32803 Africa Region Human Development Working Paper Series Notes on the Impact of the HIPC Initiative on Public Expenditures in Education and Health in African Countries Keith Hinchliffe Copyright June 2004 © Human Development Sector Africa Region The World Bank The findings, interpretations, and conclusions expressed in this report are entirely those of the author and should not be attributed in any manner to the World Bank, to its affiliated organizations, or to members of its Board of Executive Directors or the countries they represent. The World Bank does not guarantee the accuracy of the data included in this publication and accepts no responsibility for any consequence of their use. Cover design by Word Express Cover photo: World Bank Photolibrary NOTES ON THE IMPACT OF THE HIPC INITIATIVE ON PUBLIC EXPENDITURES IN EDUCATION AND HEALTH IN AFRICAN COUNTRIES Keith Hinchliffe Human Development Department Africa Region Acknowledgements This paper was prepared at the suggestion of Jee-Peng Tan who, together with Birger Fredriksen and Francis Rowe, gave useful feedback on a previous draft. Francis Rowe also provided access to a new database on priority sector spending in the Highly Indebted Poor Countries which is being constructed in the HIPC Unit of the World Bank. Thanks are also due to Elsie Lauretta Maka who has overseen the production of the paper. Financial support for this publication has been provided by the Norwegian Education Trust Fund. 2 Abbreviations and Acronyms GDP Gross Domestic Product HIPC Highly Indebted Poor Countries IDA International Development Agency IMF International Monetary Fund NGO Non government Organization PRSC Poverty Reduction Support Credit PRSP Poverty Reduction Strategy Paper 3 Foreword Over the past few years the World Bank, together with its partners, has evolved new processes and operational instruments in an effort to increase the resources of poor indebted governments and to improve the effectiveness of all domestic and donor- supplied resources in the fight against poverty. These include: a new set of procedures to allow higher levels of debt relief; the articulation by governments of Poverty Reduction Strategy Papers (PRSPs); and financial support to implement these strategies through Poverty Reduction Support Credits (PRSCs). Since human development is central to all poverty reduction efforts, these new approaches have tended to place a heavy emphasis on increasing financial resources for the social sectors, particularly education and health, and on improving sectoral performance. This paper is limited to an examination of the impact, so far, of the HIPC debt relief process on the levels and patterns of poverty- reducing expenditures in general and on expenditures in the education and health sectors in particular, in poor African countries. The Heavily Indebted Poor Countries (HIPC) Debt Initiative was originally launched in 1996 with the objective of significantly reducing the debt burden of eligible poor countries within a reasonable period of time, thereby eliminating a central constraint on economic growth. Over the following four years, public concern with excessive debt burdens, declining aid flows and a perceived over­stringency of the terms for receiving debt relief grew and, with the strong support of advocacy NGOs such as Jubilee 2000, this led to an `enhancement' of the Initiative in late 1999. The results were the promise of higher levels of debt service payment relief achievable in a shorter period of time, and a stronger link between that relief and increased poverty-reducing expenditures. In consequence, a centerpiece of the whole process was the Poverty Reduction Strategy Papers. These were intended to be prepared through highly consultative processes by national governments and used as a criteria for entry to the Initiative, as the basis for defining triggers for higher levels of relief and as the monitoring framework for additional support beyond debt relief. An earlier paper published in this Working Paper Series, Enhancing Human Development in the HIPC/PRSP Context, described the situation as of early 2001 focusing largely on the PRSPs and the nature of the triggers for debt relief as they affect the social sectors. It argued that for the process to be successful, the analytical knowledge base of the education and health sectors needed to be improved. The vehicle suggested for this was sectoral Country Status Reports, developed collaboratively. These Reports had been prepared for the health sector in three African countries and for the education sector in five. Since then another three have been completed for the health sector with another seven under way while a further four have been completed for the education sector with another four underway. Most of the Reports are being published in this Working Paper Series. At the time of this first review of the enhanced-HIPC/PRSC process, little additional debt relief had been granted but many countries were on the verge of receiving it. It is now over four years since the HIPC process was `enhanced' and over three years since the 4 `millennium rush' of countries to join the Initiative. Twenty seven countries are now `enrolled' in the HIPC process and have received some relief. Of these, 23 are African countries. The time is ripe for an updated review. This paper is a partial stock-take of the impact of the HIPC process on poverty-reducing expenditures in general, and on those in the education and health sectors in particular. The assessment is made through focusing on a series of ten questions which include: how large and significant is the debt relief; is it additional to other forms of external financial support; what have been the trends in poverty-reducing expenditures in general and in the education and health sectors in particular; how are proceeds of relief handled in government budgets and can they be tracked to specific expenditures; what activities within the education and health sectors has debt relief funded and can they be sustained; and how is the HIPC process affecting the efficiency and equity of total expenditures in the education and health sectors? The exercise reported in this paper, which focuses on expenditures, is intended as only a partial stocktaking of the full impact of the HIPC Initiative on the education and health sectors in poor, indebted African countries. A complete examination would focus on three areas; first, changes in sectoral expenditures, second the nature of the system-wide reforms which are integral to the Initiative and which aim to improve the efficiency and equity of all sectoral expenditures, and third the impact of these additional expenditures and of the reforms. It is my hope that as the new processes and instruments associated with debt relief and poverty reduction strategies are implemented and refined, future updates will focus more on the impact of increased expenditures and reforms and will provide useful lessons for future collaboration between the governments of poor countries and the World Bank and other development partners in support of poverty reduction. As a first step, the focus of this paper is on changes in sectoral expenditures with only a brief consideration of the influence of the HIPC process on improving the efficiency and equity of existing expenditures. Birger Fredriksen, Senior Education Advisor, Human Development Department, Africa Region 5 Summary and Conclusions The Heavily Indebted Poor Countries (HIPC) Initiative of 1996 was revised in 1999 due to political and public pressures arising from a perceived over-stringency in the terms for receiving DEBT relief, declining aid flows, and a campaign focused around the relative levels of debt payments and public expenditures on education and health. The framework was re-designed to increase access to debt relief and to strengthen the link between that relief and poverty-reducing expenditures with governments' Poverty Reduction Strategy Papers (PRSPs) as the centerpiece, and with a strong expectation that the education and health sectors would be major beneficiaries. This Note is a partial stocktaking of the impact of the HIPC Initiative on the education and health sectors. A complete examination would focus on three areas; first, changes in sectoral expenditures, second the nature of the system-wide reforms which are integral to the Initiative and which aim to improve the efficiency and equity of total expenditures, and third the impact of these additional expenditures and reforms. As a first step, the focus of this Note is on changes in sectoral expenditures, though some consideration is given to the specific role of the HIPC process on improving the efficiency and equity of existing expenditures. For consistency, all of the country specific data on education and health expenditures are taken from a variety of IMF publications and from a new expenditure database being constructed by the World Bank. Countries are eligible for HIPC relief if they are IDA countries and their debt is judged unsustainable on the basis of the debt stock/GDP, debt payments/exports and debt payments/government revenues ratios. They qualify for some relief (at the decision point) if they have prepared an interim or full PRSP and have a track record of macroeconomic stability. Out of 27 qualified countries, 23 are African and have been receiving interim relief. Seven of these African countries have reached the final stage of the process (the completion point) and are now receiving full relief, and another three are expected to reach this point in the immediate future. Nine African countries, mainly those which have had, or are still suffering from, civil unrest have not yet reached the decision point. Significance of the debt relief. Across the 23 African HIPCs the debt service/government revenue ratio fell from 24.2 percent in 1998 to 13.3 percent in 2003, and is projected to fall further to 10.4 percent in 2005. This is a considerable reduction which at least potentially frees up additional resources for poverty-reducing expenditures. However, within the averages there are considerable variations. In 2003, the debt service/government revenue ratio was below 10 percent in eight countries and above 20 percent in seven countries. Overall, the relief is significant. Total debt payments fell from US$2.6 billion in 1998 to US$1.9 billion in 2003. Simultaneously, expenditures on poverty-reducing programs (as defined in countries' PRSPs) increased from US$4.0 billion in 1999 to US$7.0 billion in 2003. This suggests at least the possibility that debt relief is having a catalytic effect on social sector expenditure. 