These communities have traditionally been nomadic hunter-gatherers who did not engage in arable agriculture. Many are San (Basarwa), but there are other smaller ethnic groups as well. The program is under the responsibility of the recently established Community Development Department (CDD) in the MLG, which operates in 67 designated settlements.128 113. The RADP originally financed small projects meant to raise the livelihoods of remote area dwellers. However, the program was evaluated in 2009 and recommendations to re-orient the program were approved in 2010.129 The revised program focuses on community-led development (rather than the provision of services), the creation of sustainable livelihoods systems, and an affirmative action program for disadvantage groups. With the support of the European Commission, the CDD has developed a strategic framework for community development and detailed operational guidelines for the implementation of the sustainable livelihoods approach, including for the preparation of community action plans. The CDD currently has only five trained community development 127 The RADP’s definition of remote area communities is: settlements with population of between 250 and 299 and at least 15 kilometers from the nearest recognized village, with a recognized traditional leadership, without water rights or land title, and distant from basic health and education services (White and Devereux, 2011). 128 The Ministry of Local Government’s Department of Social Services has been recently reorganized into two departments, the Department of Social Protection, which will continue oversee most transfer programs, and the Community Development Department, which is in charge of community development, the RADP, and part of the Alternative Package Program. At this point, it is still not clear where the Ipelegeng program will be located. 129 Ministry of Local Government (2009) and Ministry of Local Government (2010c). 51 officers, but there are 48 Batswana being trained in community development in Kenya. The CDD is in the process of reorganizing, training its staff, and developing a M&E system. In 2012/13, it cost P1.5 million. 114. The Alternative Packages Program (APP). The Alternative Packages Program (APP) and the Livestock Management and Infrastructure Development Program (LIMID) (see below) are part of the President’s Poverty Eradication Initiative. They seek to eliminate absolute poverty by, among other things, capitalizing the poor households and helping them to develop sustainable livelihoods. The Poverty Unit in the Office of the President is coordinating the Poverty Eradication Initiative. The APP is implemented by the MLG’s Community Development Department (CDD) and the Ministry of Agriculture. The program provides financing, training, and equipment for micro-businesses, ranging from bakeries to handwork as well as agricultural activities like beekeeping and poultry. There are 22 packages. The CDD manages 10 packages (jam and pickles, food catering, bakery, home- based laundry, leather works, textile and fashion design, hair salons, and glass recycling), and the Ministry of Agriculture (MOA) manages four packages (small stock, beekeeping, poultry, and backyard gardens) through its LIMID program (see below). Other packages are managed by the Ministry of Environment, Wildlife, and Tourism (backyard trees, nurseries, landscaping, fisheries, and the harvesting of natural products), and training and business advisory services are provided by the MESD and the Ministry of Trade and Industry respectively. 115. The program targets among others the beneficiaries of the Destitute Persons and Ipelegeng Programs with the objective of “rehabilitating” and turning them into micro - entrepreneurs. According to program guidelines, eligible people are “identified through assessment by [the] MLG; able-bodied Botswana citizens aged 18 years and above; and registered or unregistered destitute persons.”130 The program provides training, assets, and 130 Republic of Botswana (2012), page 8. 52 initial working capital as well as a transfer for the maintenance of the beneficiary until the business is operating and generating revenue (this transfer could come, for example, from the Destitute Persons Program). All assets and operating inputs are provided as grants. The cost of each package (initial investment only) varies from P4,357 (US$573) for hair salons to P34,923 (US$4,595) for jam and vegetable pickers (equivalent to 78 monthly agricultural minimum wages). As of May 2013, the APP component implemented by the CDD had identified 6,749 potential clients, had provided training and supplies to 1,969 clients who were ready to start operations, and had 785 micro-businesses in operation. The program is now putting in place a monitoring system to accompany the development of these micro- businesses. According to information received from the MFDP, the Poverty Eradication Initiative cost P104 million in 2012/13.131 116. The Livestock Management and Infrastructure Development (LIMID). The Livestock Management and Infrastructure Development Project (LIMID) is a farmer support program that is also part of the President’s Poverty Eradication Initiative. The program started in 2008 and is managed by the Ministry of Agriculture. Its objectives are to promote food security by increasing the productivity of raising cattle and small animals such as goats. It provides small stock, guinea fowls, and Tswana chickens to resource-poor farmers who own four cattle or less or 20 sheep/goats or less or who earn less than P150 per month with dependents or P120 per month without dependents. In 2012/13 the program supported 3,586 farmers. 2. Active Labor Market Programs 117. Severance Payments. Botswana has no unemployment insurance. Its formal sector enterprises are required by law to offer severance pay to workers who are dismissed. Employees with 60 months of continuous employment are entitled to a severance benefit 131 Ministry of Finance and Development Planning’s Development and Budget Division. 53 from their employer amounting to one day of basic pay per month worked. Under the 2010 amendment to the Employment Act, in case of termination of the employment contract before five years, the employee is entitled to a severance benefit that is proportional to his or her length of service.132 As is well known, in general severance payments are a poor protection against unemployment because they are not fully funded and are inefficient because they reduce labor mobility.133 118. The National Internship Program. This program was established in August 2008 following a Presidential Directive. This was done in an effort to fight unemployment among young graduates. The program is managed by the Ministry of Labor and Home Affairs’s Department of the National Internship program (DNIP). The program targets unemployed graduates who have completed their Bachelor’s Degree and Diplomas. The graduates are placed in workplaces (whether in the government, the private sector, NGOs, or parastatals) to enable them to gain work skills and enhance their employability. 119. The internship is for a period of up to 24 months of service, within which period the interns are expected to have gained the necessary skills to increase their marketability and employability. Interns are supposed to get “guidance [as] to proper behavior, expected standards, [and] pertaining policies” and to submit monthly a form to the Department so that “their progress can be evaluated.”134 The interns are paid an allowance of P2000 on a monthly basis (US$263), which is equivalent to almost three manufacturing minimum wages, with P1800 being paid monthly to the intern and P200 retained and paid at the end of the internship. Those entering the program after March 31, 2013 will be paid P1,400 (US$184). As of May 2013, there were about 3,750 interns in the program, 1,500 in the private sector, 1,500 in government offices, 600 in parastatals, and the remaining 150 in 132 http://www.ssa.gov/policy/docs/progdesc/ssptw/2010-2011/africa/botswana.html 133 See for example Ribe et al (2010), page 45 on the issue of poor protection and page 77 on the issue of labor mobility and Holzmann and Vodopivec (2012) for a full discussion. 134 http://www.gov.bw/en/Ministries--Authorities/Ministries/Ministry-of-Labour--Home-Affairs- MLHA/Department--Units/Departments--Units/National-Internship-Programme/ 54 NGOs and other organizations. The rate of absorption into permanent employment is reported to be about 45 percent in the public sector and 10 percent in other sectors for a total of 55 percent of the interns.135 The National Internship Program is currently conducting a tracer study to find out how graduates benefitted from their participation in the program and to explore ways of improving it. The program cost P72.8 million in 2012/13, including P9 million of overheads.136 120. The Apprentice Program. In 1983, Botswana passed the Apprenticeship and Industrial Training Act. The act established employer-based training through apprenticeships. In 1986, the trade testing center, the Botswana Training Center in Gaborone, was expanded and renamed the Madirelo Training and Testing Center (MTTC). The center manages the Apprentice Program. 137The requirements to enter the Apprentice Program are: having completed a minimum of junior certificate, being medically fit and at least 16 years of age; and having a sponsoring company that specializes in the same vocation. The sponsoring company and the apprentice sign a contract that stipulates that the company agrees to “employ the apprentice for the purpose of providing a course of approved instruction in accordance with approved Training Standards as per Apprenticeship and Industrial Training Regulations so that the apprentice may achieve necessary proficiency in the contracted vocation.” The contract also specifies the remuneration to be paid to the apprentice in each of the three or four years as well as the obligations of the apprentice. The contract is registered with the MTTC. This is a three-year or four-year program depending on the type of course. The courses depend on the specialization chosen by the apprentice and agreed with the sponsoring company. The apprentice goes for institutional training at a Technical College for three months and is tested on theory at the end of that time. The apprentice then goes on for practical on-the-job-training for nine 135 Information received from the program manager. 136 Republic of Botswana (2013/14) and communication received from the Ministry of Finance and Development Planning’s Development and Budget Division. 137 World Bank (2012d), page 15. 55 months and is assessed on his or her practical skills. The apprentice is allowed to progress to the next year when he/she has passed both theory and practice exams. At the end of the three or four years of training, the apprentice sits for the final National Craft Certificate Examinations. In 2012 the program graduated 949 apprentices and cost P20.1 million. 121. The Youth Development Fund (YDF). The Youth Development Fund was created in April 2009 with several goals: (i) to promote the active participation of young people in the socioeconomic development of the country; (ii) to reduce rural-urban migration by making it attractive to start growth-oriented citizen-owned youth enterprises: (iii) to encourage out-of-school, marginalized, and unemployed youths to venture into viable income- generating projects: and (iv) to create sustainable employment opportunities for young people through the development of sustainable projects. The program is implemented by the Department of Youth within the Ministry of Youth, Sports and Culture. The program targets youths between the ages of 18 and 35 who are not in school or employed and who have no particular academic background. The youths prepare simplified business plans that are appraised by local district youth officers. There are 32 youth officers in total. The appraisal takes into consideration whether the location of the proposed business is suitable, the profile of the applicant, and the commercial prospects of the venture. The District Economic Development Councils, local NGOs, and the Chamber of Commerce may also participate in the process. The program provides 50 percent of the funding as a grant and 50 percent as interest fee loans with the objective of establishing a revolving fund. However, since the recovery rate is less than 10 percent, the revolving fund does not operate. The maximum amount of the loan/grant for individuals is P100,000 (US$13,158) and for groups is P450,000 (US$59,210). There is a grace period of between 6 and 20 months and a repayment period of five years. The youth officers are charged with the responsibility of monitoring the implementation of the projects. Since 2009, there have 56 been about 3,000 beneficiaries of the program and eight beneficiary groups. In 2012/13 the program financed 1,061 projects for 1,117 beneficiaries and cost P96.7 million.138 122. The program is attempting to increase its effectiveness by: (i) hiring private consulting firms to provide short courses on business skill management to the beneficiaries; (ii) developing a business mentoring system particularly with expatriates; (iii) helping beneficiaries to market their products, for instance, at trade fairs; and (iv) promoting associations of entrepreneurs. 123. The Youth Empowerment Scheme. This is an umbrella program that coordinates various youth-related interventions. It is managed by the Ministry of Youth, Sport, and Culture under the Department of Youth. It was introduced in mid-2012 to address the youth unemployment challenge, to stimulate the spirit of volunteerism and community participation among young people, to enable youths to develop a good work ethic and to gain entrepreneurship and employment readiness experience, to empower youths with skills that will enable them to make a positive contribution to the development of their communities, and to reduce youth development and to eradicate poverty among young people. The program targets the unemployed and out-of-school youths aged 18 to 35 years with secondary or university qualifications. In 2012/13, the scheme invested P24 million in eight interventions/programs: • Life Skills and Capacity Building Program.139 This involves imparting skills to young people in areas such as horticulture, construction, character building, and the general discipline of a work environment. This program also includes bootcamps and promotion of the visual arts. 138 Information received from the project manager. 139 The major component of this program is the Construction Industry Fund (CITF) crash course that offers skills training to youths to enable them to become semi-skilled. 57 • Youth Entrepreneurship Program. Under this program, young people are provided with business awareness, counseling, mentorship, and financial planning skills to enable them to start viable and sustainable business enterprises. This program is delivered among others through the Youth Development Fund. • Sport and Cultural Development Program. This program deals with the constituency sport tournament and art competitions. Youths are engaged to mobilize and manage the teams and groups. Talented youths can receive scholarships to enable them to attend high-quality sports and arts institutions outside the country. • Basic Computer Literacy Training. This program trains young people with IT skills to provide computer lessons to members of their communities using facilities provided in the public libraries. • Public Education Program. This involves conducting public education campaigns and activities to educate and inform communities about key policies and programs using various methods such as drama, theatre, and talk shows. • Retooling Program. This entails equipping unemployed graduates with new but appropriate skills that will enable them to meet current labor market demands as well as arming them with the right attitudes to participate meaningfully in the national economy. It entails sponsoring young graduates for masters and PhD training in research in some strategic disciplines of economic demand. • Attachments to Mega Projects. Young people are attached to large construction projects in which the government has a stake. This initiative targets young people with skills such as carpentry, welding, pipe-fitting, and brick laying. Semi-skilled 58 and skilled artisans graduating from the Construction Industry Fund are employed on the projects in order to minimize the need to import skilled labor. • Attachments to Public Projects. Under this initiative, youth volunteers are placed in different government ministries or departments or in parastatals. They are attached to specific programs for a period of 12 months, in such areas as agricultural extension, health, and community services. They must have applied for admission into the scheme and must not be in any remunerated employment elsewhere. Participants are posted to different host organizations throughout the 12 months of service in accordance with their area of interest. When the participants complete the service successfully, they are issued with certificates of participation in recognition of their work. Participants are allowed to withdraw or terminate their service by giving at least one month’s notice, in which case they are not awarded the Certificate and cannot be accepted back into the scheme in any later date, except under very special circumstances and only with the approval of the Permanent Secretary. The scheme provides a living allowance as determined by the Permanent Secretary of the Ministry of Youth, Sports, and Culture. Currently the allowance is P500 per month. 3. Social Insurance 124. In 1995, the GoB created a non-contributory, defined benefit pension scheme for public officers (civil servants) financed directly from the budget. In 2001, the government decided to change the Public Officer’s Pension plan from a defined benefit pension scheme to a defined contribution pension scheme, the Botswana Public Officers Pension Fund (BPOPF). The BPOPF covers public servants at both the central and local levels. It was created in accordance with the provisions of the Pensions and Provident Fund Act (1987) and is managed by a Board of Trustees that comprises nine employers, nine employees, one pensioner, and one independent appointed by the Board of Trustees. The BPOPF is 59 supervised by the Non-Bank Financial Institutions Regulatory Authority (NBFIRA), which is responsible for licensing and supervising all public and occupational pension and provident funds, fund administrators, and investment managers. The NBFIRA has the authority to conduct investigations, impose sanctions, and disqualify service providers or plan managers.140 125. The BPOPF’s members contribute 5 percent and public employers 15 percent of the member’s salary to individual savings accounts. A member can make an additional voluntary contribution to his/her pension account of up to 10 percent of his/her salary. 141 This amount is exempted from tax. The compulsory retirement age is 60 years old, and the early retirement age is 45 years old. As of March 31, 2012, the BPOPF had 98,226 active members, 6,518 deferred (or inactive) members, and 6,619 pensioners. Table 10 presents BPOFP contributions, pensions, and lump sum payments in the last two years. The BPOPF’s total net assets as of March 31, 2012 were P33 billion or 31 percent of GDP. 142 Table 10: BPOPF contributions and expenses, 2010/11-2011/12 (P million) 2010/11 2011/12 Member contributions 1,537.0 1,602.2 Employee, statutory 383.1 396.7 Employee, voluntary 1.3 1.5 Employers/GoB 1,152.6 1,204.0 Pensions and Lump sum payments 1,844.9 1,790.9 Pensions 337.6 380.2 Lump sum payments a/ 1,507.3 1,410.7 Source: BPOPF (2012), page 12 and 13 a/ Includes lump sum payments on withdraw, death and retirement in lieu of monthly pensions 126. In addition to its contribution to the BPOPF, the GoB spent P212.5 million in pensions and gratuities in 2011/12. These included the defined benefit scheme and other pensions and gratuities for non-pensionable officers, military personnel, members of 140 International Organization of Pension Supervisors (2001). 141 http://83.143.28.92/bpopf2012/employer-information. 142 BPOPF (2012). GDP for FY 2011/12 was P107 billion (CSO, 2012). 60 parliament, and the President. For 2012/13 the budget for these pensions was P348.3 million.143 127. Members of the BPOPF who leave their job after the age of 45 years are retired from the Fund and paid one-third of their fund credit as a (tax free) cash lump sum and the remaining two-thirds is used to purchase an annuity from any registered insurer in Botswana. Members who leave their job before reaching the minimum retirement age of 45 are entitled to a cash lump sum of 25 percent of their fund credit while the remaining 75 percent of their benefits are deferred and may only be accessed when the member reaches the retirement age. The accumulated pension savings are portable so a member who leaves the Fund can exercise any of the following options: • Leave all the benefits with the Fund until he or she reaches early retirement age. • Transfer all the benefits to any registered pension fund or insurer in Botswana. • Cash 25 percent of the fund credit and leave 75 percent with the Fund only to be accessed after reaching the early retirement age. • Cash 25 percent of the fund credit and transfer the remaining 75 percent to any registered pension fund or insurer in Botswana. 128. The BPOPF, and other retirement funds, are authorized to invest up to 70 percent of their investment value in offshore assets. The BPOPF outsources the management of its portfolios (the active member and deferred pensioner portfolio, the non-profit pensioner portfolio, and the for-profit pensioner portfolio) to specialized fund managers. Its objective is to obtain a real return of 5 percent over the longer term. The BPOPF’s members nearing retirement have the option to switch from the more aggressive portfolio (in other words, the active member and deferred pensioner portfolio) to the more conservative portfolio, the main objective of which is capital preservation rather than capital growth. The option 143 Republic of Botswana (2013/14), page 403. 61 to switch can only be exercised by members who intend to retire within the following three years. 129. As mentioned, there are several private and voluntary occupational schemes that are supervised by the NBFIRA. Occupational pension plans are usually established by medium-sized and large employers. These systems together cover 49,752 formal sector employees. Pension plans are financed both by employers and employees. Contributions depend on plan rules. The total employer and employee contribution rate under an average plan is between 10 percent and 15 percent of the contributory salary (typically, 5 percent in employee contributions and 5 percent to 10 percent in employer contributions).144 According to the NBFIRA, the 97 existing private pension plans received P648.5 million in contributions from employees and employers in 2011/12 and paid out P75.1 million in pensions. 130. Other benefits such as sickness and maternity leave are provided under the amended 2010 Employment Act. The Act provides 20 days of paid sick leave a year. The maternity benefit is at least 50 percent of the basic pay and is paid six weeks before and after the expected date of childbirth. Work injury is an employer liability system normally insured with a private company. 131. To sum up, Botswana has an extensive safety net, including several universal programs (the state Old Age Pension and the Primary and Secondary School Feeding Programs) and a series of categorical programs for orphans, the destitute, needy children, the unemployed, and those the incapacitated by illness. It also has a number of large ALMP programs, though employment services and unemployment insurance are non-existent. On the other hand, it has very fragmented social insurance programs, the most important being the defined contribution pension plans for public employees. 144 Oxford Policy Management (2010), page 85. 62 C. Informal Social Safety Nets 132. There is a general consensus in the literature that, throughout Africa, informal social protection arrangements play a crucial role in assisting individuals, families, and communities to manage risk and to provide support to vulnerable groups such as older people, orphans, the sick, and those who have lost their loved ones. In Botswana there are four types of informal social protection arrangements, namely: (i) family and kin obligations; (ii) community support networks; (iii) civil society; and (iv) mutual aid associations. A brief discussion of these follows. 1. The Extended Family and Kin Support Systems 133. In Botswana, the extended family system plays a significant role in providing a safety net when individuals encounter lifecycle crises such as hunger, sickness, death, old age, unemployment, and injuries. Propelled by the principles of solidarity and generalized reciprocity, the family provides a window of support during these hard times. It has been contended that traditionally the system embraced a group of families closely united by blood or marriage.145 Members came together to carry out important domestic as well as economic activities such as building or thatching huts, clearing fields, weeding, and reaping as well as assisting one another with livestock and other commodities and exchanging gifts. This continues to happen in most parts of Botswana today. At times well-to-do families foster or formally adopt the children of their relatives. The concept of motlhoki or a destitute person has always existed in the Setswana vocabulary. 134. It has been shown that orphan households in Botswana typically received support from other household members (43 percent), from relatives (39 percent), or from the 145 Schapera (1970). 63 government (34 percent).146 However, 15 percent of households received no assistance at all. Another study that examined the views of caregivers or parents about the Orphan Care Program in Botswana found that caring for orphans was putting a heavy financial burden on many low- and middle-income households.147 The study concluded by suggesting that extended family caregiving may be sustainable with the support of an adequate public sector safety net. The study also called for the implementation of a family-centered approach by offering services specifically for orphan caregivers/parents. 2. Community Support Networks 135. Community support networks also play a critical role in meeting the spiritual, social, and economic needs of people experiencing crises. In the pre-colonial era, this assistance was provided by the village chiefs, neighbours, and other community members. Although this type of support is increasingly diminishing due to urbanization and industrialization, it is still practiced in some rural areas. For example, prior to independence in 1966, the chief occupied a position of unique privilege and power over his people,148 and it was under his auspices that certain informal social protection initiatives were arranged. The chief was responsible for maintaining law and order, presiding over cases, and administering tribal affairs as well as looking after the general welfare of the people. In many Tswana ethnic groups, two main traditional social protection schemes were regulated by the chiefs, namely masotla (tribal fields) and kgamelo (milk paid cattle). While the chiefs still wield considerable power, the situation appears to be changing, although there is anecdotal evidence to suggest that some of the traditional initiatives are still in place in some parts of the country. 146 Miller et al (2006) based on data from the Botswana Family Health Needs Study. 147 USAID (2010). 148 Wass (1969) and Schapera (1970). 64 136. The term masotla (tribal fields) refers to massive fields that belonged to the chief. These fields were prepared by the mephato (groups arranged by age) on a voluntary basis before the rest of the community was permitted to plow the land.149 Produce from masotla (including sorghum, maize, beans, maize, millet, cowpeas, and groundnuts) was stored in tribal granaries (difalana). During a drought season or in times of other natural disasters, the food was then distributed to the community. Disadvantaged people such as orphans, destitute people, and chronically ill individuals were from time to time fed from the difalana. 137. Kgamelocattle (milk paid cattle) were at times given to the poorer members of the community by the chief as a means of livelihood, but more commonly, they were entrusted to prominent members of the tribe who would herd them.150 The responsibility of the herdsman was to milk the cattle on behalf of the chief. The milk was then put daily into a milk sack (lekuka) to thicken and then brought to the chief to be distributed. Besides providing for his wives, it was a common practice for the chief to supply milk to the poor, orphans, or other vulnerable groups in the community.151 138. Apart from these kinds of assistance provided by the chiefs, other support systems were organized by neighbours or by the community as a whole. Examples include the mafisa, majako, go tshwarateu, and molaletsa systems. The mafisa system (lending cattle to the poor) was a form of special contract by which well-off relatives or members of the community placed one or more cattle into the keeping of a destitute person. The recipient herdsman took care of these cattle for an indefinite period and had the right to use them as draught animals and as a source of milk and transportation to fetch water or firewood. If the cattle multiplied, it was common practice to reward the herdsman with a heifer 149 Denbow and Phenyo (2006). 150 Schapera, (1970) and Denbow and Phenyo (2006). 151 Schapera (1970), p. 248. 65 (kgomoyamafisa) together with any offspring that they may subsequently produce. 152 Ownership of the cattle remained with the original owner who could claim back the cattle after an agreed period. 139. The majako system allowed poor people to sell their labor and work in the fields of the rich in return for a share of the harvest. Similarly, go tshwarateu or bodisa provided able-bodied poor people with an opportunity to break out of poverty by looking after other people’s cattle, and in return receiving a payment of a cow each year. Letsemaor molaletsa constituted yet another form of non-formal social protection strategy in that it allowed members of the community to perform voluntary work on behalf of a deserving family. For example, if a family wanted to clear a field to be plowed, the family would invite neighbours and other people in the community to a letsema. Any neighbor who refused to participate was made aware that she/he should not expect to be assisted when she/he called a letsema. 140. These indigenous support systems are no longer being practiced with the same intensity in most parts of the country due to the diminishing role of the chieftainship, reforms in the agriculture sector, and other social changes. No known empirical research has measured the impact of these arrangements on the well-being of beneficiaries. However, what is worth noting is that the spirit of community involvement still exists in Botswana. In general, community-based initiatives have tended to focus on mitigating the impact of the HIV/AIDS scourge as well as responding to poverty and unemployment issues. 3. Civil Society Organizations 141. Civil society organizations in Botswana also play an important role in assisting vulnerable groups. The Botswana Council of Non-governmental Organizations (BOCONGO) 152 Schapera (1970). 