6 Substitution by HIPC of other concessionary financial flows. It is possible that debt relief has increased at the expense of project and program grants and loans, and budget and balance of payments support. The World Bank's Operations Evaluation Department (OED, 2003) report suggested that in 2000 net transfers to HIPCs were no higher than in 1995 while those to non HIPCs remained lower. However, the IMF/IDA's most recent Implementation Report (September 2003) states that more recently net flows to HIPCs increased from US$4.0 billion in 1999 to US$7.0 billion in 2002 and gross flows from US$8 billion to US$12 billion. Again, there are significant variations in the experiences of different countries. In Uganda and Cameroon, the relief was very obviously additional while in Malawi and Zambia some may have been substituted for balance of payments support; and net flows declined in seven of the 23 African HIPCs. Trends in poverty-reducing expenditures. The composition of poverty-reducing expenditures is defined differently in the PRSP of each country. In some the definition covers two or three sectors while in others it covers up to 15 sectors, which is one reason why poverty-reducing expenditures are over 50 percent of total government expenditure in some countries and below 20 percent in others. In all countries these expenditures include at least some aspects of health and education; while it can also include water, social safety nets, agriculture, housing, employment and welfare, youth services, and so on. Over time, there has been a discernable trend to widen the definition. This has perhaps also been the reason for some shift in the nomenclature from `poverty-reducing' to `priority'. Expenditures on poverty-reducing expenditures have been increasing much faster than total government revenues. Overall, their share was 39 percent of the total in 1999 and 48 percent in 2003 (5.5 and 7.4 percent of GDP respectively). Of 20 African HIPCs the share of these poverty-reducing expenditures in total expenditures increased in thirteen, was constant in three and fell in four. Projections for 2005 indicate relatively little further increases in the shares. Importance of education and health in poverty-reducing expenditures. Poverty-reducing expenditures are not disaggregated in the IMF/IDAs reporting of the HIPC Initiative. Country-specific data collected for eleven African HIPCs for this Note, however, indicate that education and health expenditures dominate. In 2000, expenditures on the education activities of the packages of poverty-reducing expenditures averaged 57 percent of the total and expenditures on health were 20 percent. Two years later the share for education had fallen to 46 percent while the share for health had increased to 22 percent. The activities covered under `education' and `health' vary across countries. In a majority, `education' covers all of the sector, in some it covers primary and secondary, in some primary and in some just rural primary. The situation is similar for health. A separate database for thirteen African HIPCs covering all expenditures in education, health, social security, water and sanitation, housing, roads and agriculture from 1999 to 2002, indicates a percentage point decrease in the share of these expenditures for education and a percentage point increase for health. Trends in total education and health expenditures in African HIPCs. Much of the argument behind the public pressure to reform the processes of the HIPC Initiative in 1999 focused on comparisons made between debt service payments and expenditures on 7 education and health and on the argument that debt payments were crowding out social sector spending. In this Note an attempt has been made to collect information for each of the African HIPCs on the shares of education and health expenditure in total government expenditure and in GDP between 1998 and 2002. Education and health expenditure data for thirteen of the countries have been taken from a new World Bank expenditure database and information for another eight countries has been extracted from a variety of country-specific IMF sources. The evidence of higher shares of total government expenditures spent on education and health is very strong and, since government expenditures have been increasing faster than GDP, sectoral expenditures as a share of GDP have risen even faster. Overall, for the countries where data were available, education's median share of GDP increased between 1998 and 2002 from 3.0 to 4.2 percent and of total government expenditure from 12.8 to 15.5 percent. For health, the changes were from 1.0 to 1.7 percent and from 6.2 to 8.1 percent, respectively. In almost every African HIPC, expenditures on health have risen faster than overall government expenditure and GDP. Expenditures on education have risen faster than GDP in every country but one and faster than overall expenditures in two thirds of the countries. Overall, expenditures in education and health have increased their shares of GDP by 1.9 percentage points over the same period that the share of debt payments in GDP has fallen by 1.4 percent. HIPC Initiative relief and spending education and health. In the negotiations for debt relief, government have indicated that they would use roughly 40 percent of relief for education and 25 percent for health (HIPC Unit, 2003). It is difficult to verify that this has occurred. In most HIPCs, budgetary practices are complicated and the tracking of specific revenues and expenditures is particularly difficult. A European Commission study of five African HIPCs describes the diversity. In some countries the relief is placed in a separate account outside of the budget; in some it is held in a separate account within the budget and in the remainder it is merged into consolidated revenues. Each country has identified a set of priority sectors. Those countries where there has been an attempt to earmark and then account for the spending of debt relief have witnessed severe delays in disbursements. Overall, the report argues that all countries should shift to the Tanzanian model in which relief is amalgamated into consolidated revenues and the overall poverty- reducing expenditure program is effectively monitored. In such a case, the explicit relationship between debt relief and education and health expenditures is redundant. This is also the conclusion of the IMF/IDA report on HIPC tracking: "It is neither feasible nor desirable to track only HIPC funding. All public spending on poverty-reduction needs to be tracked." (IMF/IDA, 2001). As a consequence, the consolidated official reporting formats do not allow an explicit link to be made between the relief and sectoral expenditures. Preliminary comparisons of the total amount of debt relief with the increases in education and health expenditures within specific countries demonstrate no single pattern. In some countries, expenditures rose by a larger amount than the HIPC relief and in others by a smaller amount. Sustainability of incremental education and health expenditures. Savings from debt relief will not increase over time. At best, the Initiative will have encouraged a one time upward shift in education and health expenditures. An issue is the sustainability of such a 8 shift. The more that the relief is associated with, and seen as part of the funding for, total poverty-reducing expenditures and the more a government is committed to these expenditures, the greater the likelihood that any additional expenditures on education and health will be sustained. The easier it is to link debt relief to an individual program, the more likely that the proceeds of the relief will be switched to another program over time and that expenditures will be reduced when the relief finally stalls. However, debt relief for most countries will continue for many years. Impact of HIPC on the efficiency and equity of total sectoral expenditures. This Note focuses on changes in expenditures and has only a limited coverage of the impact of the Initiative on the composition of existing expenditures. In the HIPC process, the main mechanism for influencing existing resources in the education and health sectors is through the triggers set for countries to reach the completion point, and thereby receive full debt relief. The triggers are a set of structural (macroeconomic and budgetary) and sectoral performance benchmarks. Across countries there is a very wide variation in the benchmarks ­ from the requirement for a medium term expenditure framework and legislation on teacher employment to school mapping. Some imply more resources and some the better management of existing resources. The potential impact of the latter set is significant, and in the medium term may be a more important feature of the whole HIPC process in improving outcomes of the education and health sectors than the increase in resources from debt relief. Overview. As explained earlier, this Note focuses on only the first of three exercises which are required if the total impact of the HIPC Initiative on the education and health sectors in African countries is to be understood. The focus has been largely on changes in sectoral and poverty-reducing expenditures. To understand the full implications would require giving more attention to the effects which the process has also had on implementing sectoral reforms and on the ways in which total sectoral expenditures have been utilized; and then to the impact of both the increased resources and the reforms. The conclusion of the current analysis is that while there are significant variations between countries both in the financial benefits which have resulted from the HIPC Initiative and in commitments to increased poverty-reducing expenditures, overall the Initiative (through both the financial resources and the processes for gaining access to these) has had a positive effect on the funding for education and health in most of the highly indebted poor countries of Africa. 9 I. Introduction The Heavily Indebted Poor Countries (HIPC) Debt Initiative was originally launched in 1996 with the objective of reducing the debt burden of eligible poor countries by removing the debt overhang within a reasonable period of time, thereby eliminating a central constraint on economic growth. Over the following four years, with the strong support of advocacy NGOs such as Jubilee 2000, public concern with excessive debt burdens, declining aid flows and a perceived over­stringency of the terms for receiving debt relief grew and led to a review of the Initiative in late 1999. The results were the promise of higher levels of relief achievable in a shorter period of time and a broader set of objectives focusing on a permanent exit from debt rescheduling, the promotion of economic growth, and additional resources for social expenditures targeted at poverty reduction. The framework for the enhanced-Initiative was redesigned to provide a stronger link between the resources released from lower debt service payments and poverty-reducing expenditures. In consequence, a centerpiece of the whole process came to be the Poverty Reduction Strategy Papers. These were intended to be prepared through highly consultative processes by national governments, and used as a touchstone for entry to the Initiative, as the basis for defining triggers for higher levels of relief and as the monitoring framework for additional support beyond debt relief. It is now over four years since the HIPC process was `enhanced' and the procedures re-designed, and over three years since the `millennium rush' of countries to join the Initiative. Focus of the Note This Note is intended as only a partial stocktaking of the impact of the HIPC Initiative on the education and health sectors. A complete examination would focus on three areas; first, changes in sectoral expenditures, second the nature of the system-wide reforms which are integral to the Initiative and which aim to improve the efficiency and equity of total sectoral expenditures, and third the impact of these additional expenditures and reforms. As a first step, the focus of this Note is on changes in sectoral expenditures with only a brief consideration of the influence of the HIPC process on improving the efficiency and equity of existing expenditures. Twenty seven countries are currently `enrolled' in the HIPC process. Of these, 23 are African countries, and are the subject of this review. For consistency, all of the country- specific data on education and health expenditures are taken from a variety of country- specific IMF publications and from the first round of a new poverty expenditure database being prepared in the World Bank. While some relevant information is available for each of the 23 countries, it is somewhat limited for a small number of them. The review focuses on the following 10 questions: · How significant is the amount of debt relief in terms of total government expenditures in general and social expenditures in particular? · Have the resources from debt relief been additional to, or substituted for, other external official financial flows to HIPCs ? 10 · Has there been a substitution of aggregate donor support towards HIPCs and away from non-HIPCs? · What have been defined as poverty-reducing expenditures by different HIPCs, and what have been the trends in these expenditures compared to those for other types of expenditure? · Within poverty-reducing expenditures what has been the size and trend of education and health expenditures? · What have been the trends in overall education and health expenditure in African HIPCs? · Can any increases in education and health expenditures be linked directly to HIPC relief and how are the proceeds/savings from debt relief handled in government budgets? · In those countries where relief can be linked to the education and health sectors, on what items has it been spent? · Can measures be taken to ensure that the increased funding of those activities which have benefited from HIPC relief (and the HIPC process) will be sustained? · In what ways is the HIPC process, through the selection of items for the completion point, attempting to improve the effectiveness, efficiency, and equity of total expenditures in the education and health systems. HIPC process and coverage In this review, the processes of the HIPC Initiative are largely taken as understood and only a very brief summary of the stages and the coverage is provided.1 Countries are eligible for the HIPC initiative if they are IDA countries and would still have unsustainable levels of debt after full use of traditional debt relief mechanisms. Forty-two countries are eligible and eventually 38 are expected to qualify. Countries reach the decision point, the first stage of debt relief, based on a three year record of macroeconomic stability and the preparation of an interim or full Poverty Reduction Strategy Paper (PRSP). At that stage they begin to receive `interim' relief. Simultaneously, the criteria for the completion point are established. In addition to the maintenance of macroeconomic stability and the finalization of a full PRSP and its successful implementation for one year, performance benchmarks are set for structural and social reforms. On reaching the completion point, the remaining debt relief is scheduled and is irrevocable. To date, 27 countries, including 23 in Africa, have reached the decision point and are receiving some interim debt relief. Seven African countries 1For a fuller description see Operations Evaluation Department (2003), HIPC Unit (2003), IMF/IDA (2003) 11 have reached their completion point (Benin, Burkina Faso, Mali, Mauritania, Mozambique, Tanzania and Uganda) and another three (Niger, Ethiopia and Senegal) are expected to reach theirs during the first half of 2004. Of the remaining thirteen countries, several have encountered problems in moving towards their completion points and, at times, interim assistance has been temporarily suspended. Nine African countries, mainly those which have had, or are still suffering from, civil unrest have not yet reached the decision point. Currently the cut off date for this is December 2004. II. Ten Questions The central focus of this paper is to examine the impact of the HIPC Initiative in encouraging additional budgetary expenditures on poverty-reducing interventions in general and on education and health sectors in particular and, through the PRSPs, on increasing the efficiency and equity of existing expenditures. To aid this examination, a set of ten questions is posed. How significant is the amount of debt relief in terms of total government expenditures in general and social expenditures in particular? As of September 2003, a net present value of over US$31 billion of debt relief had been committed to the 27 countries that have reached the decision point. By 2005, when most countries are expected to have reached their completion points and be eligible for full relief, it is estimated that debt stocks will have fallen from US$77 billion to US$32 billion after traditional relief and assistance under HIPC, and to US$26 billion after the delivery of additional relief committed by several bilateral creditors. The debt sustainability indicators described in table 1 show the significant overall effects which have already resulted from the Initiative and those which are projected for 2005. By then, the debt indicators for the HIPCs are anticipated to be lower than the Table 1. Debt Indicators for HIPCs (percent, weighted average) Debt Before 2001 2002 Following the Indicators for sustainability enhanced completion non HIPCs indicators HIPC relief point 2001 NPV of debt to 274 275 214 128 143 export ratio NPV of debt to 61 65 50 30 39 GDP ratio Debt service to 16 10 10 8 15 exports Source: IMF/IDA, Status of Implementation , September 2003 p.8 indicators were for non-HIPC low income countries in 2001. Whether the ratios projected for 2005 are actually met will depend not only on whether countries meet the agreed requirements to trigger the completion point, and thereby become eligible for 12 further relief, but also on exports and GDP increasing at the rates which have been projected. A more directly relevant indicator of the debt burden for the purposes of this Note is the ratio of debt service paid to government revenues. It is this ratio, and comparisons with measures of social expenditures, which was at the center of the public campaign for a revised debt relief process in the late 1990s. Table 2 presents this ratio for each of the 23 African HIPCs from 1998 to 2003, with projections to 2005. Averaging across all 23 Table 2. Ratio of Debt Service to Government Revenues. African HIPCs. 