66 was formed in 1995 to co-ordinate the activities of all NGOs in the country. The purpose of BOCONGO is to create an enabling environment for the NGO sector to become a credible and recognized partner in the development process. BOCONGO acts as a forum for networking, advocacy, lobbying, and capacity building in order to strengthen the civil society sector in Botswana. The network is also responsible for mobilizing resources for the NGO sector. Since its formation, BOCONGO has grown to include 130 member NGOs. 4. Mutual Aid Associations 142. Mutual aid associations have become very common in contemporary Botswana. These are formed to respond to specific contingencies as determined by members in the spirit of botho, reciprocity, and solidarity.153 Burial societies also known as diswaeti are self- reliant institutions that can be found in many villages and urban settings in Botswana. These fall into three categories: (i) work-based societies that draw their membership from workplaces in urban areas; (ii) ethnic-oriented burial societies: and (iii) communal burial societies whose membership cuts across social and physical boundaries.154 In Botswana, burial societies receive no financial support from the government, NGOs, or external donors. Members collectively pay an agreed regular subscription fee that ranges from P5 to P30 (about US$1 to US$6). Older people and destitute people routinely use their monthly pensions and cash transfers from government safety net programs to pay for these subscriptions.155 143. The financial sector in Botswana has several informal organizations, including Rotating Savings and Credit Associations (ROSCAS), pawn shops, landlords, and financial transactions between friends and relatives. Apparently, this sector is dominated by 153 Ntseane and Solo (2007). 154 Ngwenya (2003). 155 Seleka et al (2007). 67 women.156 Savings and credit associations play a key role in meeting the financial needs of people who have limited access to the formal financial market. 156 Okurut and Thuto (2009). 68 IV. ASSESSMENT OF THE SOCIAL PROTECTION AND LABOR SYSTEM 144. This chapter assesses Botswana’s social protection system focusing specifically on: spending; coverage, gaps, and overlaps; generosity of benefits; targeting mechanisms; cost- effectiveness; incentive compatibility and sustainability; monitoring and evaluation; and institutional arrangements. A. Spending 145. The Government of Botswana spent P5.5 billion in 2012/13 or 4.4 percent of GDP on social protection and labor (SPL) market programs. This amount included P3.7 billion on social assistance (3 percent of GDP, including here sponsorships and scholarships for cross- country comparability purposes), P214 million on ALMP (0.2 percent of GDP), and P1.4 billion on public contributions to social insurance (BPOPF) and payment of defined benefit scheme pensions (1.2 percent of GDP) (Table 11). 69 Table 11: Botswana: Social protection and labor expenditures 2009/10 2010/11 2011/12 P millions, current price Social Safety Net 3,471 3,755 3,632 3,698 Cash and Near Cash 819 710 852 833 Orphan Care 312 267 426 301 Destitute Persons a/ 254 191 166 241 Word War Veterans 13 12 11 11 Old Age Pension 239 240 249 279 Conditional Cash Transfers 0 0 0 0 In-kind Food Transfers 355 425 462 651 Vulnerable Group Feeding 0 0 0 166 Program Primary School Feeding 219 242 290 275 Program Secondary School Feeding 135 183 172 210 Program Scholarships 2,037 2,311 1,991 1,674 Sponsorships/Scholarships 2,037 2,311 1,991 1,674 Public Works 219 277 301 409 Ipelegeng 219 277 301 409 Other Transfers 41 32 26 132 Community Home-based 32 24 16 18 Care MLG transfers to NGOs 10 8 9 8 Remote Area Development 0 0 1 1 Program Poverty Eradication Initiative 0 0 0 104 (APP) Active Labor Market Programs 87 129 184 214 National Internship Program 44 57 70 73 Apprentice Program (MTCT) 22 21 20 20 Youth Development Fund 20 50 93 97 Youth Empowerment Scheme 0 0 0 24 Pensions and Other Insurance b/ 1,112 1,243 1,329 1,435 Public Officers Pension Fund 917 1,047 1,116 1,087 Non-contributory Public Plan 195 196 213 348 Total Social Protection and 4,670 5,126 5,145 5,347 Labor Percent of GDP Safety Net 4.52 3.89 3.40 3.02 ALMP 0.11 0.13 0.17 0.17 Social insurance 1.45 1.29 1.24 1.17 Social Protection & Labor 6.08 5.32 4.82 4.36 Memo: GDP (P millions) c/ 76,753 96,411 106,848 122,500 Source: Program managers, MFDP’s Budget Books, and CSO (2012) Notes: a/ Includes Needy Students and Needy Children Programs. b/ Data from MFDP budget books. For 2012/13, revised budgeted amount. c/ For 2012/13, projected from Republic of Botswana (2013), Table3. Social assistance includes sponsorships and scholarships for cross-country comparability purposes. 70 146. Spending has been declining in recent years in GDP terms, from a high of 6.1 percent of GDP in 2009/10 to 4.4 percent of GDP in 20012/13. During this period, overall social safety net spending slightly decreased, with a relatively large cut registered by the Sponsorships / Scholarship Program and the CHBC Program. The Destitute Persons Program registered a decrease in 2010/2011 and in 2011/2012 when many beneficiaries were transferred to the public works program but increased again in 2012/2013. In constant 2009/10 prices, total SPL spending declined from P4.7 billion in 2009/10 to P4.2 billion in 2012/13 (9 percent) as spending on social safety net dropped by 15 percent (Figure 6). Figure 6: Social protection and labor spending 147. Near one-half of social assistance spending is on sponsorships/scholarships (45.3 percent). This is followed by cash and near cash programs (22.5 percent), in-kind food transfers (17.6 percent), public works (11.1 percent) and other transfers (3.6 percent) (Figure 7). In most developing countries, scholarship programs for higher education tend to favor the relatively well off because poor children have lower rates of secondary completion. 71 Figure 7: Distribution of social assistance spending by type of program, 2012/13 (percent) 148. Table 12 presents a comparison of social protection spending in Botswana with spending in other upper-middle-income countries and other African countries for which data are available. Overall social protection spending in Botswana (4.4 percent of GDP) is lower than the average for the other upper-middle-income countries (5.3 percent of GDP). Social insurance spending is significantly lower than the other countries while social assistance is substantially higher. In comparison with other African countries in the table, Botswana’s social protection spending is higher than the average both for social insurance and social assistance. This comparison should be interpreted with caution because the definitions (such as the scope of social assistance) may not be fully consistent across countries.157 In particular, in Botswana the spending for sponsorships and scholarships for tertiary students is very high. 157 For a discussion of some of the difficulties involved in making cross-country comparisons of social protection expenditures, see Marques (2002). 72 Table 12: Public sector spending on social protection (percent of GDP) Social Insurance Social Assistance Total Social Year Country SI SA Protection SI/SA Botswana a/ 1.2 3.2 4.4 2012/2012 Average of Upper-middle- 3.1 2.2 5.3 income Countries below Namibia b/ 1.3 2.8 4.1 2004/2012 c/ South Africa 2.2 3.5 5.7 2010/2011 d/ Costa Rica 2.8 1.5 4.3 2009/2004 Chile d/ 5.5 0.7 6.2 2010/2003 e/ Jamaica 2.6 1.8 4.4 2010/2010 f/ Panama 4.3 2.6 6.9 2008/2008 Average other 1.0 1.7 2.6 African countries Ethiopia 0.3 1.2 1.5 2006/2009 Kenya 1.3 0.8 2.1 2010/2010 Rwanda 0.7 1.1 1.8 2005/2011 Mozambique g/ 2.0 1.7 3.7 2010/2010 h/ Sierra Leone 0.5 3.5 4.0 2011/2011 Source: Social Insurance = Public pension spending from World Bank’ s (HDNSP) Pension database(http://web.worldbank.org/WBSITE/EXTERNAL/TOPICS/EXTSOCIALPROTECTION/EXTPENSIONS/0,,con tentMDK:23231994~menuPK:8874064~pagePK:148956~piPK:216618~theSitePK:396253,00.html). a/ Botswana: Table 8 in this report. b/ Namibia: budget amount from Ministry of Finance of Namibia (2011). c/ South Africa: Republic of South Africa (2012), Tables 1.2 and Table 8.5. d/ Chile and Costa Rica: World Bank’s Social Protection Databases- Wiegand and Grosh (2008). e/ Jamaica: Marques (2011a). f/ Panama: Marques (2009). g/ Mozambique: Marques (2011c). h/ Sierra Leone: Marques (2013). Other countries: World Bank (2013), Table 9. 149. To sum up, Botswana spends a substantial amount on social assistance programs compared to countries of the same income level or even when compared to OECD countries (2.5 percent of GDP).158 However, it spends relatively less on social insurance, in part because its mandatory pension system is limited to public employees. Sponsorships/scholarships represent near one-half of total social assistance expenditures. 158 Wiegand and Grosh (2008). 73 150. The amount spent on safety nets annually would be more than enough to completely close the poverty gap in Botswana and eradicate poverty. Assuming perfect targeting and a consumption gap of 0.38 (as in 2002/03), the annual cost of filling the consumption gap of the poor (19.3 percent of the total population) would be P407 million or 0.33 percent of GDP. The cost of bringing all families currently living in absolute poverty (16 percent of the total population) to the food poverty line (P680/family of four/month) would be P260 million, or 0.21 percent of GDP. Figure 8 shows spending on safety net transfers as opposed to spending on the one of the largest programs —scholarships (which generally benefit the non-poor) and relative to the national basic needs and food poverty gaps as estimated from the 2009/2010 BCWIS. To fill the total and food consumption gap, however, Botswana would need to put in place a targeted last-resort program that would have some inherent leakage toward non poor, which will increase the budget of the program compared to the perfect targeting scenario (see Chapter 5). Figure 8: Spending on SSN programs compared to the food and total consumption gaps of the poor, 2009/2010 74 B. Program Coverage, Overlaps, and Gaps 151. Botswana has several universal programs with good take-up rates. Table 13 presents the estimates of the size of at-risk groups and the coverage of each program. Five of the 14 programs in the table seek to cover all of the people in the defined category (the Orphan Care Program, the Primary and Secondary School Feeding Programs, the OAP, and the Veterans Program) and generally cover a large share of this population. One important exception is the Orphan Care Program, which only covers 30 percent of the target group in part because of the restrictive definition of orphans used by the MLG. Table 13: Program coverage Poor Population Target Program Targeting Age Groups Population Program 2012 group Coverage mechanism 2012 Pregnant/Lactating 90,000 N/A Women VGFP 383,392 Categorical d/ 262,271 80,595 0-5 Orphan Care 137,805 40,030 Universal/Low Needy 59,174 a/ 29,409 Low Students 12-20 404,650 88,655 Primary School 268,761 Universal/high Feeding Secondary School 161,929 Universal/high Feeding 12-20 404,650 88,655 Needy N/A 1,020 N/A Children Internship N/A 2,891 N/A Apprentice N/A 949 16,500 Sponsorships b/ 28,507 High Destitute 132,028 Means- a/c/ 30,518 21-64 1,050,648 148,465 Persons tested/Low CHBC 1,345 Means-tested Self- Ipelegeng 55,000 selection/Low OAP 93,639 Universal/high 65 and above 92,172 27,572 Veterans 2,049 Universal/high Source: Ministry of Local Government (2010), Table 3, and UNESCO Notes: a/ Absolute poor people. Assumes that three-quarters of the poor are absolute poor - see Statistics Botswana (2011a), Table 2. b/ Total number of students in tertiary education - latest available estimate from UNESCO. c/ Includes estimated absolute poor aged 21 years and over. d/ According to the MLG, beneficiaries are supposed to be screened at the clinics for need. 75 152. Other programs are not universal but have high coverage. In this category we should include the Sponsorships/Scholarships Program, which seems to cover a large proportion of students in higher education (the coverage is much higher than the total number of students in tertiary education reported by UNESCO for 2006, which is 16,500).159 Regarding the National Internship Program, which covers 3,750 youths, and the Apprentice Program, there is no information on the size of the target group. The VGFP is targeted to specific groups (children under 5 years old and pregnant and lactating women) and its coverage is high as it covers 383,400 people with food supplements. This program together with the Primary and Secondary School Feeding Programs cover 40 percent of the population. In this context, recall that the absolute poverty headcount is 16 percent and the poverty headcount is 20 percent. 153. In contrast, two programs that target the poor and are means-tested appear to have low coverage and significant exclusion errors—the Destitute Persons Program and the Needy Students Program.160 The latter program covers about one-half of the number of absolute poor children of primary school age; this may to some extent explain why a significant number of children are still out of school. The Destitute Persons Program covers about one-fourth of the estimated number of absolutely poor people (21 years and over). Finally, the labor-intensive public works program Ipelegeng also has limited coverage as there has been excess demand for the program, which has meant that a lottery is needed in most constituencies to select the beneficiaries. 159 The latest available information is from UNESCO, which estimated about 16,500 students in 2006, which compares to 28,507 beneficiaries of sponsorships/scholarships in 2011/12. 160 Note that the eligibility criteria are not clearly defined and we do not know who the beneficiaries are. Therefore, it would not possible to estimate the magnitude of exclusion or inclusion errors. 76 154. There are several small programs, none with sufficient capacity to identify all poor (or absolute poor) to provide at least a basic benefit and serve as a referral for other programs to select their own vulnerable group. None of the programs explicitly supporting Figure 9: Coverage of targeted programs and absolute and total poor (percent of total population) poor families or individuals has the potential to cover all 19.3 percent of the total poor even if perfectly targeted. Figure 9 compares the coverage of key targeted programs with the number of absolute poor and total poor. 155. There are significant program overlaps resulting from how the eligibility is established. A study by the BIDPA in 2007 using the 2002/03 HIES indicated that 9 percent of all households in the poorest quintile benefit from at least three programs and 30 percent from at least two programs.161 Even in the richest quintile, 6 percent of households benefit from at least two programs. Further analysis indicated that many of those that received benefits from the Destitute Persons Program also received benefits from the OAP, 161 Seleka et al (2007), Table 16. The analysis is restricted to the nine assistance programs covered by the survey (the Veterans Program, the OAP, the Drought Relief Program, the Destitute Persons Program, the Needy Students Program, the Orphan Care Program, Remote Area Development Program, CHBC, and the Clinic Ration). 77 Needy Students Program, the Orphan Care Program, and VGFP. These overlaps are a direct consequence of how the target group and eligibility criteria are defined as well as the fragmented approach to social protection. An approach that would focus on the family as opposed to individuals could eliminate unjustified overlaps of programs and would be more effective. 156. With regard to active labor market programs, there a large number of institutions with unclear lines of responsibility implementing a large number of programs. Intuitions implementing ALMP include the MESD, the Ministry of Labor and Home Affairs, the Ministry of Youth, Sports, and Culture, and autonomous bodies such as the Botswana Training Authority (BOTA), and the Tertiary Education Council (TEC). These institutions manage several programs including: out-of-school youth programs, training programs, youth employment and skill development programs, the Labor Market Observatory and Labor Exchange, the Scholarships/Sponsorship program, the Internship Program, and the management of the 0.2 percent turnover tax to finance vocational training (more on this on Section IV.H). 157. Social insurance is also fragmented. Of those in the formal labor force in 2012 (about 379,000), 41 percent were active members of a pension scheme. This includes 98,226 members of the Botswana Public Officers Pension Fund (BPOPF), which covers central and local government employees, and 47,970 formal sector employees who are enrolled in the 97 existing stand-alone private pension funds (including five umbrella funds that pool together another 168 small funds).162 The formal pension system only covers about 13 percent of the labor force. This compares to social security coverage in 11 other African countries ranging from 5 percent to 20 percent (Table 14). Of the elderly 60 years and above in Botswana, only 8 percent receive a retirement pension. 162 Data received from NBFIRA. In 2012, the total number of contributors to the BPOPF was 98,226 and to private funds was 47,979, excluding deferred or non-active contributors to the BPOPF (6,518) and private funds (1,782). 78 Table 14: Pension coverage Number of Labor Force ('000) Percent Year Active Contributors Botswana a/ 131,854 1,056 12.5 2011 Ghana 900 10,366 8.7 2010 Lesotho 38 883 4.4 2005 Madagascar 521 9,835 5.3 2009 Mali 341 4,297 7.9 2010 Mauritania 135 1,031 13.1 2000 Namibia 73 760 9.6 2008 Rwanda 196 4,281 4.6 2004 South Africa 1,213 18,163 6.7 2010 Swaziland 70 453 15.4 2009 Uganda 1,224 11,920 10.3 2004 Zimbabwe 1,000 5,012 20.0 2006 Source: World Bank Pension database a/ Botswana: contributors from NBFIRA and labor force from World Bank World Development Indicators http://web.worldbank.org/WBSITE/EXTERNAL/TOPICS/EXTSOCIALPROTECTION/EXTPENSIONS/0,,contentMDK: 23231994~menuPK:8874064~pagePK:148956~piPK:216618~theSitePK:396253,00.html 158. Out of 41 countries in Sub-Sahara Africa (SSA), Botswana is one of the five countries that do not have a mandatory public pension plan (Pillar 1). Most countries have a defined benefit scheme but four have provident funds/public managed defined- contribution schemes. As discussed, Botswana has a mandatory plan but only for public employees as well as a Pillar zero in its OAP, which is universal. Botswana does not have a mandatory privately managed scheme (Pillar 2) either.163 Out of the eight countries with an old age pension in Sub-Saharan Africa, one-half the programs are targeted while the other half are universal.164 163 The pillars are part of the World Bank’s pension reform conceptual framework. Zero Pillar = non - contributory basic benefits financed by the state, fiscal conditions permitting.First Pillar = mandatory with contributions linked to earnings and the objective of replacing some portion of lifetime pre-retirement income. Second Pillar = mandatory defined contribution plan with independent investment management. Third Pillar = voluntary taking many forms (such as individual savings; employer sponsored, defined benefit, or defined contribution). Fourth Pillar = informal support (such as family), other formal social programs (such as health care or housing), and other individual assets (such as home ownership and reverse mortgages) (Holzmann et al, 2008). 164 World Bank pension data basehttp://web.worldbank.org/WBSITE/EXTERNAL/TOPICS/EXTSOCIALPROTECTION/EXTPENSIONS/0,,content MDK:23231994~menuPK:8874064~pagePK:148956~piPK:216618~theSitePK:396253,00.html 79 159. To sum up, Botswana has a comprehensive social safety net that covers most of the risks that have been identified, but there are substantial program overlaps that are a consequence of the fragmented approach to social protection. On the other hand, the analysis points to three major gaps. One is a program that focuses on poor families to provide immediate relieve and overcome long term poverty through investment in human capital. Second, there is a need for a better targeted labor-intensive public works that allows poor unemployed to work for longer periods. Finally, there is need to increase the coverage of the old age social insurance as the existing pension system covers only a fraction of the labor force and to replace the inefficient and often unreliable severance payment by an unemployment insurance. C. Program Generosity 160. In assessing the generosity of the social protection programs, we should be concerned with four issues: (i) whether the existing social protection programs are sufficiently generous to have their expected impact; (ii) whether they are too generous and therefore serve as a disincentive for the beneficiaries to work and save and create a culture of “dependency”; (iii) whether they are too generous to be financially sustainable; and (iv) whether assistance programs undermine traditional social protection arrangements which in Botswana are already being disrupted by migration, rapid urbanization, and the HIV epidemic. If the programs are not generous, then the other issues should be less relevant. To address the first issue, we compared the benefits paid by the major social assistance programs with international benchmarks. Our focus is on the Old Age Pension, the labor- intensive public works program Ipelegeng, and the transfer program Destitute Persons for which there is comparative international information. Table 15 presents the benefits paid by these programs. At the end of this section we also evaluate the generosity of the public pension system. 80 Table 15: Program benefits, 2012/13 Estimated monthly Estimated monthly Program Type of Benefit value (P) value (US$) OAP Cash 250 32.9 Food P450 - 650 + Destitute 640 84.2 cash P90 a/ Cash P480 + Food Ipelegeng 580 76.3 P100 Source: Republic of Botswana (2013), page 6. For Ipelegeng Statistics Botswana (2012), Table 5 161. Table 16 shows global estimates of the generosity of various types of social assistance programs in developing countries as reported in a World Bank study. Program generosity is defined as the program transfer divided by the average consumption expenditures (excluding durables, housing, and health care) of the poorest households (those in the first quintile, Q1). Table 16: Generosity of social assistance programs in developing countries Programs Median transfer as Range transfer as (in parenthesis no. of percent of average percent of average HH Denominator countries) HH consumption consumption a/ Non-contributory (Social) 27 12-53 Q1 b/ Pensions (14) Last Resort Programs (20) c/ 23 5-45 Q1 b/ d/ Family Allowance (15) 18 6-28 Q1 b/ Source: Grosh et al (2008) Table 5.1 (page 137) Notes: a/ The two highest and lowest values were excluded. b/ Consumption expenditures of the poorest quintile excluding durables, housing, and health. c/ Last-resort programs are needs-based, usually means-tested, programs designed to help those who are not assisted or not assisted enough to keep them out of poverty by social insurance (pensions, unemployment insurance) or universal programs (such as child allowances or education). d/ Family allowances are cash transfers for families with children and can take various forms, such as means- tested child benefits, birth grants, or universal transfers for all children under a fixed age. 162. The Destitute Persons Program. This is one of the major social assistance programs in Botswana. The program may be considered a “last resort program” as defined in Table 16 above. The benefit paid by this program is equivalent to P640 per month including both the value of food and the cash allowance. Its generosity therefore is 77 percent of the 81 poorest household food consumption compared to a median of 23 percent in Table 16. 165 The benefit is equivalent to 143 percent of the agricultural minimum wage (P445 per month)166 and to 94 percent of Statistics Botswana’s BCWIS’s food poverty datum line for 2009/10 (P680 per month). Indeed, the program was characterized by some officials during our mission interviews as “very attractive.”167 There are reports that people do not want to graduate from the program, and in fact there is no record of people graduating from the program with the exception of those able-bodied recipients who were transferred to the Ipelegeng program in 2009. Finally, two new items have been added to the program— starting in 2013 all beneficiaries are to receive a blanket, and the elderly living alone will also receive a radio. The budgets for the added items for 2013/14 are P6 million and P2.7 million respectively. 163. The Old Age Pension. The state Old Age Pension provides a monthly cash transfer to all Batswana aged 65 years and older. It is a universal program. The transfer amounts to P250, which corresponds to a generosity of 30 percent of the poorest household food consumption. As can be observed in Table 16, the median generosity for such pensions in a sample of developing countries is 27 percent of the adjusted consumption of the poorest quintile, lower than in Botswana. The benefit is equivalent 37 percent of the food poverty line. 164. Labor-intensive Public Works. As can be observed in Table 17, setting the wage or compensation rate for public work programs below the minimum wage or the ongoing 165 We take food consumption to represent the adjusted consumption excluding durables, health, and housing. The mean total consumption of the poorest (Q1) household in 2002/03 was estimated by BIDPA researchers from the 2002/03 HEIS at P565 per month (P94.11 mean per capita with an average household size of six members). Mean food consumption is estimated at 65 percent of total consumption or P367 per month (CSO, 2004b, Table 91). This amount has been inflated by a factor of 2.25 to reflect the change in food consumer prices since 2003, bringing the mean food consumption of Q1 in 2012 to P826. 166 Commissioner of Labor and Social Security (2012). 167 Note further that we have assumed that the family only gets one basket when the program guidelines indicate that households will get two baskets if the number of members is five or more. 82 market wage (for the unskilled) is the norm in most public works programs in developing countries, with the exception of India, so that there is an element of self-selection. Table 17: Wage setting in public works programs Intensive Public Works Programs Wage setting Bangladesh’s Food-for-Work Program Program wage < market wage Bolivia Emergency Employment Program Program wage < market wage (about 2/3) Ethiopia Productive Safety Net Program Program wage < market wage (cash or food) India Maharashtra Employment Guarantee Scheme Program wage=minimum wage> market wage Malawi Public Works Program wage < minimum wage South Africa Expanded Public Works Program wage ≤ minimum wage Yemen Public Works Program wage < minimum wage Source: Grosh et al (2008), pages 487-492 165. In the case of Ethiopia’s Productive Safety Net Program (PSNP), the payment was determined on the basis of daily minimum food requirements rather than market rates, but its monetary equivalent was still below the market wage.168 In the PSNP in 2009, the average number of participating days per household per year was 150, and the average value of wages earned per household per year was US$137 (there is no limitation on the number of years of participation). This is estimated to represent about 40 percent of the annual food needs of the beneficiaries.169 166. In the case of South Africa, the Expanded Public Works Program (EPWP) launched in April 2009 (Phase 2) considered an average duration of employment of 100 days per year.170 The compensation or wage is set by the municipalities and provinces and is expected to be guided by the prevailing minimum wage in the area and in any ongoing poverty-relief projects. The aim is “not to displace workers from existing employment to new EPWP projects. Allowance is made for EPWP projects to pay below the minimum wage.”171 168 Subbarao et al (2010), page 9, Box 3. 169 World Bank (2010b), page 31. Note that the PSNP has two components: a public works component and a direct support transfer, which also averaged US$137 per household per year. 170 http://www.epwp.gov.za/ 171 http://www.epwp.gov.za/index.asp?c=FAQ#Rate 83 167. In Botswana, the Ipelegeng labor-intensive public work program pays compensation of P480 plus a meal that is equivalent to an additional P100 per month. Therefore, the total compensation is P580 per month (for six hours of work per day during five days of each week). This is 30 percent higher than the minimum wage for agricultural workers and equivalent to 85 percent of the absolute (food) poverty line of a family of four. When adjusted for time worked, the hourly remuneration under the Ipelegeng (included the meal) is 90 percent and 5 percent higher than the minimum hourly wage for agricultural and manufacturing workers respectively. 172 168. Retirement Pension. The average monthly retirement pension paid by the BPOPF in 2011 was P4,685 (US$616) (Table 18). The average pension is equal to over five times the poverty line datum of 2009/10. We have no information on the plan’s replacement rate, and since the average monthly wages at the central local, and parastatal levels vary significantly (P6,091, P4,629, and P13,500 respectively as of March 2011), it is not reasonable to estimate them on this basis. Nonetheless for future reference, in Chile and El Salvador, which have similar defined contribution systems as Botswana, the replacement rate is 44 percent and 39 percent respectively.173 Table 18: BPOPF average monthly retirement pensions 2011 a/ 2012 a/ Pensions (P) 337,622,957 380,192,279 Pensioners 6,006 6,619 Average monthly pension 4,685 4,787 (P) Source: BPOPF (2012) Notes: a/ Year ending in March 31 172 The Ipelegeng hourly wage is P4.4 (=P580/22 days/6 hours); the manufacturing minimum wage rate is P4.20 per hour; and the agricultural minimum wage rate is P2.29 per hour for a 45-hour working week. 173 Pallares-Miralles et al (2012). 84 169. To sum up, Botswana’s major social assistance programs are quite generou s. The benefits granted by the Destitute Persons Program and the OAP are high compared with the median of similar programs in several developing countries. Ipelegeng provides compensation for a six-hour working day that is 30 percent above the minimum wage for agricultural workers or 90 percent higher when adjusted for the number of hours worked. D. Targeting Efficiency 170. The target group and the beneficiary eligibility criteria are not clearly defined in some programs, while in others, the implementation of established criteria seems to be lax. For example, the eligibility criteria for the Needy Students and Needy Children Programs are not established. Some programs such as the Community Home-based Care use the Destitute Persons Program criteria. While the Destitute Persons eligibility criteria are clear (less than P120 income per month without dependents and less than four livestock units), in practice the criteria are not applied because the thresholds are considered outdated and most likely also because of the difficulty in verifying self-reported information. Consequently, beneficiaries are selected by MLG officials with the support of the Village or Ward Social Welfare Committees or local leaders. Also, the target groups are not clearly defined. For example, the Destitute Persons Program may include the elderly who cannot work and able-bodied persons who were “victims of an accident such as a train crash, plane crash, and motor vehicle accident or a natural disaster, ill health, or the death of the breadwinner.”174 Indeed, a large number of able-bodied adults were originally admitted into the Destitute Persons Program though they were later transferred to the Ipelegeng program. 