1998 to 2005 (percent) 1998 1999 2000 2001 2002 2003 2004 2005 Preliminary Projected Projected Benin 17.1 17.3 14.6 9.4 8.0 7.6 5.4 5.3 Burkina 15.7 15.8 18.6 11.4 11.8 9.8 6.7 6.0 Faso Cameroon 28.0 24.1 26.3 14.9 11.7 12.0 11.5 12.2 Chad 20.0 24.0 29.4 11.4 20.4 22.9 13.5 9.9 DRC 0.4 1.4 0.0 0.0 7.4 25.0 17.8 18.6 Ethiopía 8.6 11.0 10.2 15.5 8.9 6.4 7.6 5.8 Gambia 12.4 25.5 29.2 31.3 21.9 30.7 28.2 19.8 Ghana 41.0 53.4 78.1 25.7 39.1 17.7 9.9 7.9 Guinea 33.5 35.3 45.5 22.2 22.0 23.6 24.4 18.4 Guinea 62.6 15.5 31.6 1.2 7.1 13.1 11.7 9.6 Bissau Madagascar 41.7 25.0 14.4 9.7 13.8 9.5 10.3 9.4 Malawi 21.9 20.5 32.5 29.4 18.1 30.5 33.4 13.7 Mali 17.4 19.6 20.3 12.7 11.9 9.3 7.8 7.7 Mauritania 35.0 30.4 36.1 36.6 19.9 17.8 18.9 16.7 Mozambique 23.2 12.3 4.1 6.7 8.3 7.1 7.3 7.4 Níger 9.0 10.6 14.3 19.1 21.5 10.0 6.5 6.5 Rwanda 16.0 23.0 23.4 6.2 7.0 6.1 7.9 8.0 Sao Tome 84.1 21.4 38.7 46.4 39.2 24.1 67.5 10.7 and Principe Senegal 26.0 22.0 21.4 20.6 19.0 18.4 15.5 6.8 Sierra Leone 18.2 77.4 76.2 90.3 18.4 16.1 17.9 9.9 Tanzania 29.0 19.8 16.1 9.1 10.2 8.9 10.0 10.1 Uganda 16.0 12.9 14.2 11.5 9.3 14.0 15.8 16.9 Zambia 24.3 22.9 18.7 23.6 20.0 24.4 28.6 23.4 Simple 24.2 22.7 23.6 16.4 14.1 13.3 12.2 10.4 average Total debt 2633 2451 2254 1774 1719 1938 service paid (US$mill.) Total debt 2050 1915 service due (US$mill.) Source: reformulated from IMF/IDA 2004, Annex table 10.B Note: 1998-2002 actual; 2003 preliminary; 2004-2005 projected 13 countries, the share of government revenues preempted for debt servicing fell significantly from 24.2 percent in 1998 to 13.3 percent in 2003. By 2005, it is projected to have fallen to 10.4 percent. Not surprisingly, however, there are large variations across countries. Focusing on the most recent ratios, for 2003, the range is from just 6.1 percent in Rwanda to around 30 percent in the Gambia and Malawi. In all, the ratio is below 10 percent in eight countries, between 10 percent and 19 percent in another eight countries, and 20 percent and over in seven countries. The total amount of debt service paid has fallen by over a quarter from a total of US$2633 million in 1998 to US$1938 million in 2002. In the period to 2005, nominal payments will decrease further in 14 of the 23 countries while overall remaining relatively constant. The data in tables 1 and 2 indicate that the HIPC process is a major intervention with the potential of improving the fiscal situation of most concerned countries. In 1998, roughly one fourth of government revenues were required to service debt with the corollary that they could not be used for programs of public expenditure. Five years later the share had fallen to almost one eigth, and actual payments were 26 percent lower. A central question of this Note is whether this development has affected expenditures in health and education. Initially, some more aggregated figures are provided. These cover poverty reducing expenditures as defined by the IMF and IDA for their reviews of the HIPC Initiative. The definitions, in turn, are based on priority sectors as defined by individual governments in their Poverty Reduction Strategy Papers. The detailed components vary quite considerably across countries, though they are reasonably consistent over time within each country. Poverty-reducing expenditures in the African HIPCs increased from US$4.3 billion in 1999 to a preliminary estimate of US$7.0 billion in 2003 (IMF/IDA, 2004). This increase can be compared to a reduction in debt payments from $2.4 billion and $1.9 billion in debt payments over the same years. This set of initial comparisons indicates that while the conclusion cannot simply be reached that lower debt payments have caused higher social expenditures, the relationship is worth further exploration. Have the resources from debt relief been additional to, or have they substituted for, other external official financial flows to HIPC countries ? While the HIPC Initiative has undoubtedly reduced the level of debt servicing and, all things being equal, thereby allowed for a higher share of government internally generated revenues to be available for spending programs, it is possible that the relief provided by multilateral and bilateral donors has been at the expense of project and program related loans and grants, including balance of payments and budget support ­ that is, a substitution of development financing by debt relief. An independent review of this issue made within the World Bank suggested that over the period 1995-2000, such a substitution did occur (OED, 2003). For HIPCs, there was a strong decline in net transfers (gross official resource flows minus debt payments) between 1995 and 1997 followed by an increase to 1999 and a fall in 2000 back to the 1997 level. Non-HIPC low income countries suffered a far larger fall. However, the IMF/IDA Status of Implementation report for 2003 suggests that, more recently, across the 27 countries which have passed the decision point, aggregate debt relief does not appear to have been offered at the expense of grants and loans. The overall totals of external loans and grants 14 plus HIPC debt relief (gross resources) increased from about US$8 billion in 1997 to almost US$12 billion in 2002. Debt relief accounted for half of the increase and additional grants and loans for the rest. Similarly, net flows are described as having been virtually stable between 1997 and 1999 at around US$4 billion but to have then increased to over US$7 billion by 2002. Once again, aggregate trends do not reflect the situation for each country. The OED report describes how, for instance, HIPC relief in Uganda and Cameroon had been additional as donors continued to provide other forms of support while in both Malawi and Zambia, official loans and grants declined considerably. Even the more positive 2003 Status of Implementation Report describes external financial flows in 2002 as having declined in seven of the 23 African HIPCs. Guinea-Bissau, Malawi, Sao Tome and Principe, and Senegal saw interruptions to their (IMF) Poverty Reduction and Growth Facility programs as agreements were not met, while Mali and Mauritania experienced serious delays in disbursements. A reduction in Rwanda partly reflected the very large increase in external support at the beginning of the period. Another reason why debt relief savings do not automatically lead to a net transfer of resources is that in the case where a country is in default, the relief is simply reflected in a clearing of arrears. This is at least partially the case in Guinea-Bissau and Zambia. Has there been a substitution of aggregate donor support towards HIPCs away from non HIPC countries? External official financing fell significantly in the mid 1990s. Since 1998 there appears to have been a redistribution of available resources to the HIPCs. While the HIPC initiative may be resulting in additionality for the HIPCs and for all low income countries taken together, low income non-HIPCs (such as several countries in South Asia) have benefited little, if at all, from the overall increase in gross and net flows. While this issue is not one focused on in this report, the possible inequity is apparent. What have been defined as poverty reducing expenditures by different HIPCs and what have been the trends in these compared to other types of expenditure? Implementation Status reviews of the HIPC process use the term `poverty-reducing' expenditures as a measure of the expenditure shifting and enhancement effects of debt relief. What is defined as `poverty reducing' for each country reflects what is defined as such by governments in their PRSP. As a consequence, the definition differs across countries. Of the current 23 African HIPCs, five include just education and health (Benin, Cameroon, Sao Tome, Senegal and Niger). Others include water (for instance, Madagascar), social safety nets (Mali), agriculture (Ethiopia and the Gambia); and several countries include a much wider spread of activities. Examples are Zambia - housing, resettlement, urban development, health, information services, employment and welfare, education and training, and general social expenditure; and Rwanda ­ specified programs in internal affairs, agriculture, commerce, education, youth and sports, health, transport and communication, energy and water resources, gender, public service, lands 15 and settlement, and support to local government. Recently, there has been a trend to widen the definition, and a new nomenclature of `priority sectors' has begun to emerge. For instance, Tanzania now includes eight activities under `priority sectors' ­ education , health, water, agricultural extension, lands, roads, judiciary and AIDS. Table 3 describes the size of poverty reducing expenditures in aggregate and as shares of government revenues (minus grants) and GDP from 1999 to 2003, with projections to 2005, for the 23 African HIPCs as a whole. Table 3. Poverty-reducing Expenditures in 23 African HIPCs. 1999 to 2005. (US$ million.) ___________________________________________________________________________________ 1999 2000 2001 2002 2003 2004 2005 ___________________________________________________________________________________ Poverty reducing-expenditures 4267 4085 4742 5547 7034 8595 9108 - as percent government revenue 38.6 39.7 43.8 45.6 48.1 51.2 49.2 - as percent GDP 5.5 5.5 6.1 6.7 7.4 8.2 8.1 ___________________________________________________________________________________ IMF/IDA (2004) table 11A. Overall, there was little change in the level of poverty-reducing expenditures between 1999 and 2000 but in the following three years they are reported to have increased by around 16, 17 and 27 percent. These increases have been far higher than the increases in government revenue and GDP. As a result, poverty-reducing expenditures as a share of total revenues increased from around 39 percent to 48 percent between 1999 and 2003, and as a share of GDP from 5.5 percent to 7.4 percent. Projections of expenditures, revenues and GDP to 2005 suggest that the expenditures will continue to rise (by almost 30 percent), and their share of GDP will increase to around 8.1 percent. As a share of government revenues, however, there will be little change. These aggregates again hide wide variations across countries. However, because of the differences in defining the composition of poverty-reducing expenditure, there is little value in comparing the expenditure shares across countries. There is more value, however, in looking at trends within countries. Poverty-reducing expenditures as shares of total government revenues are presented for twenty countries from 1999 to 2003 in table 4, and for another three countries for selected years, with projections for all countries for 2005. Of the twenty countries, significant increases in the share of total expenditure directed towards poverty-reducing activities were evident in thirteen. Exceptions were Benin, Madagascar and Niger where the shares remained roughly constant and Ghana, Zambia, Chad and Sao Tome and Principe where they fell. Calculating the percentage point changes between the two years the variations are very wide with an average change of around 14 (column 7). This average, however, includes some very large apparent increases and decreases. The median increase is just over five percentage points ­ still a significant increase. According to projections, in aggregate two thirds of the potential increases in the share of poverty-related expenditures have already been delivered. By 2005, the assumed year of HIPC completion for most countries, the 16 share is expected to be lower in nine of the 23 countries than in 2003, slightly higher in most of the rest and significantly higher in Ethiopia and Senegal. Table 4. Poverty-reducing Expenditures by Individual African HIPCs as Shares of Total Government Revenue Actual Actual Actual Actual Estimate % point change Projection 1999 2000 2001 2002 2003 1999-2003 2005 (1) (2) (3) (4) (5) (6) (7) (8) Benin 30.0 29.4 41.7 35.4 31.0 1.0 36.5 Burkina Faso 29.6 31.9 43.5 42.0 42.7 13.1 37.6 Cameroon 15.8 17.3 19.2 19.2 21.1 5.3 21.7 Chad 149.8 130.9 153.1 129.0 120.7 -27.1 88.0 D.R. Congo .... .... ... 5.7 12.0 .... 16.4 Ethiopía 61.0 45.9 60.0 72.6 77.0 16.0 86.9 Gambia 30.2 29.7 41.4 42.1 37.3 7.1 29.6 Ghana 35.3 27.7 25.1 26.6 31.9 -3.4 29.6 Guinea 22.8 25.3 30.3 27.1 28.4 5.6 27.3 Guinea-Bissau .... .... .... 20.3 18.6 .... 19.4 Madagascar 36.7 27.9 35.2 47.1 38.3 1.6 42.1 Malawi 66.1 57.1 48.0 63.8 55.6 10.5 54.4 Mali 24.3 39.5 31.3 33.0 38.2 13.9 34.4 Mauritania 34.7 28.9 38.8 28.2 41.0 5.3 46.1 Mozambique 52.8 143.3 145.5 129.9 129.7 76.9 111.5 Níger 58.2 53.9 53.1 47.9 60.5 2.3 64.2 Rwanda 39.5 45.4 48.1 53.7 52.1 12.6 53.9 Sao Tome 87.8 77.9 117.9 83.3 82.2 -5.3 70.7 Senegal 30.8 28.5 30.0 29.2 34.5 4.5 44.5 Sierra Leone ...... 32.3 35.0 50.6 47.8 ..... 49.8 Tanzania 42.8 54.6 57.3 74.0 83.2 40.4 88.3 Uganda 40.4 62.6 71.3 79.3 100.6 60.2 91.4 Zambia 30.1 20.1 11.0 14.6 13.2 -16.9 18.4 Source: Based on IMF/IDA (2004) table 11B. Note: poverty-reducing expenditures of over 100 percent of government revenues indicate very high levels of external grant assistance Comparing real per capita expenditures in priority sectors in HIPCs and in five non-HIPC low income countries (Angola, Kenya, Lesotho, Pakistan and Sri Lanka), Rowe (2004) concludes that the gap which existed in favor of the non-HIPCs in 1997 had been eliminated by 2001. Within `poverty-reducing expenditures', what has been the size and trend of education and health expenditures? Estimates of expenditures on education and health made in African HIPC's in recent years are presented in the following section. However, in table 5, the shares of total expenditures defined as `poverty-reducing' which have been allocated to the education and health sectors are presented for eleven countries. Most of the data relate to the period 2000-2002, but are available for a longer period in some countries. It is clear that education and health expenditures dominate the set of poverty-reducing expenditures in each of these countries. In 2000 education received an average of 57 percent of the total 17 and health received 20 percent. It is also clear, though, that the share for education is tending to fall (in eight of the eleven countries) and the average for 2002 had fallen to 46 percent, while the share for health slightly increased to 22 percent. What is covered by `education' and `health' varies across countries and the definition is not always readily obvious in the sources of data used for this report. In several cases, it appears that the whole of the education and health systems is included, which is why `poverty-related' expenditures can be over half of all government expenditure (table 4 above). There are exceptions, however: Ghana includes only basic education and primary health; Guinea excludes higher education; Malawi excludes universities in education and tertiary institutions in health; the education program in Mali covers only community teachers and teaching materials in rural areas and health sector expenditures cover specific programs such as HIV/AID and malaria; in Niger rural education and health are specified; and in Uganda the Poverty Action Fund covers primary education and primary health. Table 5. Education and Health Expenditures as Shares of Total Poverty-related Expenditures 1998- 2003. Selected HIPCs. 1998 1999 2000 2001 2002 2003 Benin Education 27 Health 31 Burkina Faso Education 48 48 52 48 47 Health 46 48 39 42 35 Ethiopia Education 56 57 57 57 Health 18 17 17 16 Ghana Education 62 61 42 Health 16 11 13 Malawi Education 67 51 53 Health 18 22 15 Mali Education 45 45 Health 15 15 Mauritania Education 60 56 52 51 Health 22 23 27 30 Mozambique Education 26 29 36 26 Health 22 19 16 22 Rwanda Education 80 81 80 66 Health 15 13 15 13 Tanzania Education 46 52 51 45 43 39 Health 24 19 20 18 18 19 Uganda Education 73 70 60 49 46 42 Health 9 7 13 17 20 20 Source: IMF country reports Note: the components of poverty-related expenditures are defined in country PRSPs 18 A new database in the process of being constructed in the World Bank shows similar, though less dramatic, trends (Rowe, 2004). This database currently includes thirteen African HIPCs and measures total expenditures on education, health, social security, water supply and sanitation, housing, roads and agriculture. Of the combined expenditures across these common sectors, the median share for education fell slightly between 1999 and 2002 from 39.6 percent to 38.3 percent, while for health the share increased slightly from 17.5 percent to 18.6 percent. What have been the trends in education and health expenditures in African HIPCs? As described in the previous section, in the anti-poverty programs which have been implemented by governments in many African countries over the past few years, and which have mainly evolved from the PRSPs, expenditures on education and health have played a significant part. An important question is the extent to which this, and other influences which have highlighted the role of these sectors, has affected overall public expenditures on them and led to a relatively greater allocation of government resources, and a greater share of GDP. To answer this question, an attempt has been made to assemble the relevant data for each of the 23 African HIPCs. For a set of thirteen countries, sectoral expenditures were taken from the new World Bank priority sector spending database described above. Additional countries were added using data from recent IMF country reports. In all, estimates of education and health actual expenditures as shares of total government expenditures and of GDP between 1998 and 2002 are available for between 18 and 21 countries. No data series are presented for the Democratic Republic of Congo or for Sao Tome and Principe and only a limited amount of data is available for Guinea-Bissau, Chad and Niger. The results are described in tables 6-9 below.2 2The importance of the advocacy NGOs in adding to the pressure to reform the HIPC process in 1999, their use of comparisons of debt service payments compared to health and education expenditures and their continuing effort to argue for still wider forms of debt relief makes a recent paper by Jubilee Research (a successor of Jubilee 2000) documenting the trends in debt relief and health and education expenditures particularly interesting (Jubilee Research, 2002). Data (sources not provided) were presented for ten African HIPCs for the period 1998 to 2002. The results are striking. For these countries, debt service payments are reported to have fallen from US$ 979 million in 1998 to US$ 620 million in 2002, while education expenditures increased from US$ 929 million to US$ 1306 million and health expenditures increased from US$466 million to US$ 796 million. In other words while debt service decreased by 36.7 percent, education expenditures increased by 40.6 percent and health expenditures by 70.8 percent, with a combined increase of 50.8 percent. 19 Table 6. Education Expenditure as Share of Total Government Expenditure 1998-2002. 21 countries 1998 1999 2000 2001 2002 Benin 16.8 16.8 16.1 19.5 19.7 Burkina Faso 10.4 6.6 7.1 7.6 9.1 Cameroon 12.7 14.9 16.1 18.8 17.2 Chad 9.7 9.6 9.5 11.7 9.1 Ethiopía 13.8 11.3 9.6 13.2 12.9 Ghana 12.3 14.5 14.8 13.3 18.4 Gambia 23.2 20.1 22.6 23.4 Guinea 15.6 15.4 14.2 13.3 12.6 Guinea-Bissau 9.5 10.6 Madagascar 5.2 5.3 4.7 5.6 7.5 Malawi 15.3 12.7 12.1 13.9 13.6 Mali 12.8 12.9 13.4 13.1 11.3 Mauritania 20.0 21.1 15.0 17.0 17.6 Mozambique 11.5 11.4 18.4 13.9 9.9 Níger 12.9 15.6 15.9 Rwanda 11.8 17.2 18.2 16.5 15.5 Senegal 19.2 18.7 17.3 18.2 22.2 Sierra Leone 18.1 17.3 15.9 17.0 Tanzania 21.1 20.3 22.4 24.4 Uganda 26.9 26.3 24.9 24.1 24.8 Zambia 8.1 7.5 7.6 8.0 10.1 Median 12.8 14.9 14.9 13.9 15.5 Table 7. Education Expenditure as Share of GDP 1998-2002. 19 countries. 1998 1999 2000 2001 2002 Benin 3.0 3.1 3.2 4.1 3.7 Burkina Faso 2.6 1.9 1.8 1.9 2.2 Cameroon 2.4 2.6 3.0 3.4 3.3 Chad 1.5 1.8 1.9 2.2 1.9 Ethiopía 3.4 3.5 3.2 4.1 4.5 Ghana 3.5 3.8 4.1 4.4 4.9 Gambia 5.4 4.4 5.5 5.6 Guinea 1.4 1.4 1.2 1.2 1.5 Madagascar 1.0 0.9 0.8 1.0 1.1 Malawi 4.5 3.7 3.8 4.4 4.5 Mali 3.1 3.2 3.2 3.2 2.8 Mauritania 4.4 5.5 4.6 4.3 5.5 Mozambique 2.9 3.2 5.5 6.0 3.8 Rwanda 2.2 3.4 3.4 3.4 3.6 Senegal 3.9 4.0 3.4 4.0 4.8 Sierra Leone 4.0 5.0 5.1 5.5 Tanzania 3.0 3.3 3.7 4.2 Uganda 3.4 3.7 3.7 4.3 4.4 Zambia 2.3 2.1 2.3 2.6 3.2 Median 3.0 3.2 3.4 4.0 4.2 Source (table 6 and 7): education expenditures taken from World Bank Staff (HIPC Unit) for Burkina Faso, Ethiopia, Ghana, Madagascar, Malawi, Mali, Mauritania, Mozambique, Rwanda, Senegal, Sierra Leone, Tanzania, Zambia. For all other countries, IMF country reports. 20 Table 8. Health Expenditure as Share of Total Government Expenditure. 1998-2002. 20 countries 1998 1999 2000 2001 2002 Benin 6.5 8.3 7.2 8.8 8.1 Burkina Faso 9.8 5.4 5.6 5.9 8.1 Cameroon 3.2 3.4 4.8 5.5 7.8 Ethiopía 5.8 4.3 3.4 4.8 4.4 Ghana 2.7 3.3 3.0 3.7 5.7 Gambia 12.8 14.6 16.3 Guinea 4.4 4.4 4.2 7.5 6.5 Guinea-Bissau 4.3 3.5 Madagascar 3.6 2.8 2.7 4.4 2.7 Malawi 9.6 8.0 6.6 8.3 9.3 Mali 5.2 4.4 6.1 7.1 3.3 Mauritania 6.8 6.7 5.4 7.1 9.3 Mozambique 11.1 11.3 12.1 11.2 12.0 Níger 9.0 11.7 11.9 Rwanda 1.9 2.6 2.9 3.3 3.1 Senegal 4.9 5.2 5.8 8.7 11.7 Sierra Leone 4.7 5.4 6.8 8.1 Tanzania 8.5 7.3 9.1 10.1 Uganda 6.7 6.5 7.4 8.6 9.6 Zambia 6.9 5.5 4.7 4.7 6.9 Median 6.2 5.1 5.5 5.9 8.1 Table 9. Health Expenditure as Share of GDP. 1998-2002. 18 countries 1998 1999 2000 2001 2002 Benin 1.1 1.5 1.4 1.9 1.7 Burkina Faso 2.4 1.5 1.4 1.5 2.0 Cameroon 0.6 0.6 0.9 1.0 1.5 Ethiopía 1.4 1.3 1.1 1.5 1.5 Ghana 0.8 0.9 0.8 1.2 1.5 Gambia 3.4 2.8 3.5 3.9 Guinea 0.4 0.4 0.4 0.5 0.8 Madagascar 0.7 0.5 0.5 0.8 0.4 Malawi 2.8 2.3 2.1 2.6 3.1 Mali 1.3 1.1 1.5 1.8 0.8 Mauritania 1.7 1.7 1.7 1.8 2.9 Mozambique 2.8 3.2 3.8 4.9 4.7 Rwanda 0.4 0.5 0.5 0.7 0.7 Senegal 1.0 1.1 1.2 1.9 2.5 Sierra Leone 1.0 1.6 2.2 2.6 Tanzania 1.2 1.3 1.5 1.7 Uganda 0.8 0.9 1.1 1.5 1.3 Zambia 2.0 1.6 1.4 1.5 2.2 Median 1.0 1.1 1.3 1.5 1.7 Source (table 8 and 9): health expenditures expenditures taken from World Bank Staff (HIPC Unit) for Burkina Faso, Ethiopia, Ghana, Madagascar, Malawi, Mali, Mauritania, Mozambique, Rwanda, Senegal, Sierra Leone, Tanzania, Zambia. For all other countries, IMF country reports. 