171. Benefit incidence analysis helps to evaluate the targeting efficiency of the social assistance programs or the extent to which the poor benefit from the programs. The 174 Ministry of Local Government (2002), pages 6 and 7. 85 benefit incidence of a public program for the poor can be determined from the concentration curve, which shows the share of total benefits (vertical axis) going to the percentage of the population ranked by household income per capita (horizontal axis). 175 A program in which an equal amount of benefits is received by each quintile has a straight 45- degree concentration line. A pro-poor program is represented by a concentration curve located above the 45 degree line curve. A program is progressive if its concentration curve is above the consumption concentration curve. If the program’s concentration curve lies between the diagonal and the consumption curve, the program is not pro-poor, although it is progressive. If it lies below the consumption curve, it is not pro-poor and is regressive in the sense that its distribution is worse than the existing distribution of consumption. 172. In Botswana there is one major assistance program that is “means-tested”: the Destitute Persons Program. Figure 10 depicts the concentration curve for this program. The dotted line is the concentration curve (or Lorenz curve) of consumption. As can been seen, this curve is close to the axes, thus reflecting a large Gini coefficient or consumption inequality in Botswana. The concentration curve for the Destitute Persons Program is above the 45 degree line, which means that the program is pro-poor. Indeed, 52 percent of the Destitute Persons Program beneficiaries are from the poorest quintile (Q1). Its value on the Coady-Grosh-Hoddinott indicator176 is 2.60, which compares well with last-resort programs in Panama (0.88) and Mexico (1.61) but falls short of those in Costa Rica (2.90) and Argentina (2.68).177 175 The beneficiary incidence analysis in this section should be taken with caution. First, it relies on second-hand information reported in Seleka et al (2007). Second, the information is based on the 2002/03 HIES. Finally, since information on the resources received by those in each quintile is not available but only information on participation, it is assumed, as is customary in these circumstances, that benefits are equally distributed among the participants. In this case beneficiary incidence and benefit incidence are similar. 176 The Coady-Grosh-Hoddinott targeting efficiency indicator measures the ratio of the value of the transfer going to the poorest to the (relative) size of the same group in the population (Coady et al, 2004) 177 World Bank database on social protection indicators, Aspire:http://web.worldbank.org/WBSITE/EXTERNAL/TOPICS/EXTSOCIALPROTECTION/0, contentMDK:22986320~menuPK:8117656~pagePK:148956~piPK:216618~theSitePK:282637,00.html 86 Figure 10: Concentration curves for the destitute persons program 173. A program’s targeting accuracy can also be determined from the concentration index. This index is calculated from the concentration curve in a similar manner as the Gini coefficient is calculated from the distribution of per capita consumption. Negative values of the concentration index indicate a pro-poor program. The larger the index in absolute terms, the more accurately the program is reaching the poor. Figure 11 shows the concentration index for those social assistance programs for which there is information. We have also included programs whose primary objective may not be to target the poor but are directed to all people within a specific category to evaluate to what extent the poor benefit from these programs. As can be observed, the concentration indices are negative for all programs, which indicates that these programs are pro-poor. In Figure 11 we have also included the index for the group of programs that are means-tested and for those that are universal. As expected, the means-tested programs benefit more poor people than the universal programs. 87 Figure 11: Concentration indices for selected social protection programs 174. Finally, to evaluate whether the program helps to increase equity, a progressivity index can be calculated. The technique most commonly used to measure the progressivity of public expenditures is the method proposed by Kakwani, which is equal to the program’s concentration index minus the Gini coefficient of consumption. A program that increases equity presents negative values in its progressivity index. Table 19 presents the progressivity index for each program from the most accurate in reaching the poor (Destitute) to the least accurate in reaching the poor. As can be observed, all programs are progressive, implying that they all contributed to reducing the considerable inequity that exists in Botswana as measured by the Gini coefficient. 88 Table 19: Concentration and progressivity indices of selected programs Programs Concentration Progressivity Index Characterization DESTITUTE PERSONS -0.50 -1.13 PRO-POOR AND PROGRESSIVE NEEDY STUDENTS -0.49 -1.12 PRO-POOR AND PROGRESSIVE MEANS-TESTED -0.48 -1.11 PRO-POOR AND PROGRESSIVE DROUGHT RELIEF -0.42 -1.05 PRO-POOR AND PROGRESSIVE ORPHAN CARE -0.32 -0.95 PRO-POOR AND PROGRESSIVE OAP -0.28 -0.91 PRO-POOR AND PROGRESSIVE UNIVERSAL -0.27 -0.90 PRO-POOR AND PROGRESSIVE VGFP -0.27 -0.90 PRO-POOR AND PROGRESSIVE Source: Based on data from 2002/03 HIES 175. Summing up, the targeting of the one major social assistance program that is “means-tested” (the Destitute Persons Program) is relatively accurate, though there is always room to improve. The major problem in terms of targeting is the large number of poor people that are excluded from the Destitute Persons Program and the large universal programs that are not targeted to the poor (such as the school feeding programs).178 E. Cost-Effectiveness 176. Most programs do not have any comprehensive policies or operational guidelines, and in other cases the guidelines are out of date. This makes the implementation of the programs subject to the interpretation of the different front-line workers and therefore discretionary. On the other hand, it makes it impossible to hold these workers (and their supervisors) accountable for results. Table 20 presents the status of the guidelines for the major assistance programs. The Destitute Persons Program has a revised policy, but this still has not been approved. The Primary School Feeding Program, one of the largest social assistance programs, has guidelines that date back to 1993. 178 The errors of exclusion make the program ineffective in reducing poverty; the errors of inclusion make the program inefficient (Coady et al, 2004, page 11). 89 Table 20: Program policy and guidelines Program Type Latest Issued Destitute Persons Revised National Policy 2002 a/ Orphan Care National Plan of Action 2010-16 2010 CHBC Guidelines 2005 OAP Guidelines 1996 Veterans Guidelines 1998 RADP Revised policy, strategic framework, 2009-2011 and guidelines Ipelegeng Revised Guidelines 2012 Vulnerable Group Feeding Guidelines for health facilities 2011 Primary School Feeding Suggested guidelines 1993 Secondary School Feeding Guidelines 2005 Youth Development Fund Funding guidelines for youth 2012 development projects Source: MLG Notes: a/ Draft Revised Policy 2012 177. The instruments used in some programs do not appear to be adequate. For example, in the Destitute Persons Program, the same instrument—a food coupon—is used to provide both short-term emergency support for individuals and longer-term support for those who cannot work and have no means of subsistence. People who find themselves in a difficult situation or emergency might need some other type of support. Orphans receive food baskets or food coupons, whereas in some cases their primary needs (and of their families) may not be food but some other services. Indeed, the Needy Children Program targets, among others, orphans over 18 years old who graduated from the program but continue to need help because, presumably, they did not receive the support that they needed during their participation in the program. There is not one assistance program in Botswana that focuses on the family as a whole. Consequently, duplication of benefits is common with a family being able to benefit from the OAP, the Veterans Program, the Destitute Persons Program, and the Needy Student or Children Programs. While the lack of focus on the family as a unit makes it easier for households to receive benefits from several programs, it makes it difficult to develop the instruments and protocols to provide the services that meet the families’ needs. 90 178. The mechanisms used to deliver benefits in some programs are not the most efficient. For example, the delivery of food in-kind or even in the form of food coupons is not the most efficient because of the logistical costs involved. Also, it does not give beneficiaries any flexibility to buy what they need or even to save or invest. Moreover, these mechanisms are not consistent with the GoB’s policy of promoting self -reliance and avoiding dependency. 179. The operating cost of the programs gives an indication of whether they are cost- effective or not. Table 21 presents the operating costs of different types of social assistance programs from a variety of countries as reported in a World Bank study.179 As can be seen, the range of operating costs is quite wide, which makes comparisons difficult. Moreover, a program can have relatively high operating costs because of its small scale and its low benefits or because it is new and has high start-up costs. On the other hand, a program may have low operating costs because it does not invest in key activities and processes such as staff training, communications, auditing and controls, information management, and monitoring and evaluation systems. As a result, the World Bank study suggests developing an index of operating expenditures to take some of these factors into consideration. The proposed index is calculated by multiplying the generosity of program by the percentage of operating costs in the total program cost.180 179 Grosh et al (2008). 180 Grosh et al (2008), Box 9.3, page 392. The formula for the proposed index is I = (B/C*Q1)*CA/TC; where B is the transfer, C*Q1 the adjusted consumption for Q1 families, and CA and CT are the program’s operating costs and total costs respectively. Thus, a program with high generosity is expected to have relatively lower operating costs. 91 Table 21: Typical operating costs as a percentage of total program costs around the world No. of Type of Program Minimum Maximum Median Average Programs Food-related 6.0 55.5 22.0 25.4 19 programs a/ Cash and Near Cash b/ 2.0 16.6 8.9 8.2 16 Conditional Cash 4.0 13.0 6.7 8.2 9 Transfer Public Works 1.6 24.0 4.9 10.3 6 Source: Grosh et al (2008), page 411 Notes: A sample of international social protection programs. a/ Includes school lunches and nutrition programs. b/ Includes old age pensions, child allowance, food stamps, guaranteed minimum income. 180. In Table 21, it can be observed that food-related programs have by far the highest operating costs. In a sample of 19 programs, food-related programs average 25 percent of total program costs. We were not able to obtain the operating costs of Botswana’s feeding programs. Nonetheless, information from the MLG’s Directorate of Local Government Finance and Procurement Services (DLGFPS) indicates that the cost per student per year of the Primary School Feeding Program in 2012/13 was P829 or US$109 (Table 22). A World Bank study reports that the cost of school feeding programs usually range from US$28 to US$63 per child per year. Among the 19 sample countries, the average cost of these programs was even lower at US$20.40 per child per year.181 Some of the cost difference in Botswana may be attributed to a better menu and to the fact that in remote and drought- affected areas children receive two meals a day. Notwithstanding these, the cost per student in Botswana seems excessively high when compared to similar programs in other countries and this may require further analysis to identify cost saving options. 181 Bundy et al (2009), page 60. 92 Table 22: Cost of the primary school feeding program Total Annual Cost Annual Cost Per Year No of Students (Pula) Student (Pula) 2009/10 219,348,310 308,225 712 2010/11 242,176,885 310,757 779 2011/12 289,905,995 331,000 876 2012/13 274,806,773 331,000 829 Source: BIDPA (2012), Table 6, page 38, based on DLGFPS information. For 2012/13, information comes directly from the DLGFPS 181. The distribution of food is generally less cost-effective than the payment of benefits in cash. While the MLG now provides a food coupon (in the form of a smartcard) in some programs (the Orphan Care Program, the Destitute Persons Program, and the CHBC), which constitutes an important advance, it still limits what can be bought with the food coupons. It continues also to distribute food through the RADP and the Primary School Feeding Program. Paying benefits in cash has several advantages. First, it eliminates the need for logistic systems to procure, distribute, and preserve foodstuffs, which not only are costly and usually involve large transportation and storage food losses but also take a heavy toll in terms of the work time of social workers and other staff. Second, it gives families a choice of what to buy, when, and where, thus empowering families and making them more self-reliant, responsible, and less stigmatized. Third, the infusion of cash into the local economy stimulates activity and can help to create local microenterprises. This being said, cash may not be effective in all circumstances. For instance, cash is not appropriate when specialized foods are required, for example, for pregnant and lactating women, under- nourished children, the sick, or others in need of supplementary or therapeutic feeding. Also, cash may not appropriate in areas that are extremely food-insecure or when natural disaster strikes and crops are damaged and/or market channels are disrupted and emergency assistance is required. Finally, the value of the transfers in cash must be maintained to prevent the purchasing power of the transfers being eroded by inflation, in which case an in-kind transfer would be preferable. Box 1 presents some views of senior officials in Botswana on this issue as reported by the MLG. 93 Box 1: Views on cash versus food The majority of these senior informants support the idea of adopting cash transfers as an alternative to food baskets for the following reasons: • Beneficiaries already receive the cash component as part of the Destitute Persons Program. Experience in the districts shows that the majority of beneficiaries in this program use the cash component that they receive wisely and some have even used it to invest in a variety of ways. • Elderly people do not abuse alcohol. Most beneficiaries are the elderly. By giving them rations, the council literally takes away responsibility and accountability from them. Perhaps this explains why dependency is widespread. • In the few cases where cash is abused, social workers will make appropriate interventions. • Cash transfers will alleviate a lot of problems including the wastage of food rations that is common in families where a grandparent takes care of several orphans. • Social workers spend 90 percent of their time administering food rations. This arrangement will give them the opportunity to concentrate on providing psychosocial support to clients as well as on their other professional obligations Source: Ministry of Local Government (2010a), page 47 182. Replacing in-kind transfers with cash should help to eliminate many of the delayed and incomplete deliveries of benefits reported in a recent BIDPA study.182 BIDPA identified the late delivery of school uniforms and private clothing to orphans, noting that the children had outgrown the clothing by the time it was finally provided. Food packages were reported to be delayed and were often incomplete or contained reduced amounts, although the roll- out of the smartcard may have helped to solve this problem. BIDPA noted cases both of the inadequate provision of food to orphans and of waste when a family was given more food than they needed. 183 183. Table 23 presents the estimated operating costs of two social protection programs—the Internship Program and LIMID. We do not have any good benchmark for either of these two programs. The operating costs of the Internship Program have declined as the size of the program almost doubled during last four years. To be successful, the 182 BIDPA (2010). 183 BIDPA (2010), p. 35. 94 program would need to develop a realistic work plan for each intern and to invest in supervising the implementation of the work plan. More active participation is required from the hosting institutions. The operating costs of the LIMID program have fluctuated significantly in the last four years and generally appear to be on the high side, though relative high operating costs could be justified if they are the result of providing intensive support to the small farmers. Table 23: Operating costs of selected social protection programs 2009/10 2010/11 2011/12 2012/13 Internship Total cost (P) 44,083,127 57,337,221 70,349,138 72,758,328 Program Operating costs (P) 7,652,358 7,525,836 8,100,619 9,051,548 Operating/total cost 17.4 13.1 11.5 12.4 (percent) LIMID Total cost (P) 40,085,330 17,943,131 22,791,973 46,703,188 Operating costs (P) 20,085,329 6,025,131 2,791,974 16,703,188 Operating/total cost 50.1 33.6 12.2 35.8 (percent) Source: MLG and MOA. 184. Unfortunately we do not have any recent estimates of the operating costs for the major assistance programs, namely the Destitute Persons and Orphan Care Programs. Estimates for the 2004/05 to 2008/09 period indicate that the operating costs of these programs averaged about 5 percent of total costs, but these estimates should be treated with caution since we are not sure how they were arrived at.184 Moreover, these programs have changed from the distribution of food baskets to food coupons (smartcards), which is likely to have affected their operating costs. Nonetheless, for illustrative purposes, we can follow the procedure suggested by the World Bank study185 and calculate an index of operating costs for the Destitute Persons Program. Recall that the program’s generosity was estimated at 77 percent. Assuming that operating costs as a percentage of total costs are 5 percent, it yields an index of operating costs of 3.8 (=77*0.05), which compares with indices 184 Ministry of Local Government (2010a), page 112 (Orphan Care Program) and page 133 (Destitute Persons Program). 185 Grosh et al (2008). 95 ranging from 0.5 to 2.25 with a median value of 1.1 for 10 social assistance programs in various parts of the world for which information was reported in the World Bank study.186 Consequently, the operating costs of the Destitute Persons Program are quite high by these standards. 185. For the Ipelegeng program, the MLG reports average operating costs of 12 percent in the last five years although with a tendency to decline as the program has nearly doubled in size (Table 24). In the last two years, the Ipelegeng had operating costs below 10 percent of total costs. This compares to a median of 4.9 percent and an average of 10.3 percent in Table 21. Indeed, worldwide experience suggests that this type of program should have operating costs that are no more than 10 percent of the total costs of the program (see Box 2).187 Table 24: Ipelegeng expenditures (P millions) Operating as Operating Total percent of Year Labor Meal Material Costs a/ Expenses Total Expenses 2008/09 102.3 105.8 21.2 229.2 9.2 2009/10 118.5 57.1 43.2 218.7 19.8 2010/11 205.0 37.7 34.0 276.8 12.3 2011/12 255.0 21.4 24.4 300.8 8.1 2012/13 264.9 32.5 72.6 38.6 408.6 9.4 Source: Ipelegeng program management and MLG Note: a/ Operating expenses include personnel emoluments, motor vehicle fuel and maintenance, and small tools. 186 Grosh el al (2008), Box 9.3, page 392. 187 Subbarao et al (2010), page 25. 96 Box 2: International experience with public works In India, the National Rural Employment Guarantee Program provides a guarantee of 100 days of work per household on demand to all rural families, at a minimum local agricultural wage rate and is completely self-targeted on the basis of that low wage rate. In Ethiopia, the Productive Safety Nets Program (PSNP) employs some 1.2 million people annually for an average of 150 days each at a cost of about US$280 million per annum. It uses a combination of self- targeting through the wage rate along with geographical targeting (poor areas) and community selection of beneficiaries (because there is excess demand for employment). The Ethiopia program has been successful in increasing the long-term productivity of some areas by improving soil and water management as well as through the more usual road construction projects. While there are challenges (and costs) involved in mounting this scale of program, even countries with less ambitious programs (such as Malawi and Zimbabwe) have managed to achieve coverage of about 300,000 to 400,000 people each year. In Argentina, as much as 60 percent of expenditures on the Trabajar program are spent on wages paid to the poor, as do programs in Zambia, Madagascar, and India, while only between 38 percent and 47 percent of spending in Tanzania currently goes in wages. Focusing on road maintenance has a number of other benefits ‒ it does not require the formulation of separate projects to employ the poor, it can be administered through existing systems, and it can make use of money that is already being spent in the budget. Sources: India (Dutta et al, 2012), Ethiopia (World Bank, 2009), and Argentina (Jalan and Ravallion, 2003) 186. The pension plan for public employees (the BPOPF) has higher operating expenses than the private plans. Table 25 indicates that the operating expenses of the BPOPF are 16 percent of contribution income compared with 14 percent for the private funds. Clearly, the BPOPF, which is twice the size of all of the private funds together, is not reaping economies of scale. These rates appear to be high. A 2010 report on the retirement systems in countries in the Southern Africa Development Community (SADC) indicated that “best practice in the region demonstrates that costs of 4 percent” can be achieved in civil service plans. It found that administrative costs were 4 percent of total income in Tanzania and 7 percent of contributions in Namibia but 17.5 percent of contributions in Zambia.188 188 Oxford Policy Management (2010), page 52. Note that all of these plans are defined benefit plans. 97 Table 25: Operating expenses of pension funds Plans 2011/12 BPOPF Operating Expenses (P million) 253.7 Contribution Income (P million) 1,602.2 Adm. Expenses/Contribution Income 16 (percent) Private Plans Operating Expenses (P million) 90.3 Contribution Income (P million) 648.5 Adm. Expenses/Contribution Income 14 (percent) Source: Non-Bank Financial Institutions Regulatory Authority and BPOPF. 187. The resources allocated to the OAP might not be as cost-effective as possible in terms of poverty reduction. Because the pension is universal and the GoB has a limited budget for the program, the amount received by the non-poor could be redirected to increase the benefits of other programs targeted to the poor. Indeed, a recent study has argued that universal pensions are less cost-effective in terms of poverty reduction than targeted programs.189 188. To sum up, most programs do not have comprehensive policies or operational guidelines, and in other cases the guidelines are out of date. The instruments used in some programs do not appear to be adequate, and the mechanisms to delivery benefits in some programs are not the most efficient. Finally, the review of the operating costs of the few program for which information is available indicate that in general they appear to be on the high side, with an exception of the Ipelegeng program, which appears to be within international benchmarks. 189 Grosh and Leite (2009), page 161. 98 F. Incentive Compatibly and Sustainability 189. One area of concern for policymakers in Botswana is related to the potential work disincentives and dependency effects of the income support programs. Some government policies may be providing the wrong incentives to work, save, and be self-sufficient and may in fact be promoting dependency. This seems to be the case not because of a single program but a result of the cumulative approach to social policy which provides food assistance “from cradle to grave,” non-repayable “loans,” generous sponsorships and scholarships, generous compensation in youth programs, and generous transfers to adults and the elderly. On the social assistance side, four programs stand out as being poorly targeted to the needy or too generous: the School Feeding Programs, the Destitute Persons Program, the Old Age Pension, and, increasingly, the Ipelegeng. 99 Box 3: Concerns over work disincentives induced by targeted social assistance programs Poverty-targeted cash social assistance programs may sometimes negatively influence the labor supply of adult household members, who may be less inclined to join the work force as a result of receiving transfers. Dependency on benefits and reduced participation in the labor market may lead in turn to the exclusion of prime-age beneficiaries from the mainstream society. Second, policymakers in some countries are concerned about the effect of programs on the choice between formal and informal sector jobs. Given the size of the informal sector in some countries, income support programs may run the risk of unintentionally increasing the attraction of the informal sector. Third, policymakers have political economy concerns related to the legitimacy and political backup for income support programs. The perception by the general public that income support programs make recipients dependent on state benefits or encourage informality may limit the support for financing such programs. This in turn affects the programs’ capacity to provide effective protection against temporary income shocks or chronic poverty. Intensity of Job Search Labor Supply of Other Choice between Formal Political Economy Effort and Acceptance of Household Members and Informal Sector Jobs Job Offers Income support programs Income support Income support programs Perception by the public may negatively affect the programs may negatively may run the risk of of income support intensity of the job influence the labor unintentionally increasing programs as creating search effort and the supply of other the attraction of the dependency on state reservation wage, leading household members. informal sector. benefits or encouraging to longer unemployment informality. spells. From a theoretical perspective, a social assistance program could reduce the supply of labor of beneficiaries if the following conditions were to apply: • The beneficiary household includes members of working age who do not work. A more formal definition of this group is individuals of working age who are not disabled or in employment, education, training, or early retirement (the NEETD group). The relative size of the NEETD group will suggest whether this phenomenon is rare or widespread for a given social assistance program. • The benefit provided by the program is large enough to allow the beneficiary (and his family or household) to “live off the benefit.” A proxy for this factor is the generosity indicator: the ratio of social assistance transfers to the consumption or income of the beneficiary households (ranges from 0 to 1). Higher generosity implies stronger disincentives to work. To identify the share of households that depend on benefits for their household income, one could quantify the share of social assistance beneficiaries where the share of social assistance income(s) represents more than 50 percent of the total household income. A generosity indicator at or above 0.5 indicates that one spouse earns benefits equal to or greater than the rest of the earnings of the household. A generosity indicator at or above 0.33 indicates that one spouse gets benefits equal to or greater than half of the earnings of the household. • The social assistance transfers are reduced when the beneficiary’s earnings go up. This happens when the benefit formula includes an implicit marginal tax rate (MTR) on earnings. The MTR indicates the value of benefits lost when earnings go up by one monetary unit. The MTR will typically range from 0 to 100 percent, although certain eligibility conditions may result in values that are larger than 1 (for example, when moving from assistance to work results in the loss of multiple benefits, larger than the amount earned). Positive MTRs are found in all income- or means-tested programs. Higher values of MTR indicate fewer incentives to work. For example, a simple Guaranteed Minimum Income program that reduced the value of benefits at par with any extra earnings would have a 100 percent MTR and, therefore, would discourage work. Source: Tesliuc et al (2013). 100 190. From a theoretical perspective, a social assistance program could reduce the supply of labor of beneficiaries if the following conditions were to apply: (i) The beneficiary household includes members of working age who do not work; (ii) The benefit provided by the program is large enough to allow the beneficiary (and his family or household) to “live off the benefit.”; and (iii) the benefits are reduced when the beneficiary’s earnings go up (see Box 3). The evidence that some programs in Botswana may create dependency is rather anecdotal and more research to analyze the work and savings behavior of recipients is needed. 191. When asked to evaluate the impact of the School Feeding Programs, some beneficiaries are reported to have said: There is a lot of parental neglect. Some parents believe that the school has to take care of all the needs of the children including providing food for them.190 192. On the Destitute Persons Program, some senior civil servants indicated: By giving them rations, the Council literally takes away responsibility and accountability from them. Perhaps this explains why dependency is widespread. 191 193. The evaluation of the Destitute Persons Program in 2008 produced some important recommendations including the need for: “a paradigm shift or a serious re-think in addressing poverty and destitution in Botswana. It urged a renewed emphasis on self-reliance and a stricter approach to the registration of the destitute. It recommended a transfer from food packages to cash.”192 190 Ministry of Local Government (2010a), page 120. 191 Ministry of Local Government (2010a), page 47. 192 Ministry of Local Government (2010a), page 125. 101 194. The universal OAP may also reduce the incentives to save for old age. While one can argue whether the transfer is too generous or not, in Botswana, some elderly people also receive benefits from the Destitute Persons Program and possibly from other programs. The expectation of a food basket or coupon together with some cash might be a disincentive, at least for some Batswana, to save for their old age. On the other hand, calls to eliminate from the roster of other assistance programs those who receive a pension would penalize those who have saved for retirement. A middle course of gradually reducing the subsidy as the value of the pension increases, similar to that taken by Chile, might be a good alternative for Botswana as part of a strategy to expand the coverage of social insurance (see Chapter V). 