21 The data indicate that there are very large variations across countries in the financial efforts which governments make to fund educational and health services. For instance, in 2002, while educational expenditures were less than 2.0 percent of GDP in three countries the share was 4.5 percent and over in seven countries. Similarly, five countries allocated 10 percent and less of their total government expenditure to education while four countries allocated over 22 percent. Similarly, in health, while four countries spent less than 1.0 percent of GDP, seven countries spent 2.5 percent and over; and while four countries allocated less than 4.5 percent of total government expenditures eight countries allocated over 8 percent. While bearing in mind the variations across countries, the data from tables 6-9 are summarized in table 10. There, the medians of the country shares are recorded for each year between 1998 and 2002 together with the number of countries which increased and decreased their shares. The main conclusions are that: Table 10. Median Education and Health Expenditures as Shares of Total Government Expenditure and GDP 1998-2002. African HIPCs. 1998 1999 2000 2001 2002 Cts.with increased Cts. with decreased share 1999-2002 share 1999-2002 Education % GDP 3.0 3.2 3.4 4.0 4.2 18 1 % TGE 12.8 14.9 14.9 13.9 15.5 12 7 Health % GDP 1.0 1.1 1.3 1.5 1.7 14 4 % TGE 6.2 5.1 5.5 5.9 8.1 16 2 Educ+health % GDP 4.0 4.3 4.7 5.5 5.9 % TGE 19.0 20.0 20.4 19.8 23.6 Note: GDP is gross domestic product and TGE is total government expenditure (a) the median share of GDP devoted to education in African HIPCs increased regularly from 3.0 percent to 4.2 percent. Of the nineteen countries for which data are available the share increased in eighteen. Only in Mali does the share appear to have fallen. (b) the median share of total government expenditures devoted to education also increased, from 12.8 percent to 15.5 percent, but at a slower rate particularly since 1999. Of nineteen countries, the share increased in eleven and fell in eight between 1999 and 2002. (c) the median share of GDP devoted to health increased, again regularly year by year, from 1.0 percent to 1.7 percent. Out of eighteen countries the share increased in fourteen. The exceptions were Burkina Faso, Ethiopia, Madagascar and Mali. 22 (d) the median share of total government expenditure devoted to health increased from 6.2 percent to 8.1 percent and in this case the share increased significantly since 1999. Only in Madagascar and Mali did it fall. (e) aggregating expenditures on both education and health, the share of GDP increased by almost 50 percent (from 4.0 percent to 5.9 percent) while the share of government expenditure remained constant for much of the time at around 20 percent but increased to almost 24 percent in 2002. In summary, the information paints a picture of most African HIPC governments directing higher shares of GDP into both the education and health sectors over the past few years, and increasing the share of total expenditure spent in the health sector. In three fifths of the countries the share for education of total expenditures has also increased. Part of the increases has resulted from these two sectors being central to governments' increasing tendency to give additional priority to the `poverty-reducing' sets of activities. As described above, in 2000 across eleven countries with such programs, the share for education averaged 57 percent and for health, 20 percent. There is no doubt that while the situation varies across countries, in the period since the reform of the HIPC Initiative in 1999, expenditures on education and health in a large majority of the African HIPCs have increased faster than overall government expenditure and overall economic activity. This leads to two further issues: the extent to which the HIPC process has caused these increases, and the composition of the increases. Can any increases in education and health expenditures be linked directly to HIPC relief and how are the proceeds/savings from debt relief handled in the budget? The HIPC Initiative reduces debt payments and, in the absence of other changes, thereby increases the amount of domestically generated revenues which can be utilized for domestic expenditures. It does not result in an additional inflow of financial resources. In this situation, the question arises whether it is possible to show that the resources `saved' through debt relief are being used as an increment in the education and health systems. The HIPC Unit in the World Bank (2003) reports that HIPC governments have indicated that education will absorb around 40 percent of the total relief while health will take a further 25 percent. In table 11, the increase in the share of education and health expenditures in GDP between the year preceding receipt of debt relief and 2002 is compared with the estimates of relief as a share of GDP in 2002 for five countries. No clear pattern emerges.3 In Malawi, Uganda and Ghana additional expenditures have been around 75 percent of the value of the HIPC relief. Conversely, in Ethiopia and Tanzania, expenditures have increased by about one third more than the value of the relief. 3The percentage point (pp) increases in the share of education and health expenditures in GDP between the base year and 2002 and the share of HIPC relief as a percentage of GDP in 2002 for the five countries are: Ethiopia 1.9 pp and 1.4 percent: Malawi 1.6 pp and 2.3 percent: Tanzania 1.9 pp and 1.4 percent: Uganda 0.9 pp and 1.4 percent, and Ghana 0.8 pp and 1.2 percent. 23 Table 11. Education and Health Expenditures and HIPC Debt Relief as Shares of GDP. 1999 - 2002 1999 2000 2001 2002 Ethiopía E and H as % GDP 4.1 5.6 6.0 HIPC relief as % GDP 0.8 1.4 Malawi Pro-poor E and H as % GDP 6.0 5.9 7.0 7.6 HIPC relief as % GDP 0.5 2.5 2.3 Tanzania E and H as % GDP 4.2 4.6 5.2 6.1 HIPC relief as % GDP 0.8 1.8 1.4 Uganda E and H as % GDP 4.8 5.8 5.7 HIPC relief as % GDP 1.4 1.4 Ghana E and H as % GDP 5.6 6.4 HIPC relief as % GDP 1.2 Source: IMF country publications Notes: E is education; H is health While the data in table 11 again suggest some relationships between education and health spending and the proceeds of HIPC debt relief they are not strong enough to confirm a causal relationship. Part of the reason is that budgetary procedures tend to be complicated and often not very transparent. In an effort to understand how HIPC relief is managed and spent, ECORYS (2003), for the European Commission, recently studied the process in five African HIPCs ­ Benin, Burkina Faso, Cameroon, Ghana and Tanzania.4 The results, summarized below, indicate that there is no single system for managing HIPC debt relief. · Managing the savings. In four of the five countries, HIPC savings are lodged in a special account. In two of these countries (Benin and Ghana) this is a sub account of the consolidated government account while in two others (Burkina Faso and Cameroon) it is separate from the consolidated account. In the case of Tanzania, there is no special account and HIPC funds are placed directly into the consolidated account. HIPC savings are allocated through the regular budget process in Benin, Ghana and Tanzania. In Burkina Faso, they are allocated in an extra-budgetary way by the Ministry of Finance after publication of the national budget. In Cameroon, the expenditures are included in the budget but the allocation occurs through a separate process. 4 IDA and the IMF have published two documents on tracking poverty-reducing expenditures: Tracking of Poverty-Reducing Expenditures in HIPCs (2001), and Actions to Strengthen the Tracking of Poverty- Reducing Public Spending in HIPCs (2002). 24 · Spending the savings. Each of the five countries has identified a set of poverty - reducing sectors within the PRSP process but the extent to which HIPC savings have been allocated to these varies. In Benin, there appears to be consensus that four of the six priority sectors are eligible. In Cameroon, all HIPC resources are earmarked for the six identified priority sectors. In Ghana, 20 percent of savings are used to reduce the domestic debt with the remaining 80 percent for the PRSP priorities. Burkina Faso has indicated that 79 percent of all expenditures on PRSP priorities will be provided by HIPC resources. Conversely, in Tanzania since the savings simply enter the consolidated budget, there is no attempt to link the savings with specific expenditures in the priority sectors. · Definition of priority expenditures. In several of the case study countries, the number of activities designated as priority has broadened over time. "In Burkina Faso, the revised PRSP of 2002 has broadened the sectoral allocation of HIPC savings, as compared to the original PRSP of 2000. In Ghana, the number of poverty-reducing areas has increased from 13 to 31. Tanzania has broadened the selection of the existing four priority sectors to seven. The MTEF for 2002 further broadened the definition of priority spending. As a result, the number of poverty- reducing areas and programs has increased to 15." (p.30). · Expenditure impact on priority areas. Again, there is variation across countries. In both Benin and Tanzania, poverty-reducing or priority expenditures have increased by more than the HIPC savings perhaps implying a catalytic effect. In Burkina Faso, part of the HIPC savings has been used for regular, non-priority expenditures while spending of some of the rest has been delayed. In Cameroon as of mid 2002, none of the accumulated 2000 and 2001 HIPC savings had been spent. In the case of Ghana, the first HIPC savings were only made in 2002, and 20 percent of these are planned for reducing the domestic debt. In the countries where the impact has been least, the savings are treated off-budget. · Recurrent-investment expenditure proportion. The extent to which the HIPC savings are spent on recurrent and on investment expenditures raises issues of sustainability (discussed below). For Ghana, where the savings are recent, and Tanzania, where they are meshed with all other revenue sources, such a division cannot be described. In the case of the other three countries, in Benin a lower proportion has been allocated to investment compared to the overall budget while in Cameroon and Burkina Faso, the alternative case has prevailed. · Overall, the report argues that all countries need to move to the situation exemplified by Tanzania where HIPC funds become part of consolidated government revenues and the overall poverty expenditure program is effectively monitored. In this context, the question of the impact of HIPC relief on education and health expenditures is redundant and the focus shifts to the overall increase in expenditures in a set of priority sectors, of which education and health are invariably part. 25 On what items have increased expenditures in education and health been made? As described above it is not always possible to isolate the connection between HIPC relief and specific expenditures, since in many countries the revenues are simply consolidated within aggregate revenues and/or the reporting of expenditure focuses on poverty-reducing expenditures as a whole. The situation for a sample of five countries is described below. In addition, a brief breakdown of those education and health expenditures deemed `priority' is provided for another three countries. Information is taken from recent IMF country reports. Benin. Almost all of the HIPC savings have been allocated to education and health. Of the total in 2001 and 2002, education received 42 and 43 percent respectively, and health 56 and 54 percent. Of the education expenditures, teachers recruited to fill vacant positions in rural areas consumed a quarter of the allocation and compensation to rural schools following the elimination of school fees consumed a third. The rest was spent on classrooms and cafeterias. In the health sector, recruitment of staff for vacant rural positions took 20 percent with the remainder allocated to implementing HIV/AIDs, antimalarial and other immunization programs, and improving access to safe water. Burkina Faso. Of the total HIPC relief for 2000-2002, only 41 percent had been spent by mid 2003, but the rate had improved considerably in 2002. The overall allocation for 2002 was 39 percent for education, 33 percent for health and 28 percent for rural roads. Recurrent expenditures were one third of the total. Cameroon. By the end of 2002, all projects to be financed through the HIPC debt relief on account of savings accruing in 2001 and 2002 had yet to be effectively initiated. Guinea. A report in 2003 noted that the share of wages and salaries in total social spending remains large and may be diverting much-needed funds from other necessary goods and investments; it suggested that the situation poses questions as to the strategy's long term sustainability once HIPC relief tapers off. Niger. The Special Program for Poverty Reduction focuses on rural education, health, food security and water systems and has been fully financed through by HIPC relief. The listing of the projects completed to April 2002 is dominated by equal shares of school classrooms and health units. Malawi. Eighty eight percent of the pro-poor set of education expenditures designated by the government for 2001/02 were for primary education. Of these, almost four fifths were for teachers' salaries. Health expenditures were 22 percent of the total. Of these, primary and secondary care shared the allocations equally; and within the allocations salaries and drugs/materials were allocated equal amounts. 26 Mali. Sixty percent of total budget expenditures are defined as poverty reducing, including 88 percent of spent on primary education, 46 percent of those spent on secondary and university education and 83 percent of total health expenditure. Mozambique. Expenditure in priority areas was 67 percent of total expenditure in 2002. Of this almost one third was for education, and of this 86 percent was for primary schooling. Allocations to the health sector were not restricted by level, and were equal to one fifth of total priority sector spending. From the available documentation it is difficult to discern the specific activities which have been directly funded by HIPC debt relief. Benin, Burkina Faso and Niger appear to be exceptions to this and in each case, education and health are clearly the beneficiaries. In those countries in which the relief is identified it can be presumed that, apart from cases such as Ghana where 20 percent is explicitly meant for reducing domestic debt, the activities funded are those within the poverty-reducing expenditure program. These vary in their breadth with several countries including all education and health activities, though most restrict it to primary and secondary and some to primary only (and indeed some to only rural primary). In the case of Tanzania, and other countries where the relief is not separately identified, it is not possible to link HIPC relief with any specific activities in the education and health sectors. The lack of data on the direct relationship between HIPC relief and specific activities is not surprising. In Tracking of Poverty-Reducing Public Spending in HIPCs (IMF/IDA 2001), it is stated that: "Perhaps the most fundamental principle of HIPC tracking is that it is neither feasible nor desirable to try to track only HIPC funding. Instead, all public spending on poverty reducing activities needs to be tracked, both to understand the impact of HIPC assistance and to encourage a shift toward more poverty-reducing public spending in the overall budget." Can measures be taken to ensure that the increased funding of those activities which have benefited from the HIPC relief (and the HIPC process) will be sustained? The HIPC Initiative aims to reduce the size of a country's existing debt stock, and consequent interest payments, to a level which can be sustained domestically. The amount of relief will be more or less constant up to the point where the debt stock has been paid off. As a share of GDP and total government revenue, it will, hopefully, continually fall as these measures increase. The HIPC Initiative can be responsible for a one time shift of resources into poverty-reduction related areas (including education and health) but not for a continuing shift. The implied hope behind the Initiative is that the new higher levels of expenditures on poverty-related expenditures (as shares of total government expenditure or GDP) will at least be maintained. However, there is the danger that as HIPC relief as a share of government revenues and GDP declines that specific programs which are associated with it will experience reduced allocations. 27 The greater any expenditures funded through HIPC relief are associated with the country's declared program of poverty-reducing expenditures, and the stronger the government's commitment to maintaining or extending support to that program, the greater will be the chances that expenditures will be protected. While the management of HIPC relief outside of the consolidated budget may result in the relief being more easily tracked and more closely associated with poverty reducing expenditures, in the longer term those expenditure items will be very vulnerable. Even in the case where relief goes through the budget but to specific activities, there is the possibility that those activities will eventually be starved of resources and be the least protected, even if part of a wider poverty-reduction program. Consolidating HIPC relief directly into the consolidated budget without any earmarking, and designing and implementing an effective well targeted poverty-reducing expenditure program (including poverty-reducing education and health programs) potentially offers the best guarantee for stable funding of the types of expenditure which the HIPC initiative is aimed at. The benefits, however, are not automatic and are dependent on a level of fiscal discipline which does not exist in all countries. Expenditures are more likely to be sustained if they support recurrent activities, especially salaries which are usually treated as a first charge by governments. This poses something of a dilemma for decisionmakers in the social sectors. In most countries, the greatest resource scarcity in the educational sector is in teaching-learning materials and other non salary items. A similar situation pertains in health. In practice, the trade off may be between guaranteeing larger quantities of (highly cost effective) non salary components over the short term against (less cost effective) additional workers over the longer term. In what ways is the HIPC process, through the selection of items for the completion point, attempting to improve the effectiveness, efficiency, and equity of total expenditures in the education and health systems. Many of the arguments for reforming the HIPC Initiative in 1999 focused on the constraints on education and health expenditures and the possibility that reduced levels of debt payments would lead to increases in these. In the final design of the reforms, however, the process was also seen as a vehicle for influencing both macroeconomic management and for improving the effectiveness, efficiency and equity of total government expenditures in specific sectors including, centrally, education and health. The mechanisms for this are the decision and completion points. To reach the decision point, governments are required to prepare a PRSP which focuses on poverty reduction and commits to poverty-reducing expenditures. To reach the completion point requires maintenance of macroeconomic stability, implementation of the PRSP for at least one year, plus the achievement of a set of structural and sectoral performance benchmarks. The benchmarks for education and health are central to this exercise and are listed in table 12 for the 23 post decision point African HIPCs. 