195. With regard to the Ipelegeng, the recent increase in the compensation and the provision of a free meal makes the program more attractive than working in agriculture. Ipelegeng compensation is now 30 percent above the minimum salary in agriculture or 90 percent if the number of working hours is taken into consideration. This is not to say that the Ipelegeng is not a useful program. On the contrary, the Ipelegeng program has a series of interesting characteristics. It is a major safety net for the poor unemployed, and it has a high rate of participation by women (over 70 percent), even though in general women have lower labor force participation than men. Ipelegeng officials also indicate that most women who work in the program come from large families and are poor. However, the program also faces some important challenges. For example, there are reports that the projects being financed are of low quality. Also, the program has no links with other capacity- building programs and offers no graduation possibilities to the beneficiaries (more on this on Chapter V). The MLG, with the support of UNICEF, has conducted an evaluation of the program that is expected to be published shortly and to shed more light on these issues. 196. Other countries have similar concerns as Botswana but the available evidence shows that social safety net programs do not seem to create labor disincentives. A recent evaluation of the Family Benefit Program in Armenia carried out using both descriptive and 102 rigorous empirical evaluation techniques (regression discontinuity) found that the program did not seem to contain any labor disincentives. In general, the FBP did not appear to create any disincentives for participating in the labor force, for work effort, or for moving into the formal sector (Box 4). Box 4: Armenia – does the family benefit program create a disincentive to work? In 2010 the Government of Armenia (GOA) expressed concerns over the potential work disincentives effects of the Family Benefit Program (FBP). The profile of FBP beneficiaries revealed that the propensity for inactivity is similar for FBP and non-FBP able-bodied working-age individuals. The analysis of benefit generosity suggested that the amount of the FBP transfer does not compensate for potential earnings in the formal labor market, nor is it adequate for families to live off the FBP transfer alone. The estimates of program effects obtained by rigorous empirical analyses supported the descriptive results and showed no significant negative impact of the FBP on inactivity and informality. The only negative effect of the FBP that the evaluation identified was a reduction in hours worked in the rural sample. - The evaluation also yielded valuable insights into the employability profile of beneficiaries: - Fewer than half of all members of FBP families are of working age. - FBP families have fewer able-bodied working-age individuals and a higher dependency ratio than the average Armenian household. - FBP beneficiaries who could be expected to work appear to have less education than similarly defined non-beneficiaries. Given their lower educational attainment, members of FBP families would be expected to be, on average, less employable than their non-beneficiary counterparts. - Members of FBP families appear more likely to be working part-time and in the informal sector. However, FBP beneficiaries’ higher probability of working in the informal sector is not likely to be a matter of choice, since formal workers earn much higher wages than informal workers. Source: World Bank (2011) 197. With the expected reduction in mineral revenues, Botswana is going to be fiscally constrained and will need to ensure that social protection programs provide the best “bang for the buck.” Indeed, NDP 10 projects that government revenues from diamonds will decline sharply from P14 billion in 2021 to P2 billion in 2022 (at constant prices) and will gradually approach near zero by 2029.193 This means that policymakers will need to review some of the existing policies and provisions and develop tools that can increase the cost- 193 Ministry of Finance and Development Planning (2010), Chart 5.1, page 55. 103 efficiency of social assistance programs. Social protection programs should not create disincentives to work and save and should be targeted only to those who really need the support of the system. On the other hand, the GoB should pursue social security policies that facilitate membership in social insurance as the more people who have social insurance, the less social assistance the GoB will be required to dispense. 198. While there are concerns that some programs may be providing the wrong incentives to work and may in fact be promoting dependency, this area would require additional work to measure work and savings behavior. The recent CWIS 2009/2010 data could be used to analyze the propensity to work of households receiving social assistance relative to similar households that do not receive. G. Monitoring and Evaluation 199. The NDP 10 established a framework for the monitoring and evaluation (M&E) of policies, programs, and projects and clear lines of responsibility.194 A comprehensive Macro Development Results Framework (MDRF) has been developed for M&E purposes at the national level. The Results Framework is complemented by other tools such as logical frameworks, formal surveys, public expenditure tracking surveys, and impact evaluations, which will be used to monitor and evaluate the performance of projects, programs and policies where appropriate. 200. The national MDRF links broad national goals (contained in Vision 2016) with the key national strategies of NDP 10 and is used to measure and report on the progress that the country is making in meeting these national goals. At the national level, the Vision 2016 Council and the MFDP are responsible for M&E. The Vision 2016 Council monitors and evaluates progress at the macro level, while the MFDP covers the financial management 194 Ministry of Finance and Development Planning (2010), Chapter 16, pages 354s. 104 aspects. At the sector/ministry level, M&E is primarily the responsibility of the Government Implementation Coordination Office (GICO). The GICO was created in 2007 to improve the coordination and monitoring and evaluation of policy and programs and supports the development and maintenance of M&E capability across government institutions. The GICO’s responsibilities are as follows: • Developing and managing the application of a structured M&E system • Providing guidance and support to ministries/departments in the effective use of evaluation as a management tool. • Promoting the use of evaluation findings by ministries/departments to improve the quality of policies, programs or projects. • Undertaking and coordinating thematic/sector evaluations that cover crosscutting themes or issues. • Advising on the alignment of projects with policies and strategies. 201. Despite the establishment of this M&E system, major challenges remain. The recent report from the Vision 2016 Council on “Botswana’s Performance: Report on the Progress being Achieved against the Vision 2016 Goals” concludes:195 “Little is known about how effectively programs are performing in Botswana… While the government identified the importance of M&E of government programs at the outset of the NDP 10, little progress has been made across the system in terms of M&E.” 202. The information on social assistance programs is limited, and few programs have been evaluated. For example, the government has no way of knowing at the household level what benefits are going to whom, nor does it have any information on the 195 Vision 2016 (2013a), page 15. 105 socioeconomic characteristics of recipients, most crucially their poverty status. The monitoring and evaluation of programs is hampered by the lack of comprehensive information on who receives what benefits. On program expenditures, there is no central source of information. As the budget is not prepared in a programmatic manner, it is very difficult to keep track of how much has been allocated and/or disbursed by the individual programs.196 Very few programs keep track of their overhead costs. 203. The Alternative Packages Program, which is part of the Poverty Eradication Initiative, was recently launched without the monitoring and evaluation mechanism in place. This program donates assets and working capital to individuals to establish microenterprises, but its impact on the beneficiaries is still not being monitored. Several officials have questioned the sustainability of this program. A recent report published by Vision 2016 Council presented some of the issues being raised by the beneficiaries of the Backyard Gardens (one of the packages of the Alternative Packages Program):197 • Water connection fees are unaffordable for some. • Water connections are not always available. • Some households do not have space for gardens. • Pest and diseases can be a problem. • Some beneficiaries have changed their minds or are not committed to the project. 204. The report also presented some successful initiatives that explore the possibility of using local resources in a sustainable manner. The report recommended: “have stakeholders map their needs for affirmative action programs.” It suggested that better 196 The budget is divided in two separate parts — the recurrent budget and the development (or project- related) budget. 197 Vision 2016 (2013b), page 27. 106 coordination is needed among government entities and that “plans should be monitored by a joint structure at all levels.”198 205. Few of Botswana’s programs have been evaluated as to whether they are succeeding in achieving their goals. Of the 10 social major social assistance programs, only five have been evaluated in the last five years and four have never been evaluated (Table 26). With no rigorous evaluations of Botswana’s safety net programs, the government cannot know what is and is not working. For example, the contributions of the VGFP or the School Feeding Programs to nutrition have not been studied, and despite the large size of the feeding programs, under-nutrition in Botswana remains high. In-depth impact evaluations are also needed of the programs’ results chain, so that policymakers know not only what happened but why as well. The results of such evaluations can lead policymakers to change important aspects of the programs’ design to increase their effectiveness. In addition to impact evaluations, regular process and implementation evaluations are equally important for gathering information on whether beneficiaries and program staff clearly understand program rules and objectives and if procedures are being applied correctly. Table 26: Evaluations of social assistance programs Program Date of last evaluation Community Home-Based Care Program - Orphan Care Program 2006 Primary School Feeding Program 2012 Secondary School Feeding Program not dated Vulnerable Group Feeding Program - Destitute Persons Program 2008 State Old Age Pension - Veterans Pension - Remote Area Development Program 2010 Ipelegeng 2013 a/ Source: Ministry of Local Government (2010a), Table 10 Note: a/ Evaluated with the support of UNICEF. Report to be published shortly 198 Vision 2016 (2013b), page 40. 107 206. Survey results take long time to be published. The full results of the 2009/10 BCWIS are still not available, nor are those of the Population and Housing Census of 2011. Survey questionnaires also need to be improved. The questionnaire of the BCWIS was not designed to collect information (such as the numbers of people covered, the amounts that they receive, or the reasons why applicants are rejected) on specific social protection programs. The income portion of the 2002/03 HIES questionnaire does specify six social assistance programs (Drought Relief, the Destitute Persons Program, the Orphan Care Program, the Old Age Pension, the Veterans Program, and the RADP), but in the 2009/10 BCWIS several of the programs seem to have been lumped together under the heading “government aid”, which makes it impossible to evaluate who is benefitting from each program. If the BCWIS is to be used to evaluate these programs, the questionnaire will have to be improved. To counter the problem of small samples in the smaller programs, from time to time, the survey could organize an oversampling of people likely to participate in those programs. 207. The MLG’s Department of Social Services has taken steps to improve record keeping and reporting for the CHBC, the Destitute Persons Program, and the Orphan Care Program. However, information from the field continues to be mostly processed manually, and timeliness and coverage are still inadequate. The Elderly and Benefits Division of the DSP operates the most sophisticated information system, the so-called Social Benefits Payment and Reconciliation System (SOBERS). It records payments made by the OAP, the Veterans Program, and the Destitute Persons Program. The system can generate statements by pay point, payment mechanism, and gender of recipient for any reporting period. The Ipelegeng program also operates an information system that can generate summary statements by district, town, or constituency. The MLG’s Food Relief Services Division operates a logistics information management system and keeps detailed records of commodity procurement, stock levels, and deliveries. However, the MLG is not able to produce separate operation cost information on the Vulnerable Group Feeding Program and the Primary School Feeding Program. 108 208. To sum up, Botswana made important progress in establishing an overall M&E system for public policies and programs, and some progress has been made in developing information systems for specific social assistance programs within the MLG. However, much still needs to be done to monitor and evaluate major programs on a timely basis and to develop comprehensive management information systems for major social protection programs. 109 110 H. Institutional Arrangements 209. Several institutions play an important role in the development and implementation of social protection policies and programs (see Diagram 1). The MLG is responsible for the major social assistance programs. Through its Department of Social Protection (DSP), it manages the Orphan Care Program, the CHBP, Destitute Persons Program, the Veterans Program, the OAP, the RADP, and the Alternative Packages Program. Also, through the Department of Local Government Finance and Procurement (which includes Food Relief and Procurement Services), the MLG manages all of the major feeding programs, and through its Department of Local Government Development Planning, it manages the Ipelegeng program. The Ministry of Labor and Home Affairs (MLHA) is responsible for the major active labor market policies and programs. The Ministry of Youth, Sports, and Culture (MYSC) manages major programs for young people. The MESD manages the Secondary School Feeding Program and the Ministry of Health (MOH) with the support of the MLG manages the distribution of food baskets in the clinics and in the CHBP. The Ministry of Agriculture (MOA) manages LIMID, which is part of the Alternative Packages Program. On pensions, the Non-Bank Financial Institutions Regulatory Authority (NBFIRA) supervises public and private pension plans and insurance. The MFDP through its Economic and Financial Policy Division coordinates the financing of social protection programs, and through its Insurance Pension Section, it oversees the financing of the public pensions programs. 210. The MLG is now being reorganized with the creation of a new Department for Community Development, which is in charge of the RADP, the Alternative Packages Program (for MLG packages only) and community development affirmative action initiatives and programs. The Department of Social Services is to be renamed the Department of Social Protection. It is still unclear where the Ipelegeng is to be placed. 111 211. Coordinating all these actors and programs is a challenge. In the past, the MFDP had a prominent role in the coordination of poverty and rural development issues.199 At present, the Office of the President has taken a pro-active role in overseeing the policies and programs of line ministries and has established the Government Implementation Coordination Office (CICO) to increase the efficiency of policy and program design and implementation. The Poverty Eradication Group in the Office of the President plays an increasingly prominent role in developing social protection programs and in ensuring that they are consistent with other broad poverty reduction strategies. However, more needs to be done to clearly define the roles of different actors and to demand greater accountability for results. 212. Program implementation on the ground is the responsibility of the local councils in the districts, cities, and towns. Their Social and Community Development (S&CD) staff are responsible for evaluating, counselling, and supporting the beneficiaries of the Orphan Care Program, the Destitute Persons Program, and the CHBC. The same S&CD staff team manages these three programs. The councils also help to implement the Primary School Feeding Program. The Ipelegeng has its own field staff as does the Food Relief Services Division of the MLG to manage its food distribution system nationwide. While in the Ipelegeng the lines of accountability of the field staff and the MLG’s Department of Local Government Development Planning Division are clearly established, the lines of accountability are less clear in the other programs. Local council staff report to the councils’ Chief Executive Officers who in turn report to the Permanent Secretary of the MLG. Therefore, the staff on the ground implementing the programs do not report directly to the MLG department in charge of the program. This is the case, for example, of the Destitute Persons Program, the CHBC, the Needy Children Program, and the Needy Students Program. Social workers assigned to these programs are accountable to the local authorities and not 199 Ministry of Local Government (2010a), page 50. 112 to the MLG’s DSP, which manages the programs. This creates problems of accountability and transparency. 213. The MESD, the MLHA, the MYSC, and autonomous bodies such as the Botswana Training Authority (BOTA), the Tertiary Education Council (TEC) are all involved in active labor market programs. In 2009, the GoB approved a new Human Resource Strategy 2009- 2022, which lays out a short-, medium- and long-term human resource development strategy for the country.200 One of the first steps is to eliminate the existing fragmentation and place the responsibility for key institutions and tools under one roof, the newly established Human Resource Development Council (HRDC). The HRDC together with the Botswana Qualification Authority and the Botswana Examination Council under the MESD will be responsible for coordinating the implementation of the strategy. The HRDC will become responsible for the following programs: the Labor Market Observatory and Labor Exchange, which are currently being established; the Scholarships/Sponsorship program; the Internship Program, and the management of the 0.2 percent turnover tax to finance vocational training. Some of vocation training centers under the MESD such as the Brigades (vocational training for unskilled workers) will become autonomous. Bills have been drafted to create the Human Resource Development Council and the Botswana Qualification Authority and are expected to be approved by Parliament shortly. 214. In sum, Botswana lacks a high-level body that can dictate policy, establish goals, and coordinate the development of social protection and labor programs. Many institutions are involved in the sector, and most programs operate in silos, circumscribed by their own agendas, needs, and priorities, with little communication between them and little attempt to explore synergies. At the local level, the major institutional weakness derives from the fact that there are many departments/divisions implementing many programs 200 At the request of the Government of Botswana, the World Bank prepared four policy notes on “Skills for Economic Growth and Diversification in Botswana,” analytical work that helped to inform the design o f the strategy. 113 targeting the same groups and often using inefficient instruments. Many staff, particularly social workers, spend most of their time administering the delivery of food or food coupons or discharging other functions (such as the rehabilitation of beneficiaries) instead of doing the work for which they were trained. I. Summary Assessment 215. Botswana’s social protection system covers most of the risks that have been identified, but there are substantial program overlaps that are a consequence of the fragmented approach to social protection. At the same time, the preceding analysis points to three major programmatic gaps: a program that focuses on poor families with children; more focused labor-intensive public works; and old age pension and unemployment insurance for all workers. 216. Botswana is spending substantial resources on its safety net. However, the results are not commensurable with the amount of spending, and there is substantial room to increase the cost-effectiveness of social protection programs. While the targeting of the major program that is means-tested (the Destitute Persons Program) appears to be relatively accurate, a large number of poor people including children are excluded from the social assistance programs. Cost-effectiveness could be increased by replacing food or food stamps with cash transfers, by improving targeting to the poor, by rationalizing existing interventions, and eliminating existing disincentives to work, save, and invest in income- generating activities. 217. The assessment indicates that there is substantial room to establish links between social assistance, active labor market, and social insurance programs to exploit the synergies between programs, to improve the incentive compatibility, and make the system more sustainable while delivering more resilience, equity, and opportunity to all Batswana. 114 218. Finally, there is almost no rigorous evidence on what impact existing programs are having on the consumption and poverty status of those whom they are intended to benefit, making it difficult for policymakers to determine which represent the most cost- effective options for scaling up. 115 V. OPTIONS TO STRENGTHEN THE SPL SYSTEM 219. After briefly reviewing Botswana’s vision and priorities for social protection, this chapter presents a strategy to put the country’s vision into operation based on the above analysis. The rest of the chapter offers more detailed recommendations on how to implement the suggested strategy. A. Vision and Strategy 1. The Government’s Vision 220. Vision 2016 outlines clearly the vision that Batswana want for their country underpinned by the following five national principles derived from Botswana’s cultural heritage: democracy, development, self-reliance, unity, and botho.201 221. The vision consists of several goals or “pillars”—an educated, informed nation; a prosperous, productive, and innovative nation; a compassionate and caring nation; a safe and secure nation; an open, democratic, and accountable nation; a moral and tolerant nation; and a united and proud nation. A key policy priority is poverty eradication. To achieve “A Compassionate and Caring Nation,” Pillar III of Vision 2016, “Botswana must establish an effective social safety net to support those who are genuinely unable to obtain a minimum standard of living through their own efforts, and to assist and encourage them to become self-reliant again. Labor-intensive work programs have considerable potential as an employer of last resort, and as instruments of poverty alleviation. They should not be related to drought, as in the past, but to the existing pattern of poverty in the country. ”202 201 The concept of a person who has a well-rounded character, who is well-mannered, courteous and disciplined, and realizes his or her full potential both as an individual and as a part of the community to which he or she belongs (Republic of Botswana, 2011a, page 5). 202 Republic of Botswana (2011a), page 41. 116 222. Vision 2016 places the family in the center of development efforts: “The family will be the central institution for the support and development of people in Botswana, and for the transmission of social and moral values. The strength of the family will have been reinforced in response to the rapid social changes that are sweeping the country, the region, and the world. The emphasis on a strong family unit will encourage responsible parenting and the institution of marriage. It will provide the social foundation for the eradication of problems such as the high incidence of teenage pregnancies, adultery, prostitution, street children, and the spread of HIV.” 203 2. Building a Coherent Social Protection and Labor System 223. This assessment uses the analytical framework of the recently adopted World Bank Social Protection and Labor Strategy204 2012-2022, which recommends that countries move toward harmonized SPL systems by eliminating duplications, gaps, or inefficiencies. These systems should be able to deliver the three core functions of the SPL systems: resilience, equity, and opportunity.205 Taking a systems approach to social protection and labor interventions should strengthen coordination and integration at three levels: policy, programs, and administration.206 224. At the policy level, the country should have a clear vision and strategy for the development of the SPL system that is consistent with national goals and priorities and seeks to address the population’s major risks and vulnerabilities. The strategy should ensure consistency between instruments, financing, and institutional arrangements and should clearly state: (i) the objectives and functions of the SPL system; (ii) the different instruments needed to achieve the stated objectives; (iii) given the risks and vulnerabilities, 203 Republic of Botswana (2011a), page 12. 204 World Bank (2012a), page v. 205 World Bank (2012a, page i. 206 Robalino et al (2012), pages 9 and 10. 117 what type of programs are needed and what they are expected to achieve; (iv) financing mechanisms that are sustainable and progressive and that minimize economic distortions; and (v) institutional arrangements that provide clear operating rules, well-defined roles and responsibilities for participating institutions, incentives for performance, and accountability for results. 225. At the program level, the World Bank strategy recommends that countries harmonize or integrate similar programs or schemes, and improve the coordination and/or strengthen the links between programs (and services) within SPL and other sectors. To this end, there is a need to analyze: (i) program coverage, gaps, and overlaps; (ii) program generosity and its impact on the incentive system and program sustainability; (iii) targeting performance (benefit incidence); (iv), program cost-effectiveness; (v) opportunities to promote the harmonization of programs (in terms of benefits, services, and eligibility conditions) or to integrate programs; and (vi) opportunities to establish links between programs within SPL and other sectors, particularly education and health. 226. At the administrative level, the focus should be developing or improving the tools that enable the SPL system’s basic business processes, including: (i) targeting and beneficiary selection mechanisms; (ii) the registry of beneficiaries; (iii) the delivery of benefits or payment systems; (iv) the grievance redress mechanisms; and (v) management information systems and monitoring and evaluation. 227. According to the World Bank, “strengthening SPL systems can provide a broad set of advantages. Stronger SPL systems can enhance equity by providing individuals with the means to secure their rights. They offer efficiency, through better designed tax/financing arrangements, economies of scale, and common platforms such as registries. They impart 118 effectiveness, through harmonized incentives and fewer unintended consequence on behavior.”207 3. Elements of a Strategy 228. Botswana has a mature and complex social protection system, with many elements of a modern social protection system (see Table 1). Botswana is one of the few countries in Africa that fully funds the social protection programs out of own resources, and dedicate a large part of its GDP to this endeavour. During 2012/13 fiscal year, social protection spending accounted for about 4.4 percent of GDP, or P5,347 million. Of this, social insurance spending accounted for 1.2 percent of GDP and consisted of contributory pensions mainly for the public sector employees. Reflecting the structure of the economy and its small formal sector, contributory pensions cover less than 13 percent of the workforce. To support employment, Botswana also finances a number of active labor market programs, whose total spending represented 0.17 percent of GDP in 2012/13. Social assistance spending represents another 1.7 percent of GDP, allocated to a diverse mix of programs that address most vulnerable groups in the country. The social assistance program mix includes a public works program (Ipelegeng); a social (non-contributory) old age pension; cash and in-kind assistance for families who take care of orphans and for destitute (indigent) persons; nutrition programs for infants and pregnant and lactating women; school feeding for primary and secondary school children. A fourth area of spending is on sponsorships and scholarships for students in tertiary education, accounting for 1.4 percent of GDP spent in 2012/13. These policies lie at the border between human capital development policies and the safety net. 229. At the same time, the social protection system exhibits a number of features that are suboptimal. The majority of the programs have limited effectiveness because of their 207 World Bank (2012a page 18. 119 small size. Contributory pensions cover less than 13 percent of the workforce, which reflects the structure of the economy and its small formal sector. Targeted programs for the poor such as the Destitute Persons or Ipelegeng cover less than 3 percent of the population.208 Furthermore, safety net programs are fragmented, implemented by different ministries, diluting scarce administrative capacity. The current social assistance spending is skewed in favor of a single program—scholarships and sponsorship for tertiary students—that absorbs 45 percent of all social assistance spending. The high volume of spending for these programs shows the commitment of the Government of Botswana to ensure that every child gets access to education, including tertiary. The intended goal of these programs is to break the cycle of inter-generational poverty, by ensuring that children attain the highest level of education, so that they can compete in the labor market. However, it is very likely that these programs mostly benefit the non-poor, as secondary school enrolment rate is only 61 percent; this is probably due that low income children do not finish secondary school. Also, the high and persistent level of malnutrition reveals the ineffectiveness of the nutrition program. The current system relies too heavily on in-kind distribution of food (e.g. school feeding programs, Vulnerable Groups Feeding Program), which are known for spending a large share of their budget on administration and logistics. In addition, the safety net covers only a small fraction of the absolutely poor because of its reliance on programs targeted at individuals, not families, and the lack of a last resort family-focused anti-poverty program. 