28 Table 12. Completion Point Triggers for Education and Health in African HIPCs Country Education completion triggers Health completion triggers Benin A medium term expenditure program A medium term expenditure program which increases resources for basic which increases resources for education; reproductive health and HIV/AID; Elimination of school fees and grants to Targets for child immunization; schools to compensate; Implement Monitoring and evaluation Grants to school communities to hire system in health; teachers; Strategy for HIV/AIDS Eliminate repetition at grade 1. Burkina Faso Ensure adoption and endorsement of Targets for immunization; organization and budgetary measures of Targets for meeting health center the 1998 civil service reform; staffing norms; Abolish link between teacher training Target for incidence of insufficient colleges and civil service employment; generic drugs Establish new category of teachers for community hiring; Consolidate school promotion and limit grade repetition Cameroon Targets for new classrooms; Targets for immunization; Decentralization of teacher management; Targets for use of impregnated bed Adopt and implement new teacher nets for pregnant women statutes Develop strategy for HIV/AIDS Chad Targets for boys and girls primary gross Target for the proportion of all health enrolment rates, and for reduced districts and health centers which are repetition operational; Targets for immunization and assisted delivery; Targets for sale of condoms and for treatment of STDs Democratic Adoption of satisfactory education Adoption of satisfactory education Republic of strategies and plans strategies and plans Congo Ethiopía Target for the primary repetition rate and Targets for immunization and for the for the gross enrolment rate for girls utilization of health outreach facilities Gambia Target for graduating teachers; Targets for attended births Establish a trust fund for girls Increase the share of the total health scholarships with a target number budget for primary and secondary health care Guinea Targets for the overall primary gross Targets for immunization and for enrolment ratios and for girls prenatal consultation Specified increase in the number of new primary school teachers hired per year 29 Guinea-Bissau Eliminate fees for primary pupils Targets for immunization Target for the overall primary gross Targets for the use of impregnated bed- enrolment ratio nets by pregnant women Adopt a strategic framework to fight against HIV/AIDS Ghana Target for overall primary gross Target for percentage of households enrolment ratio with access to safe water targeted share of the health sector budget allocated to district and lower level governments Madagascar Increase the availability of teachers in Operationalize a generic essential drug rural areas through remuneration and supply system to district pharmacies allocation policies Malawi Target for education's share of total Target for health's share of total recurrent budget recurrent budget Target for teacher training intake and in- Target for recruitment of selected service training occupations Reallocation of expenditure from Completion of set of reforms for boarding too teaching-learning materials Central Medical Stores Direct supply of donor-supplied Appropriate budget for drugs and textbooks from suppliers to schools medical supplies Mali Reforms in budget allocation, teachers' Reforms in budget allocation and recruitment and scholarship policy personnel policy Mauritania Targets for overall and girls primary and Establish a central procurement lower secondary enrolment ratio facility for drugs and contraceptives Target of immunization Maintain HIV prevalence rate Mozambique Target set for increased share of Development and approval of new education in total current expenditure Health Sector Strategic Plan Implementation of multi-sectoral plan on HIV/AIDS Target set for increased share of health in total current expenditure Niger Targets for new classrooms and Target for increasing staff in rural volunteer primary school teachers health centers Country-wide school mapping Adopt a plan to improve availability of Limitation of repetition in grade 6 essential drugs in rural health centers Target set for immunization Rwanda Target set for primary gross enrolment Target set of full staffing and ratio equipment of district health centers Opeationalize 6 teacher training colleges Develop national plans to reduce Establish a framework for community malaria, infant mortality and maternal participation in primary and secondary mortality schools Implement framework for coordination Design and implement a capacity public, private and NGO providers. 30 building program for the management of education Sierra Leone Target for girls' primary gross enrolment Target for distributing treated bed nets rate Staffing of the HIV/AIDS Secretariat Target for training of unqualified Targets for immunization teachers Sao Tome Targets set of additional classrooms and Targets for immunization and new teachers primary health centers Targets for reducing malaria and child mortality rates Senegal Targets for teacher recruitment Targets set for immunization and pre New employment contracts for teachers natal care Maintain recent budgetary increases for Targets set for utilization of primary primary education as a share of education health care centers budget Tanzania Target set for coverage of school Targets set for immunization mapping Implementation of the HIV/AIDS campaign Uganda Increase budgetary allocation to primary Increase budgetary allocation to education at least at the rate of increase primary health at least at the rate of of nominal GDP. increase of nominal GDP. Zambia Target set for increased share of Implement action plan for malaria education in the discretionary budget Reorganize mechanisms for Raise teacher salaries in rural areas procurement of drugs above the poverty line for a household Earlier release of health expenditure Prepare action plan for increasing student budget data retention Minimum target set for actual releases to district management boards as a percent of amount budgeted Source: Country Decision Point and Completion Point documents As is clear from table 12 the setting of conditions in the education and health sectors, as part of a larger set, for triggering the final portion of debt relief, vary quite considerably across countries. In education, the outliers range between ambitiously requiring medium term expenditure programs and legislation to alter conditions of teacher employment to simply requiring school mapping. There is also a mixture of conditions which have a direct (upward) budgetary implication and those which focus more on management and the better use of existing resources. In education, it is clear that the overriding focus is on primary education with the triggers ranging from compensating schools for abolishing fees, to constructing additional school building, to implementing measures to reduce repetition, and to changing the legal statutes which govern teacher employment. Regarding spending, some countries face the condition of increased recurrent expenditures while for others increases in capital expenditures are required. In the health sector, the focus is on the primary health care system. More of the conditions appear to have direct budgetary implications, and there is a greater emphasis on setting quantitative 31 targets. Most countries have targets for child immunization. Overall, the conditions set for triggering the completion point are not insignificant and in the longer term are likely to have a greater impact on the development of the education and health sectors than any increases in expenditure which might be directly attributable to HIPC debt relief. III Next Steps As described in the Introduction, this Note has focused on only the first of three exercises which are required if the total impact of the HIPC Initiative on the education and health sectors in African countries is to be understood. The focus has been largely on changes in sectoral and poverty-reducing expenditures. To understand the full implications would require giving more attention to the effects which the process has also had on implementing sectoral reforms and on the ways in which total sectoral expenditures have been utilized; and then to the impact of both the increased resources and the reforms. The conclusion of the current analysis is that while there are significant variations between countries both in the financial benefits which have resulted from the HIPC Initiative and in commitments to increased poverty-reducing expenditures, overall the Initiative (through both the financial resources and the processes for gaining access to these) has had a positive effect on the funding for education and health in most of the highly indebted poor countries of Africa. 32 REFERENCES AFTHD (2002). Status of HIPC Debt Relief in African Countries: what scope does this Initiative offer for financing education (mimeo). World Bank ECORYS-NEI (2003). The Management of HIPC Funds in Recipient Countries : a comparative study of five African countries. Rotterdam. Garrett A. (2003). A Background Paper on the Enhanced Heavily Indebted Poor Country Initiative (HIPC) and the Poverty Reduction Strategy Papers. (mimeo). HDNED, World Bank HIPC Unit (2003). Financial Impact of the HIPC Initiative : first 26 country cases. World Bank IMF/IDA (2001), Tracking of Poverty-Reducing Public Spending in HIPCs. IMF/IDA (2002), Actions to Strengthen the Tracking of Poverty-Reducing Public Spending in HIPCs IMF/IDA (2003). Heavily Indebted Poor Countries (HIPC) Initiative ­ Status of Implementation IMF/IDA (2004). Heavily Indebted Poor Countries (HIPC) Initiative ­ Statistical Update Jubilee Research (2002). Relief Works : African proposals for debt cancellation ­ and why debt relief works. New Economics Foundation, London Operations Evaluation Department (2003). Debt Relief for the Poorest : an OED Review of the HIPC Initiative. World Bank Rowe F. (2004). Priority Sector Spending in the HIPCs: Has Debt Relief Helped? HIPC Unit, World Bank. (mimeo) Tan J-P., A. Soucat and A. Mingat (2001), Enhancing Human Development in the HIPC/PRSP Context : progress in the Africa Region during 2000. Africa Region Human Development Working Paper Country documentation Poverty Reduction Strategy Papers IMF country documents including: Article IV Consultations 33 Reviews of the Poverty Reduction and Growth Facility Statistical Appendix Selected Issues and Statistical Appendix Recent Economic Developments Joint Staff Assessments of PRSPs and Progress Reports HIPC Documents including Decision Point Documents Completion Point Documents 34