230. These weaknesses could be corrected over the next few years and by 2016 Botswana could emerge with a modern, effective and efficient social protection system, capable of eliminating absolute poverty. Using the analytical framework of the recently adopted Social Protection Strategy of the World Bank, this report offers punctual suggestions on how to strengthen the social protection system in Botswana. Based on an in- 208 The low coverage of the Destitute program is due to the fact that it targets only the poor who are unable to engage in sustainable economic activities due to various reasons, and have insufficient assets and income sources. The work-able poor are expected to apply to Ipelegeng, the public works program. 120 depth assessment of the social protection programs, this report recommends three strategic directions and associated policy measures that would strengthen the social protection system over the next three to seven years. First, the report proposes a number of options to tighten the safety net and contribute to the elimination of absolute poverty in a budget neutral way. These policy measures are focused on filling existing gaps in coverage, eliminating program fragmentation and curtailing over-generous benefits. Second, the report spells out a comprehensive but feasible administrative reform agenda, targeted to develop the administrative tools and systems that underpin an efficient and effective service delivery across all social protection programs. Third, the report illustrates a set of recommendation focused on strengthening institutional arrangements and coordination mechanisms. B. Eradicating Absolute Poverty in a Budget Neutral Way 231. Recently, the Government set an ambitious goal to eradicate absolute poverty by 2016, relying on its safety net as one of the key pillars. In 2011, the Government has developed Vision 2016, an overarching strategy that identifies the goals for Bots wana’s development and provides a timeline for achieving them by 2016. One of the pillars of the Vision calls for “a more equitable income distribution and the eradication of absolute poverty in Botswana”. To achieve this goal, Vision 2016 and the accomp anying Social Development Policy Framework recognize the need for a social safety net for the poor and vulnerable that can withstand unexpected disasters or crises. As of 2009/10, 16 percent of the total population lives in absolute poverty. If the share of population living in absolute poverty has not changed by 2012/13, it would represent about 84,000 families or 336,000 persons. 232. The goal of lifting 84,000 families from absolute poverty by 2016 is achievable by redirecting 0.4-0.6 percent of GDP toward families living in absolute poverty and implementing a targeting system. Tactically, this will require a better weaving of the safety 121 net through the introduction of a last-resort, poverty-targeted program; a reduction in the wage rate of the public works program, Ipelegeng, to improve both the targeting and the coverage of the program; and accompanying cost-cutting measures to ensure budget neutrality. 233. The suggested road map rests of three pillars: • A first pillar will consist of a new family-based last resort anti-poverty program, tentatively called the Family Support Grant (FSG). This program will provide last resort cash assistance to all absolute poor families that are currently not adequately supported by other social protection instruments. • A second pillar includes the Ipelegeng public works program, which will continue to focus on poor adults without employment opportunities. Currently, the program achieves modest targeting results due to poor regional allocation of its funding and by offering a relatively high wage rate, which makes the program attractive for the non-poor. The report recommends a set of parametric reforms to improve the targeting and the effectiveness of this program. • A third pillar consists of programs with good design, coverage and budget that will continue to operate in the same manner and will cover a part of the absolute poor. In this pillar we include the Old Age Pension and the Community Home- based Care, which will continue to cover all elderly poor and the chronically ill that require special diets and care. 122 1. Introducing the Family Support Grant program 234. While Botswana has numerous safety net programs aimed at vulnerable individuals, it lacks a last-resort anti-poverty program that would protect all individuals against absolute poverty. The focus on programs serving vulnerable individuals is likely to lead to overlapping assistance for some families, while others would not be covered at all. In particular, most families with orphans, destitute persons, elderly over 65 years old, or whose breadwinners can get a place in Ipelegeng, receive enough cash assistance to bring their total income or consumption above the absolute poverty line. The rest of the families in absolute poverty are left out. Economic growth, especially if pro-poor, could bring some of these families out of poverty. But not all families can be lifted out of poverty solely with economic growth. 235. The reliance on programs focused on vulnerable individuals is not an efficient way to achieve the Vision 2016 target. First, it could lead to overlaps and gaps in coverage. An average family of four that takes care of an orphan, a destitute person and a senior of over 65 years old could obtain more than P1500 only from social transfers, more than twice as much than the absolute poverty line of P680. An equally poor family of five without such members would receive no assistance. Second, the current system offers too much protection for small families and too little for large families, as needs vary with family size. For example, a family of three caring for an orphan would get the same amount as a family of ten carrying for an orphan, P625 per month. The current benefit level would be enough to bring the first family over the (family) absolute poverty line of P420, but will fall short of eliminating absolute poverty for the family of ten (whose family absolute poverty line is P1400). Third, programs with very generous benefits could discourage the beneficiaries from working, along with the other family members with whom he/she shares these resources. 123 236. A more efficient way to address absolute poverty would be to offer a family benefit to all families in absolute poverty, with a per capita benefit level set around the average consumption gap209 of these families. We called this program the Family Support Grant (FSG) program. Putting in place such program would help achieve the objective of Vision 2016, elimination of absolute poverty. The (minimum) target group for such program would be the number of people in absolute poverty. The 2009/10 CWIS estimated that 16 percent of the population consume less than the food poverty line of about P170 per person per month, and hence live in absolute poverty. This is equivalent to 336,000 persons or 84,000 families in 2012/13. There is no published information about the average consumption gap of these households in 2009/10. For illustrative purposes, we rely on the estimate from HIES 2002/03, which found that the consumption of the poor was 38 percent below the poverty line. If this gap has persisted in 2009/10, it would represent P65 per capita per month.210 We suggest a slightly higher benefit, of P85, equal with half of the poverty line, to increase the share of absolute poor who will be lifted over the food poverty line by such transfer. This is equivalent to P340 per month for a family of four. The annual cost of a FSG program offering P85 per capita per month to the 84,000 poorest families, and operating with 15 percent administrative costs, would be P394 million, or 0.32 percent of GDP. This program variant is illustrated in Table 26, column 2, and named “theoretical scenario”. 237. To implement the FSG, the Government would need to develop a mechanism to identify (target) the families living in absolute poverty. This report recommends the development and use of a targeting system based on a proxy-means test (PMT), described in the next section and Annex 1. Because the consumption (or income) of a family is difficult to measure, the PMT method estimates it through a regression on a set of 209 The consumption gap is the difference between the poverty line and the average per capita consumption of the poor. 210 The consumption gap of P53 per capita per month represents 38 percent of the per capita absolute poverty line of P140 in 2009/10. 124 observable characteristics that are easy to verify and hard to conceal or falsify, such as characteristics of the dwelling, ownership of large durable items or household size and composition. This estimation could be done based on a recent household survey, such as the 2009/10 CWIS. The coefficients of this regression are then multiplied with the characteristics of each applicant family and they generate a score (in this case, the predicted level of the consumption of a family with those characteristics). Families whose predicted consumption is lower than a threshold are then considered eligible for the program. 238. To cover all families in absolute poverty, the PMT targeting mechanism would have to cover a larger share of the population (than the 16 percent), to account for some inherent leakage. All social assistance programs targeted to the poor end up including some beneficiaries that are non-poor, i.e. inclusion error or leakage (of program funds toward non-poor). Even well performing programs in developed economies have between 10-33 percent inclusion errors. A recent review of social assistance programs in developing programs (Grosh et al, 2008) shows that well performing programs have between 20-50 percent inclusion error. Many of those included but not poor are, however, in the vicinity of the poverty line. 239. To cover all or most absolute poor households, the FSG coverage should explicitly take into account that some non-poor families would be predicted poor. The simulations included in Annex 1, based on the 2002/03 HIES, suggest that this leakage rate could range between 33 and 50 percent of the total number of beneficiaries. Thus, a moderate FSG scenario based on 33 percent leakage rate should target 24 percent of the population (16 percentage points absolute poor, plus 8 percentage points leakage). If the predictive power of the PMT model would be lower, i.e. amounting to 50 percent leakage rate, then the FSG program should target 32 percent of the population (to be sure to include the poorest 16 percent of the population within the program). These program variants would be 50 percent and 100 percent more expensive than the “perfectly targeted” program sketched 125 above. They are termed Option 1 and respectively Option 2 in Table 27 below. The FSG program with a coverage of 24 percent of the population, Option 1, would cost P577 million, or 0.47 percent of GDP. The variant with a coverage of 32 percent of the population, Option 2, would cost P770 million, or 0.63 percent of GDP. Option 1 would be budget neutral, while Option 2 will require an additional 0.2 percent of GDP. Table 26 below makes explicit the assumptions used in the calculations. 240. To maintain a budget neutral stance and avoid program fragmentation, the Government could replace the programs for Orphans and Destitute Persons with the FSG program. In 2012, these two programs covered 70,548 individual beneficiaries for a total cost of about P543 million, or 0.44 percent of GDP. Currently, these two programs cover only a fraction of the population in absolute poverty. If these programs are replaced by the FSG Option 1, the new program would cover all orphans or destitute persons living in absolute poverty, and some of those living in (moderate) poverty. Option 1 will require an additional P53 million savings and will offer benefits to half a million persons, including the 16 percent poorest, with a benefit that is sufficient to lift them out of absolute poverty. Option 2 will cost an additional P246 million per year, or an additional 0.2 percent of GDP, and will cover about one third of the poorest population . 241. If replacing entitlements for Orphans and Destitute Persons is politically unfeasible, the Government still has the option to cover the safety net gaps with a complementary FSG program, at a moderate fiscal cost. One way to weave the FSG with the current programs is to offer the FSG benefit (e.g. P85 per capita per month) only to those family members who are not receiving other, more generous, individual benefits. This is equivalent to P340 per month for a family of four. This is Option 3 in Table 27. These other benefits include the food rations for orphans and destitute persons, the old age pension and the wages earned in Ipelegeng. Those who receive these benefits will continue to receive them. The other family members, who do not receive any other individual level benefit, will receive the P85 Family Support Grant. The implementation of the 126 “complementary FSG program” would require using an administrative definition of family that excludes members who receive other benefits. For example, a family of 7 members that includes a destitute person and an elderly on old age (social) pension would receive the FSG grant only for the remaining 5 members. Assuming that 25 percent of the individuals eligible for the FSG with moderate coverage would receive other individual-level benefits, this will results in an additional FSG cost of P433 million, or 0.35 percent of GDP. This increase in the social safety net budget could be compensated, for example, by the introduction of a 25 percent co-payment for scholarships or sponsorships for tertiary education, as discussed in the next paragraphs. 127 Table 27: Illustrative options and cost of the family support grant Assumptions about program leakage Theoretical Option 1 Option 2 Option 3 outside the group of absolute poor: Scenario Leakage of 33%, excluding Perfect Leakage of Leakage of family Targeting 33 percent 50 percent members covered by other programs* Percent Number of beneficiaries 16 24 32 24 population '000 Persons 336 504 672 378* P/Per Benefit Level 85 85 85 85 capita/Month Transfers (P/ Million) Per Month 29 43 57 32 Per year 343 514 685 385 Admin Costs Assumed @ 15% 51 63 85 48 Total costs Mil. P/ Year 394 577 770 433 Percent of GDP % 0.32 0.47 0.63 0.35 Savings by discontinuing the Mil. P/ Year 524 524 433 Orphans and Destitute programs: Mil. P/ Year 53 246 Savings (-) / Add'l Percent of GDP 0.32 0.04 0.20 0.35 Budget Required (+) Pro Memoria GDP P/ Million 122,500 Population 000 persons 2100 Source: Team estimates Note: a/ Evaluated with the support of UNICEF. Report to be published shortly Notes: Option 1 and 2 assumes that FSG program would replace the Orphans and Destitute Persons programs, thus are budget neutral reforms. Option 3 assumes that all existing SSN programs continue to operate. The theoretical scenario illustrates the minimum size and cost of such a program, and is presented only for comparability purposes; it cannot be implemented as such. *) The objective of the program is to cover all families in absolute poverty, i.e. the 16 percentage points poorest. The program target group is the sum of absolute poor + an assumed leakage rate. **) Under Option 3, all eligible families receive P85 per capita per month, except members on other SSN benefits (Orphans, Destitute Persons, Old age pension and Ipelegeng). 242. To implement the FSG program in the next 3 years, the report proposes a gradual approach, where the design of the program could be developed during 2013 based on the recent Core Welfare Indicator Survey 2009/10. Piloting would take place during 2014 and the full program would roll out in 2015. This gradual approach would allow the Government 128 to put in place adequate administrative systems, including targeting, payment of benefits, grievances redress mechanisms, and monitoring and evaluation. 243. The FSG should preferably be provided in the form of cash instead of food or food coupons. Replacing in-kind transfers or food coupons with cash transfers would eliminate the existing logistical and accountability problems associated with the distribution of food. Cash transfers would also allow the beneficiaries to buy what they need in terms of foodstuffs or other basic necessities, and would free up social workers to concentrate on providing support to families and other vulnerable groups. Nonetheless, this may raise opposition because of fear that the cash may be misused. There are two alternatives for addressing these fears. One would be to do a small pilot in a few constituencies to evaluate whether cash transfers would defeat the purpose of the program. The other option would be to gradually increase the amount of cash transferred together with the food coupon and evaluate whether it results in widespread misuse. At the same time, policymakers could consider giving the transfer to the women in the house as in many countries they have been shown to be more conscientious caretakers and to make better use of the money than men, which might also be the case in Botswana. 211 Similarly, social workers would need to work closely with the families to quickly detect any misuse of the transfers. 244. Over medium term, the Government could consider enhancing the program by introducing work requirements for work-able individuals (in areas where the public works program does not operate) and by conditioning cash assistance in exchange for certain behavioral changes related to the use of nutrition, health and education services. A program of this type offers the opportunity to promote behavioral changes to help the family increase the human capital of their children and thus break the intergenerational transmission of poverty. Conditional cash transfers (CCTs) have been very successfully 211 If it is decided to pilot the full cash transfer, then the question of who in the household should receive the transfer could also be evaluated. 129 implemented in a variety of middle-income countries (Box 5). They have been rigorously evaluated and found to be highly effective in improving child outcomes, including poverty status, nutrition, health, and schooling.212 Most recently, Burkina Faso conducted a randomized program evaluation of pilot cash transfer program in rural area with the goal of comparing the impact of conditional versus unconditional cash transfers. The CCT was very effective at getting parents to invest in children who they normally would not. The conditionality pushes parents to enrol marginalized children —those not enrolled at baseline, girls, and low ability children.213 Box 5: Conditional Cash Transfers Conditional cash transfers (CCTs) have been very successfully implemented in a variety of middle-income countries. CCTs consist of small cash payments to the mothers of poor children in return for a commitment from the mothers to ensure that the children will attend school (or health clinics in the case of pre-school children). CCTs have been rigorously evaluated and found to be highly effective in improving child outcomes, including poverty status, nutrition, health, and schooling. CCTs have two objectives: (i) to alleviate poverty in the short term by providing households with a cash transfer and (ii) to reduce inter-generational poverty by requiring families to invest in their human capital,by keeping their children in school or maintaining them in good health through preventive health and nutrition programs. The value of the benefit is usually equivalent to about 15 to 25 percent of households’ baseline consumption. The fiscal impact varies depending on the size of the program, but nationwide programs such as those in Brazil, México, or Colombia cost about 0.4 percent of GDP per year. Evaluations have shown that CCTs can be effective in improving health, education, and malnutrition while also increasing the consumption of the poor in both middle- and low-income countries. For instance, in Brazil, between 2003 and 2008, CCTs explained 35 percent of the reduction in absolute poverty and 13 percent of the fall in inequality. In Colombia during the first two years of the program, beneficiaries increased their total consumption by 15 percentage points more than households who did not participate in the program. In India, a conditional cash scheme increased ante-natal care and in- facility births significantly and has been responsible for a reduction of 4.1. peri-natal deaths per 1,000 pregnancies and 2.4 neo-natal deaths per 1,000 live births. In Turkey, the enrollment of girls in secondary school increased by 11 percent. Evaluations of the programs in Malawi, Mexico, and Ecuador have demonstrated that conditional programs have a greater impact on school enrollment and attendance than unconditional transfers. However, CCT programs have to be carefully designed and implemented if they are to succeed. Experience has shown that it is critically important to: (i) define the level and type of transfer (that is, whether it is paid per family, per beneficiary member, or per condition); (ii) define the beneficiaries’ co-responsibilities and set up a verification process to monitor their compliance; (iii) choose appropriate targeting mechanisms (CCT programs usually combine geographical, proxy means tests, and community targeting methods); (iv) maintain the frequency of payments; (v) ensure payment mechanisms are transparent and are subject to controls and audits at the local level; and (vi) regularly re-evaluate and re-certify program beneficiaries. Finally, it is critical that beneficiaries can access the services that they are required by their participation in the CCT program to use (usually schools and health providers) and that these services be effective. Source: Fiszbein and Schady (2009) 212 Fiszbein and Schady (2009). 213 Damien de Walque (2013). 130 245. The Family Support Grant could provide cash transfers to families in return for behavioural changes that would build their human capital and help them to rise out of poverty. The persistence of high rates of under-nutrition may be an argument for ensuring that transfers are conditional on behavioural changes such as attending health clinics for growth monitoring, participating in feeding and caring practices education, giving babies micronutrients, and taking severely undernourished children to clinics for therapeutic feeding. Also, the relatively large number of primary school age children who are out of school may be another argument to require parents to send their children to school. A conditional cash transfer could also help reduce the large drop in secondary school enrollment by giving parents an additional incentive to ensure that their children continue studying and complete secondary education. However, conditionality adds to the operating cost of the program. If the government decides to implement the Family Support Grant differently, then each community should be fully briefed about the purpose of the program. Communities should help ensure that the families selected as beneficiaries are those that are most in need and that families headed by children, OVC, children with nutritional problems, and poor pregnant women are included. Furthermore, the beneficiary families must be made aware of what is expected of them—that the diet of the family must improve and that the children must attend and progress in school and visit health clinics regularly. These communication campaigns will require trained staff and resources. 2. Strengthening the Public Works Program Ipelegeng 246. Over the short term—in 2014—the Government should focus on improving the targeting and214 effectiveness of the public works Ipelegeng program, a key intervention for improving the employment opportunities for work-able absolute poor. This will require adjustments to three parameters: a reduction of the wage rate; and increase in the 214 At the time of writing of the report, the CWIS micro data was not available to estimate the targeting accuracy of the Ipelegeng in 2009/10. 131 number of months of employment offered and/or places in the program; and an allocation of the program budget proportional with the regional poverty rate. 247. To improve its targeting, the program needs to offer lower wages for an extended period of time. The current Ipelegeng wage rate (about P580 per person per month, including the food component) is above the minimum wage, which diminishes its self- targeting potential by increasing the demand for the program from otherwise non-poor individuals. In turn, this leads to rationing the excess demand through a lottery system. In addition to the wage rate issue, the government could also review the number of days for which work is offered. The rule is not clear. For some parts of the program (Green Scorpions), the maximum is one year, while for others (wildlife volunteers), it is three months, but for the majority, it is one month with the possibility of being renewed but with no guarantee. This period is significantly lower than that used in in most countries, and is not enough to help the poor to cope with unemployment (seasonal or otherwise). For example, reducing the wage rate by one third, or P390 per month (in line with the minimum wage for agricultural work), would allow extending the duration of the program from 8.7 to 12 months under the same budget, and serve 10 percent more beneficiaries (Option A in Table 28 below). Alternatively, it could expand the coverage of the program to 81.8 thousand places under the same budget and duration of the program (Option B in Table 28 below). Both these option would improve the targeting of the program, i.e. would ensure that the applicants to the program are among the poor. Table 28: Reforming Ipelegeng: Offering a lower wage rate would not only improve targeting, but cover more poor people for longer period of time 2012/13 2013/14 Actual Option A Option B Overall Budget P/ Million 409 409 409 Labor costs Percent 68 68 68 P/ Million 278 278 278 Wage rate P/ month 580 390 390 No beneficiaries Persons 55,000 60,000 81,800 No of months 8.7 11.9 8.7 Source: Team estimates. 132 248. The effectiveness of the program could be increased by allocating budget to regions proportional with the prevailing or estimated poverty rate. The program also suffers from a suboptimal allocation of its funds across regions and areas. The number of places in the program should be related to the incidence of poverty and food insecurity in each constituency rather than to their share of the population, as the formula now specifies. This rigid allocation formula for the Ipelegeng often leads to excess supply in Gaborone and excess demand in most other places. To increase its effectiveness, the size of the program (and, implicitly, its budget) needs to be determined based on the total number of poor work-able adults; and its regional distribution needs to be determined based on the regional distribution of absolute poverty (e.g. based on the regional estimates of absolute poverty of the CWIS, or even better through a poverty map). Furthermore, in rural areas the duration of the works and their location should be related to seasonal needs, particularly during the period of the year of greatest food insecurity. 249. A better targeted Ipelegeng would help reduce absolute poverty among families with work-able adults who either lack employment opportunities or work in low- productivity, low-return sectors. For example, a monthly wage rate of P390 could lift a family of 4 people out of absolute poverty with a consumption gap of 50 percent (whose consumption, in the absence of the program, would have been P340, or half of the household-level food poverty rate of P680). With the program, the total household income would increase to P730. However, the public works income would not be sufficient to lift larger families (6 or more people) out of poverty. For such families, especially those which currently do not receive other types of individual assistance (e.g. Orphan Care, Destitute Persons or Old age pensions), there is a need for a family-based anti-poverty program. 250. In addition to above suggestions which would help the program achieve its primary objective of poverty relief and a bridge to more permanent employment, the Ipelegeng might also consider actions to improve on the quality and utility of the public 133 works. Some countries have used public works to help communities suffering from severe environmental degradation and limited access to infrastructure (see Box 6). Box 6: Ethiopia Productive Safety Net Program The Productive Safety Net Program (PSNP) in Ethiopia is a public works program based on an integrated watershed development approach. This approach is recognized both nationally and internationally as an appropriate method for soil and water conservation. The public works projects result in the creation of community assets that contribute to sustainable livelihood and long-term developments such as soil and water conservation structures, feeder roads, social infrastructure, water supply for human and livestock consumption, and small-scale irrigation facilities. A guideline authored by the Ministry of Agriculture and Rural Development provides development agents—the government staff members employed by woreda agricultural office to work at kebele level on agricultural extension activities—and rural communities with a workable and adaptable planning tool. Project selection starts from the bottom up and depend on the participation of community members living in the relevant micro watershed,as they are the people who will live with the results, and the program is one of their tools for addressing food insecurity. A key principle is to optimize the use of existing natural resources and untapped potential in degraded watershed areas. There is a planning hierarchy consisting of the watershed, the watershed unit, and the micro-watershed. Development agents must ensure that only sustainable projects are selected. Therefore, all projects should have defined owners after completion—people or groups with rights of use, and therefore the obligation to maintain and manage the assets. In many cases, user rights arrangements will involve the creation of group bylaws governing access, management, and responsibility obligations. They may also require the establishment of a system to collect user fees. These user rights and maintenance responsibilities need to be defined during the planning process to establish a sense of ownership by the benefiting community or group. Source: Subarrao et al (2013) 3. Nutrition Programs 251. To address the high and persistent level of chronic under nutrition, we recommend a thorough review of the current Vulnerable Group Feeding Program, which does not seem effective. The review should focus on both the design and implementation of these programs. In particular, it should determine whether the design of the program conforms to good practice principles (e.g. focusing the support on the first 1,000 days, using proven interventions), and whether the actual implementation of the program follows these principles. The recommendations of the review should help improve the effectiveness of the program by 2016. 134 252. The supplementary and therapeutic feeding (currently provided by the VGFP) should probably be redesigned to limit the interventions to the age group most likely to benefit from this type of interventions (the first 1,000 days, pregnant and lactating women) and focus on nutrition interventions which were proven effective. These include nutrition hygiene education, growth monitoring and therapeutic feeding for severely malnourished children. Some MLG staff appears to be dissatisfied with existing arrangements in which they are responsible for procuring the foodstuffs for those programs but have no control over how these are managed and used. The staff is blamed when things go wrong, and the divided responsibilities lead to lack of accountability. 4. School Feeding Programs 253. The approach to primary school feeding program should be re-evaluated. First, it must be determined whether it would be preferable to provide schools with financial support rather than with foodstuffs. As argued above, having schools buy their own foodstuffs locally would help to develop the local economy as well as marketing channels, and in the process reduce existing isolation of some communities. According to a recent study, the budget that was used for the procurement of local produce in the 2011 cropping season was only 1 percent of the total budget of the Primary School Feeding Program. 215 The study recommends among other things: • Including more local foods and improving the nutrition quality of the menu. • Exploring different procurement options and modalities to allow the program to benefit more small farmers. • Promoting more private sector involvement in the program. • Improving coordination and information management systems. 215 BIDPA (2012), page xiii. 135 • Integrating school feeding with other activities within the schools’ health and nutrition interventions for a more cost-effective strategy to support positive educational and nutrition outcomes.216 254. Procurement should be moved to the schools and financed by direct transfers to the schools, as is already the case in the Secondary School Feeding Program. Also, food coupons and cash are already being given to the beneficiaries of other programs to buy similar products in these communities. Consideration should be given to targeting the free school feeding programs to poor students, particularly poor child beneficiaries of other transfer programs (such as the proposed Family Support Grant). According to a recent World Bank study, “in high- and middle-income countries, free school meals are generally integrated within social protection programs targeted to individual children on the basis of vulnerability and means-based proxies. Children not considered at risk would normally pay for the meal, though often at subsidized cost.”217 The GoB could transfer a given amount per poor student to the school /parents association. In the poorest areas, the program could cover all children while in other areas, some cost recovery could be considered. The schools and parent associations would help to determine how cost recovery was to be implemented, provided that the meals (in quantity and quality) are guaranteed for the targeted students. Box 6 reviews some international experience with these types of programs. 216 BIDPA (2012), page xiv. 217 Bundy et al (2009). 136 Box 7: International Experience with School Feeding Programs Every country for which information is available now operates some kind of program to provide food to its schoolchildren. Middle-income countries run near-universal school feeding programs, with India, for example, covering 120 million children annually and Brazil covering 35 million. The programs provide meals, fortified biscuits, or take-home rations. On-site meals are a relatively expensive way of feeding the poor because they provide meals to many children who may not need them. Take-home rations cost slightly more but can be better targeted, allowing more coverage for the same expenditure. Globally, costs average US$40 per child per annum for on-site programs, slightly more for take-home rations (US$52 for a sample of four programs), and much less (US$13 per student) for fortified biscuits. Targeting is often geographical. In Indonesia, 530,000 students in vulnerable areas receive fortified biscuits at a cost of about US$18 each per annum. Bangladesh provides high-energy biscuits to about 400,000 students in food-insecure areas (for a cost of US$12 per student per annum). The evidence on how school feeding affects educational outcomes and attendance is positive but not very strong. As a nutrition intervention, school feeding has been shown to have some positive impact, but it reaches children later than the critical under-30 month age and is nowhere near as cost-effective as supplemental feeding or food fortification. On balance, school feeding programs are not a particularly cost-effective way of achieving either educational outcomes or consumption transfers, but they have the benefit of contributing to both goals simultaneously and of being politically and socially acceptable. 5. Sponsorships and Scholarships 255. Some of the measures recommended would require additional budget. To finance the additional budget for the FSG program, savings could be obtained over the short term by reducing the generosity of costly programs (e.g. scholarships).The current social assistance spending is highly skewed in favour of a single program: scholarships and sponsorship for tertiary students. This program costs the budget P1.674 million in 2012/13 or 1.4 percent of GDP and absorbs 45 percent of the total social assistance spending. To achieve the Vision 2016 goal, the country could reallocate a small share of these resources away from this program and channel them through programs with good targeting accuracy. The government could reduce the generosity of scholarships for tertiary education by about 25 percent starting from 2015 for high income students. These programs are likely to be regressive, benefiting mostly rich and upper-middle-income students. Their use results in a private benefit, with fewer positive externalities than in primary or secondary education. 137 While these programs have multiple objectives to support skills development, there is a weak rationale for an across-the-board state intervention in this market from a public policy perspective. Students from high-income families could cover a share of the cost of their tertiary education, while the report recommends that poor students maintain their merit- needs based scholarships covering full cost. 256. Given the high cost of the sponsorship / scholarship program in the present, a reduction of the generosity of the program could pay for a large share of the budget of the FSG program. A 25 percent reduction in the generosity of the sponsorship / scholarship program would result in savings of about P420 million per year. Assuming that one third of these students do come from the poorest half of the income distribution, and would continue to enjoy full payment of their sponsorship and scholarships in the future, the savings that could be obtained by such a measure amount to P280 million per year. These savings are sufficient to pay for the FSG program, Option 3. 6. Active Labor Market Programs 257. Botswana operates a number of active labor market programs (ALMP) that enhance youth employability and worker skills, but have low coverage and high benefits (e.g. for internship programs), are disconnected from SSN interventions, and have not been evaluated. Under its Human Resource Development Strategy, which is currently being implemented, the GoB has already decided to streamline the existing institutional arrangements, eliminate existing fragmentation, and assign clear roles and responsibilities to the sectoral institutions. The Human Resources Council, which should be established shortly, plans to begin operating an effective web-based labor market exchange (already developed), to keep track of labor market developments through the Labor Market Observatory, and to rationalize youth programs such as the Internship Program while continuing to support programs that have reportedly shown promising results such as the Apprentice Program. To get more value for money from these programs, the Government 138 could link the ALMPs and SSN beneficiaries, recalibrate the benefits of the internship program, and combine skills training with internships in the private sector in youth employment programs. 7. Contributory Pensions 258. The contributory pension schemes protect the covered population against the drop in income during old age, but have low coverage and are excessively fragmented. The National Development Plan 10 calls for a review of the social insurance system to increase social security coverage through comprehensive social security policy and legislation. At the request of the GoB, the ILO is currently evaluating the possibility of establishing a mandatory pension plan in Botswana. Currently only public employees are required to contribute to the BPOPF. The limited amount of information that we have on social insurance indicates that Botswana faces three challenges in this area: (i) low coverage, as social insurance covers less than one-half of the formal labor force and about 13 of the labor force; (ii) large-scale fragmentation of the private system (with over 97 private pension plans in existence); and (iii) lack of unemployment insurance. A revised social insurance system in Botswana should: (i) eliminate the existing fragmentation of the private pension system; (ii) continue to allow for voluntary savings; (iii) facilitate the enrollment of the nearly 60 percent of the formal labor force with no insurance; and (iv) facilitate the participation of informal sector workers. Eventually, the existing universal OAP could link to the pension plan and be transformed into a minimum pension guarantee for all those who were not able to contribute to their retirement or contributed only very little. 259. There is substantial room, therefore, to exploit synergies between programs by linking programs and services and by building an effective social protection system. On social assistance, transfers to families should be linked with services provided to the children as well as training provided to the parents to enhance their skills and nutritional education. Transfers to the elderly and the disabled should be linked to services that they 139 require, particularly health services, but also to other programs that ensure their continued meaningful participation in the community’s affairs. The public works program should be linked to skill-enhancing programs and potentially to financial literacy, savings, and investment services. Active labor market programs should provide youths with employability skills and on-the-job training to facilitate their transition to work but should be very parsimonious in the benefits paid to instill a work ethic in young people. Finally, the social insurance programs could be eventually linked to the OAP. C. Improving Administrative Systems and Tools 260. To obtain more value for money from its current investment in social protection programs, Botswana could consider improving the administrative systems and tools that underpin service delivery across key programs, building on the already significant progress made in this area. These tools include targeting methods, beneficiary registration and management information systems (MIS), payment mechanisms, grievance systems, monitoring and evaluation (M&E). In addition, the Government could benefit from a greater use of administrative and survey data in assessing current programs, simulating new programs (e.g. the FSG program), or modifying the parameters of current programs (e.g. the parametric reforms to Ipelegeng). 261. Making Targeting More Accurate. Targeting could be further improved to ensure that most of the benefits reach the intended beneficiaries. Botswana’s safety net is largely either self-targeted (the Ipelegeng program) or categorically targeted (the school feeding programs, the Old Age Pension, and the Orphan Care Program). The only means-tested benefit, provided by the Destitute Persons Program, is extremely small, covering only 2 percent of the population, and the means testing is rarely done because of difficulties involved in verifying the declared information. This leaves the selection of beneficiaries to the best judgment of the social workers and the local community. 140 262. In countries with large informal sectors, proxy means testing (PMT) has been used to narrowly target safety net programs. The advantage of PMT is that it requires less information than means testing. In PMT, rather than trying to measure total income perfectly, an index or a “score” that predicts poverty is generated using information on the characteristics of the applicant households that are correlated with poverty. These characteristics are much easier to measure and verify than income and can include the number of children in the family, the location and quality of the family’s dwelling, their ownership of durable goods, the demographic structure of the household, and the education and, possibly, the occupations of the adult members. Being easy to measure, these proxy variables cost relatively little to gather and verify. Another advantage of a PMT is that, because it does not measure money income itself, it may be less likely to discourage beneficiaries from seeking work than an actual means test would.218 PMTs do require some administrative capacity to implement, mostly a good MIS and administrative staff to enter the applicants’ data. Most countries have used temporary workers (often college students) to help with the large inflow when a system is first set up or for a mass recertification, but existing staff can usually handle the new cases thereafter. 263. If introduced, a PMT targeting scheme would outperform existing categorical schemes identifying the poor. To examine this, we compared a hypothetical PMT targeting scheme to some of the categorical schemes already being used in Botswana, using household-level data from the 2002/03 HIES to simulate what would happen if we use these alternate schemes. Annex 1 describes the PMT model and presents a detailed simulation analysis. We compared the PMT model to an elderly allowance, an orphan grant and also a universal child grant. Targeting the elderly, orphans, or children may be a policy 218 The indicators and the formula used to combine them into a single PMT score are derived from statistical analysis (usually regression analysis or principal components) of data from detailed household surveys of a sort too costly to be carried out on all applicants to large programs. The information provided by the applicant is usually partially verified either by collecting the information on a visit to the home by a program official or by asking the applicant to bring written verification of part of the information to the program office. Eligibility is determined by comparing the household’s score against a predetermined cutoff (http://go.worldbank.org/55Q72CB9Q0). 141 objective in itself; but in this exercise, our focus was on evaluating these hypothetical programs in terms of how well they target the poor. 264. Compared to any single categorical targeting, a PMT scheme is more flexible. The target population is essentially pre-determined in any categorical targeting scheme. For instance, targeting all households with an elderly member automatically means covering 24 percent of the population, which is the share of the populating residing in households with elderly members. In a PMT, we can vary the coverage by varying the cutoff PMT score that determines eligibility for the transfer. For a given budget, we can reach more individuals but with a lower transfer per beneficiary or vice versa. The key finding of the simulation analysis was that PMT targeting compares favourably to categorical targeting schemes such as the elderly allowance and the orphan grant in that it more accurately identifies the poor. This was true in both rural and urban areas. 265. Our simulation analysis showed that the PMT was more accurate at identifying the poor. The presence of an elderly member or an orphan is strongly correlated with poverty. Specifically, the poverty rate in households with elderly members is 43 percent, as opposed to 26 percent in other households. The poverty rate in households with an orphan is 38 percent, as opposed to 27 percent in other households. These correlations mean that transfers to households with elderly members or orphans will be pro-poor. However, because the PMT score incorporates more information, it is potentially even more correlated with poverty, and hence more accurate. The PMT targeting scheme would have an error rate of about 31 percent, as compared to 34 to 37 percent for the categorical schemes that we considered.219 It is important to keep in mind that all administrative targeting mechanisms have a certain degree of error. 219 The error rate is the percentage of individuals who are misclassified by the scheme in question. Misclassification occurs when a poor person is assumed to be non-poor according to the scheme or when a non-poor person is assumed to be poor. 142 266. The superior accuracy of PMT can be captured in an “effectiveness rate” or the percentage of spending that goes into reducing the poverty gap.220 Our analysis estimates that the effectiveness ratio of a PMT scheme is about 36 percent, while those of the categorical schemes are in the 23 to 28 percentage range. To get a sense of the trade-offs, let us consider the following hypothetical exercise. In 2002-03 terms, an elderly grant of 150 pula per month221 would mean a total transfer budget of about 13 million pula. With an effectiveness rate of 28 percent, 9.4 million pula would be “wasted.” For the same total transfer budget, the wastage in the PMT scheme would be about 8.2 million pula, or 1.2 million pula lower. These “savings” in terms of targeting efficiency have to be weighed against the potential higher cost of administering the PMT system. As the efficiency savings from the PMT increase with the coverage rate while administrative costs are relatively flat, the PMT looks increasingly more effective with wider coverage. 267. Botswana could develop a PMT-based targeting scheme to target the poor more accurately in the next two years. The calibration of the PMT formula could be done based on the relatively recent CWIS 2009/10. The simulations presented in Annex 1 based on 2002/03 HIES show that developing such a scheme is feasible and cost-effective. However, the information from the HIES 2002/03 is old and does not reflect the current welfare distribution in Botswana. To jumpstart the process, the Government could place the CWIS data under open access and contract a consultant to develop the PMT formula, as well as the operational manual for administering eligibility under the PMT scheme. 268. While the PMT is a precondition for the development of the FSG program and filling the gaps in the existing safety net, Botswana could use this targeting instrument 220 The effectiveness ratio depends on targeting accuracy and also the transfer amount relative to the distance from the poverty line. To understand why, consider that even perfectly accurate targeting will have a less than 100 percent effectiveness ratio if the transfers are so large that some of the beneficiaries cross the poverty line after the transfer. The part of the transfers that lifts the formerly poor up to the poverty line (or just below) is considered to have been effectively spent, while anything beyond that is considered a “waste” according to this measure. 221 An amount comparable to the actual Old Age Pension amount in 2007 in real terms. 143 over time across a larger set of programs. The beneficiaries of the Poverty Eradication Initiatives could be selected among those identified with the PMT as poor. Also, the PMT system could be used to determine which students would be entitled to a fully paid sponsorship or scholarship (e.g. if their family income ranks them in the bottom half of the income distribution), or else have to pay a co-payment.222 269. Creating a Central Beneficiary Registry and an Integrated MIS. Collecting information about the beneficiaries of all safety net programs in one central beneficiary registry (CBR) would make it possible to get a clearer picture of the poor in Botswana and of what benefits they receive and for how long. Currently, the government has no way of knowing at the household level what benefits are going to whom, nor does it have any information on the socioeconomic characteristics of recipients such as their poverty status. The monitoring and evaluation of programs is hampered by the lack of comprehensive information on who receives what benefits. 270. A central registry would consolidate key information on potential program beneficiaries into a single common database. This would: (i) identify the poor and vulnerable and make it possible for programs to identify their specific target populations quickly; (ii) provide consolidated information on program beneficiaries to policymakers; (ii) enable the cross-checking of beneficiary lists to reduce the possibility of recipients double dipping; (iii) make it easier to verify beneficiaries’ identities by linking the single registry with other administrative databases including the central civil registration system; and (iv) act as the foundation for creating other common delivery systems, including payment systems, which would reduce the costs associated with delivering benefits. 222 The PMT could be used to identify disadvantaged students who currently are not enrolled but who could benefit from the scholarship and sponsorship programs. The higher-income students will continue to be subsidized, but at a lower rate. This targeting mechanism is compatible with the provisions of the Childrens’ Act (2009), which establishes the right to education for every child in Botswana. 144 271. The key steps that would be involved in developing a central registry would be: (i) to define the population to be targeted by various safety net programs and to determine what information needs to be collected to establish their eligibility; (ii) to disseminate information about the programs’ objectives, benefits, registration process, and eligibility rules to communities of potential applicants; (iii) to collect information from potential beneficiaries and enter the information in the registry; (iv) develop information systems to capture, verify, and transfer data between the registry and the databases of individual programs; and (v) set up systems for data management and maintenance. For an applicant, being included in the registry would not automatically guarantee that he or she would be enrolled in the program. Each program would use its own criteria to screen potential eligible beneficiaries. Good practice has shown that a registry must be inclusive so that all individuals who consider themselves as poor could register. The requirement for applicants to produce proper means of identification (such as ID cards, tax numbers, birth certificates, or driver licenses) should not preclude potential beneficiaries from registering. Botswana has a national identification card known as the omang, the presentation of which is a prerequisite for participation in all social assistance programs. By law, all Batswana above the age of 16 must have one of these cards. The existence of the omang should make it easier to establish and manage the registry of all assistance program beneficiaries. 272. Several countries have developed this sort of registry to identify beneficiaries of their safety net programs, including Brazil, Colombia, the Dominican Republic, El Salvador, Costa Rica, Kenya, and Lesotho. In Brazil, the Ministry of Social Development and Fight against Hunger administers the Cadastro Unico (CADÚNICO), a central registry that collects information on all poor and vulnerable families to facilitate the distribution of federal social assistance programs. In 2011 CADÚNICO registered about 20 million families (one-third of the Brazilian population), of whom 13 million families are beneficiaries of the Bolsa Familia cash transfer program, 4 million are beneficiaries of Prestacao Continuada, an old age pension, and another 8 million are benefitting from other social assistance programs (Box 8). 145 Box 8: Building the National Registry in Brazil (CADÚNICO) CADÚNICO was created in 2001 for the Bolsa Escola, a scholarship program. In 2003, the CADÚNICO—with 5.5 million families enrolled in 5,190 municipalities—became the backbone of the Bolsa Familia conditional cash transfers program that unified the main four social assistance programs in Brazil. In 2005, the government began transforming the CADÚNICO from a database of Bolsa Familia applicants into a database of the entire poor and vulnerable population in Brazil. In 2007, the government passed the CADÚNICO law that defined the registry’s goals, processes, and operational rules and reaffirmed its role as a single entry point for all federal social assistance programs. The current CADÚNICO can provide timely and updated information about potential beneficiaries to all social assistance programs (with recipient families being recertified every 24 months). In 2011, the MIS of the CADÚNICO went online with pre-defined security levels for data entry and maintenance to avoid data manipulation. The registry allows: - Different municipalities to consult with each other about families in the national database. - Beneficiaries to be included in, updated, or removed from the single registry. - Programs to verify the identities of their applicants or to identify any inconsistencies or missing information. - Program supervisors to obtain timely information about the interviewer or data entry clerk. - The identification of potential beneficiaries by agencies implementing social assistance programs. - Program implementers to obtain information about families and the payments that they have received from government programs over time. - The generation of monthly reports and descriptive statistics about population in the CADÚNICO. CADÚNICO helped the rapid expansion of the conditional cash transfer program Bolsa Familia which now reaches 13 million families (over 50 million people) or 30 percent of the population, making it the largest program of this type in the world. The program has a very impressive targeting accuracy with over 75 percent of the benefits reaching the bottom 20 percent of the population. Moreover it has demonstrated significant impacts on reducing poverty and inequality. Source: Leite et al (2012). 273. The GoB may consider establishing a comprehensive registry for the proposed Family Support Grant and linking it to the existing beneficiary databases of the existing Destitute Persons Program, the Ipelegeng Program, the Vulnerable Group Feeding Program, and the Orphan Care Program. Botswana has the advantage that several programs keep comprehensive lists of beneficiaries and this could facilitate their registration and updating of information. The wide coverage of national identification card should make it easier to detect duplication and ghosts beneficiaries in the initial stage of combining individual program databases. 146 274. Another administrative improvement for Botswana’s safety net would be the development of an integrated management information System (MIS). The main existing management information system (which contains data on the Old Age Pension and on the Veterans Program) is actually partly paper-based. In the absence of a fully computerized MIS, the government cannot receive regular updates on how many beneficiaries are receiving benefits, and there is no mechanism for tracking payment delays or complaints and grievances. The MIS for the Orphan Care Program is also paper-based and runs on a different platform than the main MIS. 275. Each social assistance program must carry out a series of routine transactions including those related to the targeting and enrollment of beneficiaries, the payment and reconciliation of benefits, the handling of appeals and grievances, the updating of family histories, and monitoring and management reporting. All of these transactions could be done more efficiently and in real time if there were a single integrated MIS for each program or a set of programs. A modern MIS building on the existing systems should be developed for the suggested Family Support Grant. 276. Once the registry and the program MISs become fully operational, they should be able to communicate with each other and update household information in real time as shown in Figure 12. 147 Figure 12: Single registry of beneficiaries and program management information system 277. Payment Mechanism. Botswana has started using modern technology to deliver benefits, and this effort should be expanded. Smartcards are already being used to deliver cash or pay for food in the Destitute Persons, the Orphan Care, and the CHBC Programs, and check book-like leaflets have been introduced to facilitate the delivery of the OAP and the Veterans Pension. Beneficiaries are reported to be satisfied with the new smartcards. Logistics are much smoother with the smartcards, as the previous paper system required recipients to pick up their entire food package once every month, and their food would often spoil before they received their next month’s allotment. 148 278. The next step would be to move from food coupon to cash in all transfer programs with the exclusion of the supplementary and therapeutic feeding programs because of the requirements of specialized foods. As discussed, this could be done gradually and with appropriate complementary measures including communication and education to minimize the possibility of misuse of funds. This would empower the poor and give them much more flexibility to purchase what they need and even to save or invest in income-generating activities. 279. In a number of countries governments have sought to increase the use of electronic means for payments of cash transfers and to promote greater financial inclusion. On the one hand, electronic payments were seen as likely to reduce the cost of payment for government and make delivery more convenient for recipients, compared to the prevalent cash-based schemes, which require recipients to be in a particular place at a particular time to receive payment. On the other hand, a bank account was seen as the portal into the wider world of formal financial services, such as savings, insurance, and credit. Using these services appropriately would enhance developmental benefits from social cash transfer schemes (Bold 2012). More recently a review of the evidence from four countries with large well established cash transfer schemes —Brazil, Colombia, Mexico and South Africa—has reiterated this move towards electronic payments and financial inclusive mechanisms (Box 9). 149 Box 9: Social Cash Transfers and Financial Inclusion: Evidence from four countries Brazil, Colombia, Mexico and South Africa have pursued the twin objective of electronic government payments and financial inclusion. Three main approaches to making social cash transfer payments can be distinguished in these countries: (i) Cash, (ii) Limited Purpose Instrument, and (iii) Financial Account (see Figure below). There has been a move overtime from approach (i) to approach (ii) and (iii) which offer financial inclusion. Type 3 payment mechanisms offer programme’s the greatest flexibility and beneficiaries the greatest benefits. In Brazil, a small portion (1 percent) of Bolsa Familia recipients is paid in cash, while 84 percent use the limited purpose Social Card from which they must withdraw the benefit within 60 days and onto which they cannot deposit additional funds. The remaining 15 percent are paid into a mainstream Caixa Facil bank account, the same account that is used by 10 million non-Bolsa Familia recipients. In Mexico, 66 percent of recipients are still paid in cash, while 34 percent are paid into mainstream financial accounts. Of these, 16 percent are paid via magnetic-stripe cards linked to Bansefi no-frills interest-bearing savings accounts (Debicuenta). Payments through Debicuenta accounts are offered only to recipients in areas (mostly urban) where there is bank infrastructure since payments are made online from any point in the country’s national payments network. Twelve percent are paid via smart cards linked to Bansefi prepaid noninterest bearing no-frills accounts. This modality is used where there is limited payment and communication infrastructure since it enables offline transactions. The remaining 6 percent are paid into fully transactional Bansefi savings accounts, but are accessible only at Bansefi branches with no accompanying card (this is being phased out as a payment mechanism). In South Africa, the Sekulula account offered by AllPay, a subsidiary of ABSA Bank, is a specialized debit card-linked bank account that can be used as widely as any debit card in the country. The recipient can also deposit funds into the account, suggesting a mainstream financial account. 150 Programme Brazil Colombia Mexico South Africa Child Care Familias en Bolsa Familia Oportunidades Grant, Old Age Accion Pension Revamped 1998 (OAP Year started 2000 1997 2003 1928) # hhs/recipients 12.9 million 2.4 million 5.8 million 9.0 million Percent of 30 11 20 30 population Avg grant per US$70.10 US$55.10 US$118.2 US$144.7 rec. Fee per payment US$0.84 US$6.24 US$2.52 US$3.50 Fee as % avg 1.2% 11.3% 2.1% 2.4% grant The four large well established programmes demonstrate the economies of scale associated with payment costs. All programmes use at least two types of delivery mechanism. The cost of a mainstream account is generally less than limited purpose instruments that may use proprietary technology (technology that is owned by the Payment Service Provider and is not therefore following international standards). 280. Grievance Mechanisms. In theory, all programs in Botswana already have grievances and redressing mechanisms. Regarding cash transfers (the pensions and the Destitute Persons cash allowances), complaints are first presented to the staff of the Social and Community Development Department (S&CD) in the Local Authority and Village Pension Committees. Grievances can be presented more formally to pension officers. If satisfaction is still not obtained, appeal may be made to the District Commissioner or, ultimately, to the Commissioner of Social Benefits. Regarding other social assistance programs (the Orphan Care Program), complaints are presented to S&CD staff or the Head of the Social and Community Development Department. Higher appeals may be made to the Chief Executive Officer of the council or above him/her to the council’s Education, Health, and Social Services Committee.223 Since most of these programs have not been evaluated, it is not clear whether these mechanisms work in practice or not. 223 Ministry of Local Government (2010b), page 47. 151 281. The GoB should ensure that all social assistance programs have functioning grievance and appeal mechanisms so that applicants who believe that they have been incorrectly ruled as ineligible for the program or that they are receiving an incorrect amount of benefit can complain. There should also be procedures for reporting program fraud or other irregularities. All potential beneficiaries should be given information on their rights and obligations with regard to their recertification, their continued eligibility, how to leave the program, and how to file a complaint or appeal. The individuals who hear the appeals should be completely independent from those managing the program. The process for resolving appeals should be made clear to all stakeholders. 282. Monitoring and Evaluation (M&E). Botswana should invest more in evaluations of its social protection programs because the results of such evaluations are invaluable for informing subsequent policy decisions regarding the safety net. Only five of the major safety net programs have so far been evaluated, and some of those were not carried out recently. Without rigorous evaluations of Botswana’s safety net programs, the government cannot know what is and is not working. New programs like the Alternative Packages Program were introduced with no provision being made in their design for impact evaluations to be carried out to assess their effectiveness. In addition to impact evaluations, regular process and implementation evaluations are equally important for gathering information on whether beneficiaries and program staff clearly understand program rules and objectives and if procedures are being applied correctly. 283. The BCWIS should include the questions required to evaluate the major social assistance programs, including the numbers of people covered, the amount of transfer that they receive, or the reasons why applicants are rejected. The income portion of the 2002/03 HIES questionnaire did specify several safety net programs including the Destitute Persons Program, the Orphan Care Program, the Old Age Pension, the Veterans Program, and the RADP, but in the 2009-2010 BCWIS, these programs seem to have been lumped 152 together under the heading “government aid.” It is not clear how “government aid” is defined, and its usefulness for analysis is very limited. 284. The Government should facilitate the increased frequency and utilization of administrative and survey data for evidence-based policy making, by promoting an open access policy. Policy analysis and planning in Botswana is hindered because of the lack of up-to-date survey information. Labor force and Botswana Core Welfare Indicators Survey (BCWIS) surveys are done only every seven to nine years, which is too infrequent to allow for timely analysis and policy corrections. The detailed results of the 2009/10 BCWIS are still not available, and some of the analysis in this assessment had to rely on data from the 2002/03 survey. In particular, the Government could benefit from a greater use of administrative and survey data in assessing current programs, simulating new programs (e.g. the FSG program), or modifying the parameters of current programs (e.g. the parametric reforms to Ipelegeng). D. Institutional Arrangements 285. The Ministry of Local Government, as a key implementing agency, will need to further strengthen its capacity through developing operational systems for improved targeting, electronic payments of benefits, and better case management and monitoring. As the MLG’s front-line staff will be called to focus on social work/case management rather than spend much time on overseeing food baskets/food coupon distribution, they will need additional training and support. 286. Local authorities (district, city, and town councils) should continue to play a key role in the implementation and monitoring of social protection programs. They will need to be strengthened to perform these functions in a more effective manner. 153 GLOSSARY224 Absolute poverty lines. Poverty lines anchored in some absolute standard of what households should be able to count on in order to meet their basic needs. For monetary measures, these absolute poverty lines are often based on estimates of the cost of basic food needs, that is, the cost of a nutritional basket considered minimal for the health of a typical family, to which a provision is added for nonfood needs. Active labor market programs (or labor activation programs). Programs aimed at increasing the skills, employment, and long-run earning potential of participants through training, apprenticeships, job search assistance, subsidized job placements, and the like. Adequacy. A program is adequate if it provides sufficient benefits to enough people for long enough. Information on which to judge adequacy is usually provided in positive rather than normative terms—the transfer as a share of income of the recipients, the share of recipients in the population or among the poor, and so on. Administrative costs. All the costs required to deliver the transfers (and, in some cases, other related services). These activities include the identification of target population receiving and processing applications, dealing with appeals, processing payments, undertaking monitoring and evaluation, and exercising oversight over how program resources are used. Cash transfer programs. Programs that transfer cash to eligible people or households. Common variants include child allowances, social pensions, needs-based transfers, and conditional cash transfers. Categorical targeting. A targeting method in which all individuals in a specific category (for example, a particular age group, geographic location, gender, or demographic composition) are eligible to receive benefits. Conditional cash transfers. Provide money to poor families contingent on them making investments in human capital, such as keeping their children in school or taking them to health centers on a regular basis. 224 Based on Grosh et all (2008). 154 Demographic targeting. A targeting method in which eligibility is based on age. Disability. A physical, mental, or psychological condition that limits a person’s activities. The social model of disability emphasizes people’s ability to function in their particular physical and social environment. Disability therefore arises when barriers prevent people with functional limitations caused by age, disease, injury, or other causes from participating fully in society. Error of exclusion. The exclusion of a person who meets eligibility criteria from a program. Error of inclusion. The inclusion of a ineligible person in a program. Generosity. The level of a program benefit as a share of the poverty line or other type of indicator, such as the minimum wage, the average wage, or the total consumption of ben- eficiary households. Geographic targeting. A targeting method in which location determines eligibility for benefits or allocates budget to concentrate resources on poorer areas. In-kind food transfers. Provide additional resources to households by making food available when they need it most in the form of food rations, supplementary and school feeding programs, or emergency food distribution. Labor disincentives. Features of program design that discourage labor effort by potential beneficiaries (who may reduce work effort in order to qualify for a benefit) or actual beneficiaries (who may choose a different combination of labor and leisure once they have income from the program benefit). Last resort programs. Needs-based, usually means-tested, programs designed to help those who are not assisted, or not assisted enough to prevent poverty, by social insurance (pensions, unemployment insurance) or universal programs (child allowances, education, and the like). Leakage. In discussions of targeting, the leakage rate is the proportion of those who are reached by the program who are classified as non-poor (errors of inclusion). In discussions of accountability, the term is often used more broadly to include funds that, through 155 various forms of negligence or malfeasance, are diverted from legitimate (though possibly non-poor) beneficiaries to other uses. Management information system (MIS). Includes all the databases kept by the various program units in the performance of their functions—registry of beneficiaries, payments, and so on. Means test. A targeting method based on income that seeks to collect comprehensive information on household income and/or wealth and verifies the information collected against independent sources. Needs-based social assistance. Provide transfers for poor populations based on need. Non-contributory pensions (or social pensions). Benefits paid to the elderly from tax- financed (rather than contribution-financed) sources and without regard to past participation in the labor market. Orphans and vulnerable children. Orphans (children who have lost one or both parents) and other groups of children who are more exposed to risks than their peers such as children with HIV and those with sick caregivers, street children, children in institutions, child soldiers, child prostitutes, and others who are not cared for in a family setting or who are involved in the worst forms of child labor. Poverty gap. The mean difference between the poverty line and household income divided by the poverty line (the nonpoor have a gap of zero) calculated over the whole population. The income gap multiplied by the headcount equals the poverty gap. Poverty lines. Cutoff points separating the poor from the non-poor. They can be monetary (for example, a certain level of consumption) or nonmonetary (for instance, a certain level of literacy). The use of multiple lines can help in distinguishing among different levels of poverty. Proxy means test. A targeting method by which a score for applicant households is generated based on fairly easy-to-observe household characteristics, such as the location and quality of the household’s dwelling, ownership of durable goods, demog raphic structure, education, and so on. 156 Public works programs (or workfare). Where income support for the poor is given in the form of wages (in either cash or food) in exchange for work effort. These programs typically provide short-term employment at low wages for unskilled and semiskilled workers on labor-intensive projects such as road construction and maintenance, irrigation infrastructure, reforestation, and soil conservation. Generally seen as a means of providing income support to the poor in critical times rather than as a way of getting the unemployed back into the labor market. Quintile. One-fifth of an ordered population; for example, the poorest or richest one-tenth of the population. Risk management strategies. Strategies introduced by individuals, households, or communities dealing with risks that may temporarily or permanently affect their well-being. Ex ante strategies look to avoid the risk’s occurrence (prevention strategies) or, if this is not possible, to reduce its impact (mitigation strategies). Ex post strategies are aimed at dealing with the shock once it occurs (coping strategies). Safety nets. Non-contributory transfer programs targeted in some manner to the poor and those vulnerable to poverty and shocks. Analogous to the U.S. term “welfare” and the European term “social assistance.” Safety net system. A collection of programs, ideally well-designed and well-implemented, complementing each other as well as complementing other public or social policies. School feeding programs. In-kind food transfers that provide meals or snacks for children at school to encourage their enrollment and improve their nutritional status and ability to pay attention in class. Social pensions. Non-contributory pensions. Social insurance. Contributory programs designed to help households insure themselves against sudden reductions in income. Types of social insurance include publicly provided or mandated insurance against unemployment, old age (pensions), disability, the death of the main provider, and sickness. 157 Social protection. The set of public interventions aimed at supporting the poorer and more vulnerable members of society, as well as helping individuals, families, and communities manage risk. Social protection includes safety nets (social assistance), social insurance, labor market policies, social funds, and social services. Supplementary feeding programs. In-kind food programs intended to provide food to mothers and young children. Target group (or target population). The intended beneficiaries of program benefits. Targeting. The effort to focus resources among those most in need of them. Targeting assessment (or benefit incidence analysis). Describes how public spending is distributed across population groups, whether defined as deciles, poor versus non-poor, geographic areas, ethnic groups, and so on. Targeting errors. When program eligibility is based on imperfect information, program officials or the targeting rules they use may mistakenly identify non-poor people as poor or poor people as non-poor. When the former are admitted to a program, it is an error of inclusion; when the latter are denied access to the program, it is an error of exclusion. Targeting method. Approach taken to identify the target group and thus determine eligibility for program benefits. Vulnerable groups. Typically including the elderly, orphans, widows, people with disabilities, people with HIV/AIDS, refugees or internally displaced persons, among others. Vulnerable groups face special difficulties in supporting themselves because of some particular aspect of their situation. 158 REFERENCES Anderson, Stephen (2011). "Productive Safety Nets- Issues for MMAS’s Social Action Component." Paper commissioned by DFID UK, The Food Economy Group, June. BIDPA Briefing (2006). "Review of Existing Welfare Monitoring and Evaluation System. Final Report." Botswana Institute for Development Policy Analysis, Gaborone. December. BIDPA Briefing (2008). "Rising Global Food Prices: Causes and Implications for Botswana." September. Botswana Institute for Development Policy Analysis, Gaborone. December. BIDPA Briefing (2009). "Child Poverty and Inequalities in Botswana." Botswana Institute for Development Policy Analysis, Gaborone. BIDPA (2010). "Review of the NSPR: Social Protection for Poverty and Vulnerability Reduction." Botswana Institute for Development Policy Analysis, Gaborone. BIDPA (2012). "The Botswana School Feeding Program - A Case Study. Prepared for the Partnership for Child Development and Africa Union’s New Partnership for Africa Development." Botswana Institute for Development Policy Analysis, Gaborone. Bold, Chris, David Porteous, and Sarah Rotman. (2012). “Social Cash Transfers and Financial Inclusion: Evidence from Four Countries.” Focus Note 77. Washington DC: CGAP BPOPF (2012). "Trustee Report 2011/12." http://www.bpopf.co.bw.Botswana Public Officers Pension Fund, Gaborone Bundy, Donald, Carmen Burbano, Margaret Grosh, Aulo Gelli, Matthew Jukes and Lelsey Drake. (2009). "Rethinking School Feeding: Social Safety Nets, Child Development, and the Education Sector." World Food Programme, Rome and World Bank, Washington D.C. Coady, David, Margaret Grosh, and John Hoddinott. 2004. "Targeting Transfers in Developing Countries: Review of Lessons and Experience." World Bank and IFPRI, Washington D.C. 159 Commissioner of Labour and Social Security (2012). "Press Release. Minimum Wages for 2012/13." Gaborone. Coudouel, Aline, Jesko S. Hentschel, and Quentin T. Wodon (undated). "Chapter 1: Poverty Measurement and Analysis. Poverty Reduction Strategy Papers Sourcebook." World Bank, Washington D.C. http://siteresources.worldbank.org/INTPRS1/Resources/383606- 1205334112622/5467chap1.pdf CSO (2001). "Botswana Population Census 2001." Central Statistics Office, Gaborone. CSO (2002). "2002/03 Household Income Expenditure Survey." Central Statistics Office, Gaborone. CSO (2003). "Botswana Statistical Year Book 2003." Central Statistics Office, Gaborone. CSO (2004a). "Analytical Report 2001 Population and Housing Census." Central Statistics Office, Gaborone. CSO (2004b). "Household Income and Expenditure Survey 2002/03. Main Report. Volume I." Central Statistics Office, Gaborone. CSO (2005a). "2001 Population Census Atlas." Central Statistics Office, Gaborone. CSO (2005b). "Population Projections for Botswana 2001-2031." Central Statistics Office, Gaborone. CSO (2008). "Poverty Datum Line for Botswana 2003." Central Statistics Office, Gaborone. CSO (2009a). "2007 Botswana Family Health Survey IV Report." Central Statistics Office, Gaborone. CSO (2009b). "2008 Botswana AIDS Impact Survey III Statistical Report." Central Statistics Office, Gaborone. CSO (2009c). "Natural Disaster Digest 2008." Central Statistics Office, Gaborone. 160 CSO (2011). "Preliminary Results of the Botswana Core Welfare Indicators (Poverty) Survey 2009/2010. No. 2011/15. November." Central Statistics Office, Gaborone. CSO (2012). "National Accounts Statistics Report 2011." Central Statistics Office, Gaborone. DSS (2005). "Guidelines for Provision of Social Safety Net for Community Home-based Care Patients." Department of Social Services, Ministry of Local Government, Gaborone. DSS (2011). "Orphan Care Project Memorandum." Department of Social Services, Ministry of Local Government, Gaborone. de Ferranti, David, Guillermo E. Perry, Indermit S. Gill, and Luis Serven (2000). "Hacia la Seguridad Económica en la Era de la Globalización." World Bank, Washington D.C. Dembow, J. and C. Phenyo (2006). "Culture and Customs of Botswana." Greenwood Press, London. Dutta, Puja, Rinku Murgai, Martin Ravallion, and Dominique van der Walle. (2012). "Does India’s Employment Guarantee Scheme Guarantee Employment? Policy Research Working Paper, No. 6003." World Bank, Washington D.C. March. Ferrer, Ana and W. Craig Riddell (2009). "Unemployment Insurance Savings Accounts in Latin America: Overview and Assessment." World Bank Social Protection Discussion Paper, No. 0910. June. Fiszbein, Ariel and Norbet Schady (2009). "Conditional Cash Transfers. Reducing Present and Future Poverty." World Bank, Washington D.C. Garcia, Marito and Charity Moore (2012). "The Cash Dividend: The Rise of Cash Transfer Programs in Sub-Saharan Africa." World Bank, Washington D.C. Government of Botswana (2005). "Guidelines for the Provision of Social Safety Net for Community Home-based Care Patients." Government of Botswana, Gaborone. 161 Government of Botswana (2008). "National Situation Analysis on Orphans and Vulnerable Children in Botswana."Department of Social Services, Ministry of Local Government, Gaborone. Grosh, Margaret, Carlo del Ninno, Emil Tesliuc, Emil and Azedine Ouerghi (2008). "For Protection and Promotion: The Design and Implementation of Effective Safety Nets." World Bank, Washington D.C. Grosh, Margaret and Phillipe Leite (2009). "Defining Eligibility for Social Pensions: A View from a Social Assistance Perspective" in Robert Holzman, David Robalino, and Noriyuki Takayam, editors (2009). "Closing the Coverage Gap." World Bank, Washington D.C. Holzmann, Robert and Steen Jorgensen (2000). "Social Risk Management: A New Conceptual Framework and Beyond." World Bank, Washington D.C. February. Holzmann, Robert, Richard Paul Hinz, and Mark Dorfman (2008). "Pension Systems and Reform Conceptual Framework." World Bank Social Protection Discussion Paper No. 0824, June. Holzmann, Robert and Milan Vodopivec (eds) (2012). "Reforming Severance Pay: An International Perspective." World Bank, Washington D.C. ILO (2010). "World Social Security Report 2010/11. Providing coverage in times of crisis and beyond." International Labour Organization, Geneva. ILO (2011). "Republic of Botswana-Decent Work Country Program 2011-2015. February." International Labour Organization, Geneva. ILO (2012). "Social Security for All: The ILO Social Protection Floors Recommendation. Social Security for All Briefing Note. June." International Labour Organization, Geneva. IMF (2010). "Botswana-2009 Article IV Consultation. Country Report No. 10/172. June." International Monetary Fund, Washington D.C. IMF (2012a). "Boswana-2012 Article IV Consultation. Country Report No. 12/234. August." International Monetary Fund, Washington D.C. 162 IMF(2012b). "Botswana-2012 Article IV Consultation- Select Issues. Report No. 12/235. August." International Monetary Fund, Washington D.C. IOPS (2001). "IOPS- Country Profile- Botswana. International Organisation of Pension Supervisors." http://www.oecd.org/site/iops/research/46896420.pdf International Social Security Association. (2008). "Dynamic Social Security for Africa: An Agenda for Development. Developments and Trends." Geneva International Social Security Association (2011). "Africa: A New Balance for Social Security. Developments and Trends." Geneva.http://www.issa.int/Resursy/ISSA-Publications/Africa- A-new-balance-for-social-security Jalan, Jyotsna and Martin Ravallion (2003). "Estimating the Benefit Incidence of an Antipoverty Program by Propensity-Score Matching," Journal of Business & Economic Statistics, American Statistical Association, vol. 21(1), pages 19-30, January. Leite P., C. Costela and R. Quintana (2012) "Developing and Improving Social Safety Net Systems, mimeo." Lundberg, Mattias and Alice Wuermli Editors (2012). "Children and Youth in Crisis. Protecting and Promoting Human Development in Times of Economic Shocks. Directions in Development-Human Development." World Bank, Washington DC. Marques, José Silvério (2002). "Central America, Cross-Country Evaluation of Social Safety Net Assessments (SSNAs) - Issues Paper." Prepared for the World Bank, Washington D.C. Marques, José Silvério (2007). "Honduras-Social Protection Assessment." Background paper for the World Bank’s Public Expenditure Review, Washington D.C. Marques, José Silverio (2008). "Nicaragua-Social Protection Public Expenditure Review." Prepared for the Government of Nicaragua and the World Bank, Washington D.C. 163 Marques, José Silvério (2009). "Panamá- Evaluación de Programas de Asistencia Social y Propuesta de Reordenamiento y Racionalización. Preparado para o Ministerio de Desarrollo Social Programa de Protección Social Red de Oportunidades." Synthesis ConsultoresInternacionales, El Salvador. Marques, José Silvério (2010). "Belize-Issues and Options to Strengthen the Social Protection System- Policy Note." Prepared for Government of Belize and the World Bank, Washington D.C. Marques, José Silvério (2011a). "Jamaica- Social Protection Assessment." Prepared for the Government of Jamaica, May. Marques, José Silvério (2011b). "Antigua and Barbuda- Social Protection Assessment." Prepared for the Government of Antigua and Barbuda and the World Bank, Washington D.C. June. Marques, José Silvério (2011c). "Mozambique- Social Protection Assessment." Prepared for the Government of Mozambique and the World Bank. September. Marques, José Silvério (2013). "Sierra Leone- Social Protection Assessment." Draft report prepared for the Government of Sierra Leone and the World Bank. June. Ministry of Finance and Development Planning (2010). "National Development Plan 2010." Ministry of Finance and Development Planning, Gaborone. Miller, C., S. Gruskin, S. Subramanian, D. Rajaraman, and J. Heymann (2006). "Orphan Care in Botswana’s Working Households: Growing Responsibilities in the Absence of Adequate Support." American Journal of Public Health 96(8), 1429–1435. Ministry of Health (2011). "The Vulnerable Group Feeding Program-Guidelines for Health Facilities." Public Health Department. Nutrition and Food Control Division. June. Ministry of Local Government (2002). "Revised National Policy on Destitute Persons. July." Ministry of Local Government (2009). "Revised Remote Area Development Program (RADP). February." 164 Ministry of Local Government (2010a). "Social Development Policy Framework for Botswana. Phase I Situation Analysis. May." Ministry of Local Government (2010b). "Social Development Policy Framework for Botswana. Phase II Framework and Strategy. June." Ministry of Local Government (2010c). "Recommendations of the Report of the Task Force on the Review of the Impact of the Remote Area Development Program (RADP) on the Livelihoods of Remote Area Dwellers (RAD). September." Ministry of Local Government (undated). "Botswana National Plan of Action for Orphans and Vulnerable Children 2010-2016." Ministry of Finance of Namibia (2011). "2011/12 Citizens Guide to the National Budget." National Institute of Development Research and Documentation (1998). "Study of Poverty and Poverty Alleviation in Botswana, Volume 3: Stakeholders’ Views on the Conclusions and Recommendations of the Study Reports." Ministry of Finance and Development Planning, Republic of Botswana. Ngwenya, B. N. (2003). "Redefining Kin: Burial Societies in Botswana." Journal of Social Development in Africa, 18 (1), 85-110. Non-Bank Financial Institutions Regulatory Authority (2012). "2011 Annual Report." http://www.botswanaebooks.com/nbfira/nbfira2011annualreport/nbfira_2011_annual_rep ort.html Ntseane, D. and K. Solo (2007). "Social Security and Social Protection in Botswana." Bay Publishing, Gaborone. Okurut, N and B. Thuto (2009). "Informal Financial Markets in Botswana: A Case of Gaborone City." www.http://bidpa.bw/documents/informal. Accessed 11 October 2011. Oxford Policy Management (2010). "Evaluation of Retirement Systems of Countries within the Southern Africa Development Community. June." 165 Pallares-Miralles, Montserra, Carolina Romero, and Edward Whitehouse (2012). "International Patterns of Pension Provision II A Worldwide Overview of Facts and Figures. Social Protection and Labor Discussion Paper No. 1211, June."World Bank, Washington D.C. Ravallion, Martin. 1990. "Reaching the Poor through Rural Public Employment: A Survey of Theory and Evidence. World Bank Discussion Paper No. 94, July." World Bank, Washington D.C. Republic of Botswana (2008). "National Situation Analysis on Orphans and Vulnerable Children in Botswana." Department of Social Services, Ministry of Local Government, Gaborone. Republic of Botswana (2009). "2009/2010 Estimates of Expenditure from the Consolidated and Development Funds." Republic of Botswana, Gaborone. Republic of Botswana (2009). "2009 Budget Speech." http://www.finance.gov.bw, Gaborone. Republic of Botswana (2010). "2010/2011 Estimates of Expenditure from the Consolidated and Development Funds." Gaborone. Republic of Botswana (2010). "2010 Budget Speech." http://www.finance.gov.bw Gaborone. Republic of Botswana (2011a). "Vision 2016 Booket. Full Version. March." http://www.vision2016.co.bw/vision-publications.php?flag=pub Gaborone. Republic of Botswana (2011b). "2011/2012 Estimates of Expenditure from the Consolidated and Development Funds." Gaborone. Republic of Botswana (2012). "Poverty Eradication Guidelines. Implementation of Packages 2013." Gaborone. Republic of Botswana (2013). "2013 Budget Speech." http://www.finance.gov.bw. Gaborone. 166 Republic of Botswana (2013/14). "Estimates of Expenditure From the Consolidated & Development Fund." Gaborone. Republic of Botswana and United Nations (2010). "Botswana Millennium Development Goals-Status Report 2010." Gaborone. Republic of South Africa (2012). "Budget Review 2012." National Treasury, 22 February. Pretoria. Ribe, Helena, David A. Robalino, and Ian Walker (2010). "From Right to Reality-Achieving Effective Social Protection for All in Latin America and the Caribbean." World Bank, Washington D.C. Robalino, David, Laura Rawlings and Ian Walker (2012). "Building Social Protection and Labor Systems-Concepts and Operational Implications." Background Paper for the World Bank 2012–2022 Social Protection and Labor Strategy. Social Protection and Labor Discussion Paper No. 1202, March. World Bank, Washington D.C. Ruiz-Casares, M., and J. Heymann (2009). "Children Home Alone Unsupervised: Modelling Parental Decisions and Associated Factors in Botswana, Mexico, and Vietnam." Child Abuse and Neglect 33: 312–23. Schapera, I. (1970). "A Handbook of Tswana Law and Custom." Frank Cass, London. Seleka, T., H. Siphambe, D. Ntseane, N. Mbere, C. Kerapeletswe, and C. Sharp (2007). "Social Safety Nets in Botswana: Administration, Targeting and Sustainability." Lentswe La Lesedi (PTY), Gaborone. Social Welfare Division (2002). "Revised National Policy on Destitute Persons." Ministry of Local Government, Republic of Botswana, Gaborone. Subbarao, Kalanidhi, Carlo del Ninno, Colin Andrews and Claudia Rodríguez-Alas (2013). "Public Works as a Safety Net. Design, Evidence and Implementation," World Bank, Washington D.C. 167 Subbarao, Kalanidhi and Carlo del Ninno, Colin Andrews, Claudia Rodríguez-Alas (2010). "The Design and Implementation of Public Works Programs-A Toolkit for Practitioners. Social Protection Social Protection Discussion Paper Series, April." Human Development Network, World Bank, Washington D.C. Statistics Botswana (2011a). "Preliminary Results of the Botswana Core Welfare Indicators (Poverty) Survey 2009/2010." November. Gaborone. Statistics Botswana (2011b). "Botswana Statistical Yearbook 2010. December." Gaborone. Statistics Botswana (2012a). "March 2011 Formal Sector Employment. March." Gaborone. Statistics Botswana (2012b). "Stats Brief -Botswana-Maternal Mortality Ratio 2007-2011." Gaborone. Statistics Botswana (2013a). "Stats Brief Growth Domestic Product 4th Quarter of 2012." Gaborone. Statistics Botswana (2013b). "Primary Education Statistics 2012. February." Gaborone. Statistics Botswana (2013c). "Botswana Core Welfare Indicators Survey 2009/2010, 2013/11. July." Gaborone. Tesliuc Emil, M. Grosh, L. Pop, and R. Yemtsov (2013). "Income Support for the Poorest: A Review of Experience in Eastern Europe and Central Asia." World Bank, Washington D.C. UNDP (2005). "Poverty Status Report for Botswana: Incidence, Trends and Dynamics." International Poverty Centre, United Nations Development Programme, New York and Botswana Institute for Development Policy Analysis, Gaborone. UNICEF (2004). "The Framework for the Protection, Care, and Support of Orphans and Vulnerable Children Living in a World with HIV and AIDS." United Nation’s Children’s Fund, New York. http://www.unicef.org/aids/files/Framework_English.pdf UNICEF (2011). "Rapid Assessment of Vulnerable Households-Local - level Development. Policy Brief, April 13." United Nation’s Children’s Fund, New York. 168 UNICEF (2012). "Rapid Assessment of Vulnerable Households - Nutrition. Policy Brief, March 19." United Nation’s Children’s Fund, New York. UNICEF (2013). "Improving Child Nutrition-The Achievable Imperative for Global Progress. April." United Nation’s Children’s Fund, New York. UNICEF, UNAIDS, UNFPA, and WHO (2009). "Children and AIDS: Fourth Stocktaking Report, 2009." United Nation’s Children’s Fund, New York. United States Department of Labor (2011). "2011 Findings on the Worst Forms of Child Labor." Washington D.C. USAID (2010). "Assessing the Implementation of Botswana’s Program for Orphans and Vulnerable Children." United States Agency for International Development, Washington D.C. Vision 2016 (2013a). "Botswana’s Performance-A Report on the Progress being Achieved against the Vision 2016 goals." Vision Council. Gaborone. February. Vision 2016 (2013b). "Report on March 2013 RADs Smart Partnership Dialogue in Ghanzi." Vision Council. Gaborone. Wass, P. (1969). "A Case History: Community Development gets Established in Botswana." International Review of Community Development No 21/22. Weigand, Christine and Margaret Grosh (2008). "Levels and Patterns of Safety Net Spending in Developing and Transition Countries." Social Protection Discussion Paper No. 0817, World Bank. June. White, Philip and Stephen Devereux (2011). "Social Protection in Botswana—A Model for Africa? Frontiers of Social Protection Brief, Number 9, March." Regional Hunger and Vulnerability Project. www.wahenga.net WHO and World Bank (2011). "World Report on Disability." World Health Organization Geneva, and World Bank, Washington D.C. World Bank (2005). "World Development Report 2006." World Bank, Washington D.C. 169 World Bank (2008). "The Financing and Delivery of HIV/AIDS Prevention, Treatment, and Social Support Services in Botswana: An Economic Analysis." World Bank, Washington D.C. World Bank (2009). "Ethiopia: Third Phase of the Productive Safety Net Program. Report No. 48633-ET, September 25". World Bank, Washington D.C. World Bank (2010a). "The Social Safety Net in Botswana: Spending, Coverage, Challenges. Botswana Public Expenditure Review." World Bank, Washington D.C. World Bank (2010b). "Southern Africa HIV/AIDS and Tuberculosis Plan of Action, 2010-2011." World Bank, Washington D.C. World Bank (2010c). "Designing and Implementing a Rural Safety Net in a Low Income Setting Lessons Learned From Ethiopia’s Productive Safety Net Program 2005-2009." World Bank, Washington D.C. World Bank (2010d). "Strengthening Caribbean Pensions-Improving Equity and Sustainability." Human Development Sector Management Unit, Caribbean Country Management Unit, Latin America and the Caribbean Region, Report No. 47673-LAC, March. World Bank, Washington D.C. World Bank (2011). "Can Targeted Assistance Help the Very Poor? From Evidence to Policy: A Notes Series. No. 7." World Bank, Washington D.C. World Bank (2012a). "The World Bank’s Social Protection and Labor Strategy 2012-2022. Resilience, Equity and Opportunity." World Bank, Washington D.C. World Bank (2012b). "The World Bank’s Africa Social Protection and Labor Strategy 2012- 2022- Managing Risk, Promoting Growth: Developing Systems For Social Protection in Africa." World Bank, Washington D.C. World Bank (2012c). "Safety Nets and HIV/AIDS in Botswana, Namibia, and Swaziland. December." World Bank, Washington D.C. World Bank (2012d). "Raising Botswana’s Human Resource Profile to Facilitate Economic Diversification and Growth." Policy Note 1. World Bank, Washington D.C. 170 World Bank (2012e). "Botswana: Increasing the Role of the Safety Net to Improve Households’ Resilience in the Face of Crisis. Policy Note, November." World Bank, Washington D.C. World Bank (2013). "Social Safety Nets in Africa. A Review of the Experience in 21 Countries. April." Washington D.C. 171 ANNEX 1: TECHNICAL NOTE ON THE PROXY MEANS TEST AND ITS COMPARISON with Categorical Targeting Can a targeting scheme based on a proxy means test (PMT) improve the targeting the poor? To examine this, we compare a proxy means test targeting scheme to some of the categorical schemes already being used in Botswana. These regressions and simulations use data from the 2002-2003 Household Income and Expenditures Surveys (HIES) and can be quickly updated when the 2009-2010 Botswana Core Welfare Indicators Survey (BCWIS) data become available. We recognize that schemes targeting particular segments of the population (such as the elderly or orphans) have other objectives as well and should not necessarily be judged purely on poverty targeting criteria. However, given that the focus here is on reaching the poor, we will judge each scheme solely on how it performs in this respect, using information on actual poverty status from the 2002/03 HIES to simulate what would happen if we used these alternate schemes. I. Alternative Transfer Schemes Considered a) PMT: We constructed a proxy means indicator for poverty based on an OLS regression predicting per capita consumption from a set of household characteristics (proxies) that are correlated with poverty. The OMT model included demographic information (such as the dependency ratio in the household), asset information (such as having a caror livestock), housing quality (such as the number of rooms), and spatial information (the region of residence).225 Among those available in the survey data, we chose variables that we expected to be correlated with poverty, that could be collected through a short questionnaire, and that would not be easy to manipulate. The results of the PMT regressions are shown in Annex Table 1.5 at the end of this annex. Together with the PMT score, the choice of a cut off value for the PMT score determines how many households will be targeted by the PMT scheme. We examined two targeting schemes based on alternative cut off values: 225 The PMT was estimated separately for urban and rural areas. 172 I. PMT 1: The cutoff is set at the poverty line; that is, all households whose predicted per capita consumption is below the poverty line are targeted by the PMT. II. PMT 2: The cutoff is set at the 20th percentile of the PMT scores. In other words, only the poorest 20 percent of households (according to the PMT score) are targeted. For the simulations, we set the transfer at 50 pula per person per month based on the fact that the median poor person’s poverty gap (the distance of per capita consumption from poverty line) was about 45 pula in 2002/03. Thus, if perfectly targeted, this would at least halve the poverty headcount. As an alternative, we also considered a more generous transfer of 100 pula per capita per month. If perfectly targeted, this amount would be sufficient to bring 95 percent of the poor above the poverty line. b) Elderly Allowance: Every person above the age of 65 years receives a transfer of 150 pula per month. This amount is comparable to the actual Old Age Pension amount in 2007 in real terms. c) Universal Child Grant: Every child under 15 years of age is given a transfer of 100 pula per month. d) Orphan Grant: Every orphan under the age of 18 receives a transfer of 150 pula per month. This amount is roughly comparable with the value of the food basket provided under the Orphan Care Program in real terms. In our analysis of how the categorical grant schemes (b-d) would perform in reaching the poor, we assumed that the total grant amount per household was split evenly among all household members. AS ANNEX TABLE 1.1 SHOWS, POVERTY IN BOTSWANA IS CORRELATED WITH THE PRESENCE OF THE ELDERLY (PEOPLE AGED 65 AND OLDER), CHILDREN, OR ORPHANS IN THE HOUSEHOLD. FOR EXAMPLE, THE POVERTY RATE IN HOUSEHOLDS WITH ELDERLY MEMBERS IS 43 PERCENT AS OPPOSED TO 26 PERCENT IN OTHER HOUSEHOLDS. THESE CORRELATIONS IMPLY THAT TRANSFERS MADE TO HOUSEHOLDS WITH ELDERLY MEMBERS, CHILDREN, OR ORPHANS WILL BE PRO-POOR. THUS, SUCH CATEGORICAL TARGETING CAN HELP TO REACH THE POOR. 173 Annex Table 1.1: Having elderly members, children, or orphans in the household is correlated with poverty Poverty rate (percent) All (National) 30.6 In households with elderly members 43 In households with no elderly members 26 In households with children 34 In households with no children 13 In households with orphans 38 In households with no orphans 27 Source: 2002/03 HIES However, there are limits to the extent to which we can target poverty using the presence of elderly members, children, or orphans as a unique proxy indicator. Of all poor households, only about 35 percent live in households with elderly members (Annex Table 1.2). The same is true for households with orphans. Using either of these as the sole targeting criteria would mean reaching a maximum of only 35 percent of the poor. On the other hand, if we take all households with children as the target, we could reach 92 percent of the poor. However, this would be a costly way to target the poor because, since households with children are so numerous, 78 percent of the non-poor also live in such households. These statistics suggest that targeting can be improved by using a proxy index that incorporates information from multiple correlates of poverty. Annex Table 1.2: Distribution of the poor and non-poor across different types of households Percent of the poor living in households with elderly members 35 households with children 92 households with orphans 34 Percent of the non-poor living households with elderly members 20 in households with children 78 households with orphans 24 174 II. Key Findings of Targeting Accuracy Comparisons using Simulations Annex Table 1.3 describes the results of our simulation analysis of targeting accuracy using 2002/03 household survey data. Annex Table 1.3: Accuracy of alternative targeting scheme (urban + rural) SCHEME Coverage Coverage Inclusion Exclusion Total Error Effectiveness Total of Poor Error Error Ratio Spending Percent of Percent of Percent Percent of Percent Fraction of Million population the poor covered the poor population spending pula per covered covered by who are not incorrectly that goes month under scheme not poor covered targeted into scheme) reducing poverty gap PMT1 38.23 61.2 51.1 38.8 31.4 0.36 31.2 PMT2 20 32.4 50.6 67.6 30.8 0.37 16.3 Elderly 24.2 34.8 56.1 65.2 33.5 0.28 13.1 Allowance Orphan 27.1 33.9 61.2 66.1 37 0.23 19.2 Grant Child 82.4 92.5 65.6 7.5 56.4 0.25 62.5 Grant Source: Simulation analysis using 2002/03 HIES Overall, the analysis shows that the PMT is superior at targeting the poor. It has the lowest error rate, meaning it is least likely to misclassify a household’s poverty status. It also has the highest effectiveness ratio, meaning that, compared to the others, a larger share of the PMT spending would be spent on reducing poverty. Note that this is true for both PMT 1 and PMT 2 as they rank individuals the same way and differ only in terms of how many they cover. However, the simulations also suggest that a simple categorical scheme such as targeting the elderly might not be significantly worse than a PMT if coverage were limited to reaching only 20 percent or so of the population. The PMT2, which covers 20 percent of the population, would reach about 32 percent of the poor. The elderly allowance managed to reach 35 percent of the poor, at a slightly higher overall coverage rate of 24 percent. Even so, it is possible that the PMT2 is markedly better at reaching the absolute poor, but the simulations do not support this (as discussed later). 175 We also found that a grant targeting orphans looks similar to the elderly allowance in terms of its poverty targeting accuracy and the coverage indicators being considered. The elderly allowance performed slightly better, reaching more of the poor on a smaller budget. Hence, note that the statements below regarding the elderly allowance would also apply to the orphan grant. Not surprisingly, the universal child grant had a markedly different targeting profile from the other schemes. It has much broader coverage, and while it managed to reach most of the poor, by virtue of its wide coverage, it was not very accurate. Consequently, it would have limited effectiveness and would be a costly way to target the poor. In more detail, our key findings are: a) Coverage: The PMT1 scheme would cover about 38 percent the population, a rate higher than the actual poverty headcount of about 30 percent. Note that we can reduce this coverage rate by setting a lower cut off for the PMT score, as the coverage rate of PMT 2 shows. The elderly allowance covers a smaller fraction of the population than PMT 1, reflecting the fact that most households do not have elderly members. The child grant has the widest coverage as most household have eligible children. b) Coverage of the Poor and Exclusion Errors: PMT 1 would cover 61 percent of the poor. In other words, it would miss about 39 percent of the poor (exclusion errors). This is a much lower exclusion error than the PMT 2 and the elderly allowance, which would miss about 65 percent of the poor. The reason for the low coverage of the poor in the PMT 2 is that it covers about half as many people as the PMT 1, and not because it has more errors. The main reason for the low coverage of the poor in the elderly allowance is its overall low coverage: even though households with elderly members are poorer than other households, they constitute a small fraction of all households. The universal child grant covers 93 percent of the poor, but this reflects not targeting accuracy but its wide coverage of 82 percent of all households. c) Inclusion Errors: Inclusion errors (the percentage of people targeted who are not poor) are one indicator of the leakage of the grant, inasmuch as any transfer to a non-poor individual is considered a leakage. It is here that the disadvantage of the universal child grant becomes apparent. Being universal in coverage, about 65 percent of those covered by it are not poor. The PMT has the lowest inclusion errors. 176 d) Total Errors: A targeting error occurs when a poor household is not targeted (exclusion error) or when a non-poor person is targeted (inclusion error). This error rate is the lowest for the PMT and the highest for the universal child grant. e) Effectiveness: What fraction of the total spending in each scheme would contribute to a reduction in the poverty gap (defined as the amount needed to bring every poor person up to the level of the poverty line)? Reflecting its superior targeting accuracy, the PMT is the most effective. Even though the elderly allowance has a much lower error rate than the universal child grant, its effectiveness is not that much higher. This is because, even though it is better at identifying poor people, it gives many of them “too much” in the sense that they overshoot the poverty line. Similarly, if we made the PMT more generous by setting the transfer per head at 100 pula instead of 50, its effectiveness ratio would drop from 0.36 to 0.23 because it would be giving“too much” to many poor households. f) Total Spending (excluding administrative costs): Not surprisingly, the universal child grant would be the most expensive, but note that if we made PMT1 more generous by setting the transfer per head at 100 pula instead of 50, it would be just as expensive as the child grant. The PMT 2 and the elderly allowance have the lowest costs because of their low coverage rates. III. Accuracy by Poverty Level The PMT1 scheme had an error of about 35 percent in terms of the reaching the poorest 20 percent of individuals. Given a total coverage rate of 40 percent of the population, this suggests that the PMT model predicted consumption with substantial error. However, it was still the best at reaching the poorest households for a given total coverage rate. To understand how, consider the child grant. Since this would cover 82 percent of the population, it would be good at reaching the poorest, as any near-universal transfer would. However, this would come at a high cost as it had the highest error rates at other points in the distribution and targets most non-poor households as well (see Panel B in Annex Figure 1.1). 177 The targeting error rate of the PMT1 increased to over 50 percent when reaching the poorest 30 percent of individuals (the third deciles of consumption). Since these individuals are close to the poverty line, even relatively small errors in predicting consumption using proxies can lead to a misclassification. As expected, this misclassification became progressively less likely as we looked at richer individuals. 178 Even though the PMT1 and the PMT2 used the same proxy score and hence had the same magnitude of error in predicting individual consumption levels, the PMT2 missed twice as many of the poorest 20 percent due to its stricter cut off. Moreover, the categorical targeting scheme had error rates of above 60 percent for the poorest 20 percent of individuals. This suggests that, if reaching the absolute poor is a policy prerogative, we should prefer the PMT but should also consider raising the cut off for the PMT to above 20 percent while at the same time lowering the value of transfers per head to limit total costs. The PMT2 and the elderly allowance had roughly similar error profiles across the consumption distribution, even when considering the poorest 10 or 20 percent of households. However, the PMT 2 was better at avoiding rich households and hence was less wasteful. IV. Accuracy in Rural and Urban Areas Considering just rural areas, the PMT remains superior in terms of errors and effectiveness. Even though lowering the PMT cut off to just the bottom 20 percent (PMT2) would raise the exclusion rate of the poor to nearly 70 percent, the PMT continues to be substantially more effective than categorical targeting with less wastage. 179 Annex Table 1.4: Accuracy in rural and urban areas SCHEME Coverage Coverage Inclusion Exclusion Total Error Effectiveness Total of Poor Error Error Ratio Spending Percent of Percent of Percent Percent of Percent Fraction of Million population the poor covered the poor population spending pula per covered covered by who are not incorrectly that goes month under scheme not poor covered targeted) into scheme reducing poverty gap) RURAL PMT1 42.13 60.8 47.3 39.2 34.27 0.4 29.5 PMT2 20 30.4 44.5 69.6 34.3 0.42 12.6 Elderly 29.3 36.4 54.8 63.6 39.3 0.29 12.3 Allowance Orphan 29.7 35.2 56.9 64.8 40.6 0.26 16.3 Grant Child Grant 85.2 93.7 59.8 6.3 53.3 0.29 51.3 PMT1 42.13 60.8 47.3 39.2 34.27 0.4 29.5 URBAN PMT1 24.9 65.4 72 34.6 21.7 0.17 4.6 PMT2 20 55.3 71.2 44.7 18.9 0.18 3.7 Elderly 6.8 15.5 76 84.5 14 0.16 0.8 Allowance Orphan 18.1 19.0 89.1 81.0 24.5 0.07 2.8 Grant Child Grant 72.7 77.9 88.8 22.1 66.9 0.07 11.2 The universal child grant performs slightly better in rural areas as opposed to all areas. Because poverty rates are significantly higher in rural than urban areas, any scheme with wide coverage would end up reaching many poor individuals, even if it were ineffective at identifying the poor. For the same reason, the universal child grant has a huge error rate in urban areas, with only 0.07 of every pula spent on it contributing to reducing the poverty gap. Thus, even though it would cover more of the poor than the other schemes, it would be extremely wasteful as well. The PMT1 and PMT2 have similar coverage rates in urban areas. In general, because there are many more non-poor in urban areas, the PMT scheme ends up having high inclusion 180 errors, with about 70 percent of the targeted beneficiaries being non-poor. This high error rate drives its effectiveness ratio down in urban areas. Nonetheless, it is still the best overall, having much lower wastage (inclusion errors) than the universal child grant, while being better at covering the poor than the elderly allowance (reaching about 55 percent to 65 percent of the urban poor as opposed to only 16 percent for the latter). While the elderly allowance and the orphan grant look similar in rural areas, in urban areas, targeting orphans would lead to a markedly higher overall coverage rate. Since this would also mean a higher error rate given the lower incidence of poverty in urban areas, the orphan grant includes more of the non-poor and is less effective than the elderly allowance. Annex Table 1.5: PMT OLS regression results (1) (2) Dependent Variable: Per capita consumption Sample: RURAL URBAN 254.6*** 150.5* Male head of household (55.60) (77.03) 23.04** 32.87* Age of head (9.149) (18.79) -0.151* -0.138 Age squared (0.0900) (0.217) Number of children (0-14 years 300.3*** 594.9*** old) (40.40) (79.08) Number of elderly (65 plus years 232.8*** 699.6*** old) (81.10) (216.2) -1,310*** -2,561*** Dependency ratio (158.4) (297.2) -237.9*** -472.8*** Household size (22.28) (35.07) -354.0*** 65.82 Own home (82.52) (89.62) 529.8*** 448.1*** Van (104.5) (112.1) 680.9*** 848.7*** Car (121.0) (114.1) 360.6* -355.2 Tractor (213.5) (392.0) 107.6 261.2** TV (92.54) (106.9) 431.2** 507.2*** Microwave (177.3) (123.9) 308.9*** 427.1*** Fridge (88.41) (115.7) 11.68 105.8* Number of cattle (36.09) (59.70) 181 -98.22** -49.81 Number of goats (40.08) (71.65) 75.38 159.9 Number of sheep (74.33) (114.9) 10.42 31.05 Number of poultry (38.48) (67.04) 17.53 -175.8 Head never attended school (67.63) (131.2) Head secondary or above 309.6*** 370.2*** educated (76.46) (94.81) 5.380 28.91 House has piped waterx`x` (71.32) (106.1) -12.79 9.356 House has no toilet (70.45) (209.5) 64.33** 153.6*** Number of rooms (30.66) (48.85) Use wood, coal, paraffin, cow- -116.5 -81.07 dung charcoal or (77.63) (112.7) crop waste as cooking fuel Use wood, coal, paraffin, cow- -486.6*** -212.0* dung charcoal or (128.8) (109.0) crop waste as heating fuel 161.5 17.26 Use electricity for lighting (98.67) (121.8) Region Dummies Y Y Observations 3,205 2,818 R-squared 0.258 0.287 Notes: Standard errors in parentheses, *** p<0.01, ** p<0.05, * p<0.1. OLS regressions on household level data from HIES 2002/03 Variables are dummies unless their names indicate otherwise. For heating and cooking fuel variables the defaults (omitted dummies) are households using electricity, solar power or gas for cooking and heating, respectively. 182 183 Social Protection & Labor Discussion Paper Series Titles 2012-2014 No. Title 1415 Sudan Social Safety Net Assessment by Annika Kjellgren, Christina Jones-Pauly, Endashaw Tadesse and Andrea Vermehren, May 2014 1414 Tanzania Poverty, Growth, and Public Transfers: Options for a National Productive Safety Net Program by W. James Smith, September 2011 1413 Zambia: Using Social Safety Nets to Accelerate Poverty Reduction and Share Prosperity by Cornelia Tesliuc and W. James Smith, March 2013 1412 Mali Social Safety Nets by Cécile Cherrier, Carlo del Ninno and Setareh Razmara, January 2011 1411 Swaziland: Using Public Transfers to Reduce Extreme Poverty by Lorraine Blank, Emma Mistiaen and Jeanine Braithwaite, November 2012 1410 Togo: Towards a National Social Protection Policy and Strategy by Julie van Domelen, June 2012 1409 Lesotho: A Safety Net to End Extreme Poverty by W. James Smith, Emma Mistiaen, Melis Guven and Morabo Morojele, June 2013 1408 Mozambique Social Protection Assessment: Review of Social Assistance Programs and Social Protection Expenditures by Jose Silveiro Marques, October 2012 1407 Liberia: A Diagnostic of Social Protection by Andrea Borgarello, Laura Figazzolo and Emily Weedon, December 2011 1406 Sierra Leone Social Protection Assessment by José Silvério Marques, John Van Dyck, Suleiman Namara, Rita Costa and Sybil Bailor, June 2013 1405 Botswana Social Protection by Cornelia Tesliuc, José Silvério Marques, Lillian Mookodi, Jeanine Braithwaite, Siddarth Sharma and Dolly Ntseane, December 2013 1404 Cameroon Social Safety Nets by Carlo del Ninno and Kaleb Tamiru, June 2012 1403 Burkina Faso Social Safety Nets by Cécile Cherrier, Carlo del Ninno and Setareh Razmara, January 2011 1402 Social Insurance Reform in Jordan: Awareness and Perceptions of Employment Opportunities for Women by Stefanie Brodmann, Irene Jillson and Nahla Hassan, June 2014 1401 Social Assistance and Labor Market Programs in Latin America: Methodology and Key Findings from the Social Protection Database by Paula Cerutti, Anna Fruttero, Margaret Grosh, Silvana Kostenbaum, Maria Laura Oliveri, Claudia Rodriguez-Alas, Victoria Strokova, June 2014 1308 Youth Employment: A Human Development Agenda for the Next Decade by David Robalino, David Margolis, Friederike Rother, David Newhouse and Mattias Lundberg, June 2013 1307 Eligibility Thresholds for Minimum Living Guarantee Programs: International Practices and Implications for China by Nithin Umapathi, Dewen Wang and Philip O’Keefe, November 2013 1306 Tailoring Social Protection to Small Island Developing States: Lessons Learned from the Caribbean by Asha Williams, Timothy Cheston, Aline Coudouela and Ludovic Subran, August 2013 1305 Improving Payment Mechanisms in Cash-Based Safety Net Programs by Carlo del Ninno, Kalanidhi Subbarao, Annika Kjellgren and Rodrigo Quintana, August 2013 1304 The Nuts and Bolts of Designing and Implementing Training Programs in Developing Countries by Maddalena Honorati and Thomas P. McArdle, June 2013 1303 Designing and Implementing Unemployment Benefit Systems in Middle and Low Income Countries: Key Choices between Insurance and Savings Accounts by David A. Robalino and Michael Weber, May 2013 1302 Entrepreneurship Programs in Developing Countries: A Meta Regression Analysis by Yoonyoung Cho and Maddalena Honorati, April 2013 1301 Skilled Labor Flows: Lessons from the European Union by Martin Kahanec, February 2013 1220 Evaluating the Efficacy of Mass Media and Social Marketing Campaigns in Changing Consumer Financial Behavior by Florentina Mulaj and William Jack, November 2012 1219 Do Social Benefits Respond to Crises? Evidence from Europe & Central Asia During the Global Crisis by Aylin Isik-Dikmelik, November 2012 1218 Building Results Frameworks for Safety Nets Projects by Gloria M. Rubio, October 2012 1217 Pension Coverage in Latin America: Trends and Determinants by Rafael Rofman and Maria Laura Oliveri, June 2012 1216 Cash for Work in Sierra Leone: A Case Study on the Design and Implementation of a Safety Net in Response to a Crisis by Colin Andrews, Mirey Ovadiya, Christophe Ribes Ros and Quentin Wodon, November 2012 1215 Public Employment Services, and Activation Policies by Arvo Kuddo, May 2012 1214 Private Pension Systems: Cross-Country Investment Performance by Alberto R. Musalem and Ricardo Pasquini, May 2012 1213 Global Pension Systems and Their Reform: Worldwide Drivers, Trends, and Challenges by Robert Holzmann, May 2012 1212 Towards Smarter Worker Protection Systems: Improving Labor Regulations and Social Insurance Systems while Creating (Good) Jobs by David A. Robalino, Michael Weber, Arvo Kuddo, Friederike Rother, Aleksandra Posarac and Kwabena Otoo 1211 International Patterns of Pension Provision II: A Worldwide Overview of Facts and Figures by Montserrat Pallares-Miralles, Carolina Romero and Edward Whitehouse, June 2012 1210 Climate-Responsive Social Protection by Anne T. Kuriakose, Rasmus Heltberg, William Wiseman, Cecilia Costella, Rachel Cipryk and Sabine Cornelius, March 2012 1209 Social Protection in Low Income Countries and Fragile Situations: Challenges and Future Directions by Colin Andrews, Maitreyi Das, John Elder, Mirey Ovadiya and Giuseppe Zampaglione, March 2012 1208 World Bank Support for Pensions and Social Security by Mark Dorfman and Robert Palacios, March 2012 1207 Labor Markets in Middle and Low Income Countries: Trends and Implications for Social Protection and Labor Policies by Yoonyoung Cho, David Margolis, David Newhouse and David Robalino, March 2012 1206 Rules, Roles and Controls: Governance in Social Protection with an Application to Social Assistance by Lucy Bassett, Sara Giannozzi, Lucian Pop and Dena Ringold, March 2012 1205 Crisis Response in Social Protection by Federica Marzo and Hideki Mori, March 2012 1204 Improving Access to Jobs and Earnings Opportunities: The Role of Activation and Graduation Policies in Developing Countries by Rita Almeida, Juliana Arbelaez, Maddalena Honorati, Arvo Kuddo, Tanja Lohmann, Mirey Ovadiya, Lucian Pop, Maria Laura Sanchez Puerta and Michael Weber, March 2012 1203 Productive Role of Safety Nets by Harold Alderman and Ruslan Yemtsov, March 2012 1202 Building Social Protection and Labor Systems: Concepts and Operational Implications by David A. Robalino, Laura Rawlings and Ian Walker, March 2012 1201 MicroDeterminants of Informal Employment in the Middle East and North Africa Region by Diego F. Angel-Urdinola and Kimie Tanabe, January 2012 To view Social Protection Discussion papers published prior to 2012, please visit www.worldbank.org/sp Abstract The report lays out options to use social safety nets to meet the goal set by the Government of Botswana to eradicate extreme poverty by 2016 and in a budget neutral way. Introducing a Family Support Grant is argued to be a more efficient way to address absolute poverty by offering a family benefit to all the families in absolute poverty. Reforms in the Ipelegeng public works program would allow addressing existing shortcomings associated with coverage, targeting, generosity, and wrong labor incentives. The combination of the above will allow for a better weaving of the Social Safety Nets to ensure that the goal of lifting 16% of population from absolute poverty by 2016 is achieved. About this series... Social Protection & Labor Discussion Papers are published to communicate the results of The World Bank’s work to the development community with the least possible delay. The typescript manuscript of this paper therefore has not been prepared in accordance with the procedures appropriate to formally edited texts. The findings, interpretations, and conclusions expressed herein are those of the author(s), and do not necessarily reflect the views of the International Bank for Reconstruction and Development/The World Bank and its affiliated organizations, or those of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. For more information, please contact the Social Protection Advisory Service, The World Bank, 1818 H Street, N.W., Room MSN G8-803, Washington, D.C. 20433 USA. Telephone: (202) 458-5267, Fax: (202) 614-0471, E-mail: socialprotection@worldbank.org or visit the Social Protection website at www.worldbank.org/sp. Printed on recycled paper