53959 Botswana Public Expenditure Review March 2010 PREM 1 Africa Region Document of the World Bank This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization. CURRENCY EQUIVALENTS (Exchange Rate as of March 30, 2010) Currency Unit = Botswana Pula US$ 1.00 = BWP 6.xx GOVERNMENT FISCAL YEAR April 1 – March 31 Vice President: Obiageli Katryn Ezekwesili Country Director: Ruth Kagia Sector Director: Sudhir Shetty Sector Manager: John Panzer Team Leader: Zeinab Partow This Report has been prepared by a team consisting of Zeinab Partow (TTL), Naoko Kojo (Senior economist, PRMED – Chapter 2 on Diamonds and Medium Term Sustainability); Michael Stevens (Consultant, Chapter 3 on Planning and Budgeting); and Jeanine Braithwaite (Senior Social Protection Economist – Chapter on Social Safety Nets). List of Acronyms HRM Human Resource Management AICD Africa Infrastructure Country Diagnostic AfDB African Development Bank AG Auditor General AIS Accumulated Investment Surplus ARV Anti-Retroviral IBRD International Bank for BEAC Business and Economic Advisory Reconstruction and Development Council IFMIS Integrated Financial Management BIDPA Botswana Institute for Information System Development Policy Analysis IFS International Financial Statistics BOB Bank of Botswana IMF International Monetary Fund BOBC Bank of Botswana Certificate KPP Keynote Policy Paper BSP Budget Strategy Paper KRA Key Results Area BURS Botswana Unified Revenue Service LIPW Labor Intensive Public Works BWP Botswana Pula M&E Monitoring and Evaluation CCT Conditional Cash Transfer MDA Ministry, Department, Agency CF Consolidated Fund MFDP Ministry of Finance and CGE Computable General Equilibrium Development Planning CHBC Community Home Based Care MIC Middle Income Country CSO Central Statistics Office MLG Ministry of Local Government CY Calendar Year MTBPS Medium Term Budget Policy DF Development Fund Statement DPSM Directorate of Public Service MTEF Medium Term Expenditure Management Framework DSF Debt Sustainability Framework MTFF Medium Term Fiscal Framework DSS Department of Social Services MTR Mid-Term Review DTC Diamond Trading Company NDDC National District Development ECC Economic Committee of Cabinet Conference FAD Fiscal Affairs Department NDP National Development Plan FDI Foreign Direct Investment NPV Net Present Value FY Fiscal Year O&M Operations and Maintenance GABS Government Accounting and OAP Old-Age Pension Budgeting System OECD Organization of Economic GDP Gross Domestic Product Cooperation and Development GFCF Gross Fixed Capital Formation OVC Orphans and Vulnerable Children GIA Government Investment Account PAC Public Accounts Committee GICO Government Implementation PDL Poverty Datum Line Coordination Office PDSF Public Debt Service Fund GNP Gross National Product PEFA Public Expenditure and Financial HIES Household Income and Expenditure Accountability Survey PER Public Expenditure Review HR Human Resource PETS Public Expenditure Tracking Survey PFM Public Financial Management PIP Public Investment Program PPG Public and Publically Guaranteed PPP Public Private Partnership PSFP Primary School Feeding Program PSIA Poverty and Social Impact Analysis PSRU Public Service Reform Unit RADP Remote Area Development Program RBM Results-Based Management RSF Revenue Stabilization Fund SACU Southern Africa Customs Union SBI Sustainable Budget Index SNA System of National Accounts SOE State Owned Enterprise SSA Sub-Saharan Africa SSN Social Safety Net TB Tuberculosis TEC Total Economic Cost TFP Total Factor Productivity TNS Thumbnail Sketch VAT Value Added Tax VGFP Vulnerable Groups Feeding Program WEO World Economic Outlook Executive Summary The unifying theme of this Public Expenditure Review (PER) is the impending need for Botswana to incorporate into its policy making and planning framework a series of measure to help it transition into an era of diminished diamond revenues. While diamond revenues have enabled rapid development, they have also allowed government to swell, and the public sector is today, with diamonds, a mainstay of the economy. With the decline of diamonds on the horizon, it has become increasingly urgent for Botswana to review and question its management of public finances and the role and objectives of the public sector in the economy, particularly given the large remaining human and social development needs, and the requirements of diversification. While the global economic crisis has helped to bring these issues to the fore, they are long-term challenges that are independent of today’s temporary economic difficulties. Chapter 1 reviews the pillars of Botswana’s economic management that have supported its macroeconomic, fiscal and development success and outlines some emerging challenges. Rapid mineral development has been by far the major source of government revenue in Botswana since the 1970s, and it remains the country’s primary revenue source today, although its contribution has declined to just below 20 percent of GDP. Part of Botswana’s success in managing its resource windfall has been the acceptance by policymakers that mineral revenues should be treated as the proceeds of an asset sale to be used to finance investment rather than recurrent spending, in order to sustain development for future generation. In addition to channeling resources towards investment, savings from boom times were accumulated in order to smooth public expenditure and have provided a cushion against frequent periods of drought as well as weak diamond markets. As government spending is the main link between the mining sector and the rest of the economy, the role of fiscal policy is critical in Botswana. Yet fiscal policy is today facing a number of challenges. Revenues from diamonds and the SACU pool are stagnating. Revenues are likely to remain at about 30 percent of GDP over the next 5 years, compared to an average of 38 percent in NDP 9 (2003-2009). Given Botswana’s historically strong revenue stream, there have always been equally strong pressures to spend and spending and the size of government have grown. And while the planning system has been quite successful in maintaining aggregate fiscal discipline and resisting some of the compulsion to spend excessively, money has been available for most identified public expenditure needs, and today public spending account for 46 percent of GDP. The combination of a decline in revenues and an edging upward of spending has resulted in more frequent deficits. Looking to the coming decades, projections are for an accelerating decline in diamond production and associated revenues. Spending pressures, however, will remain high. High poverty rates, the need to address the costs of the HIV/AIDS epidemic, the backlog of unmet basic social needs, and investments still required to strengthen Botswana’s competitiveness and spur the diversification of the economy ensure that the pressures on the public purse will continue. Chapter 2 analyzes Botswana’s medium term fiscal sustainability in view of the expected declining revenues. The analysis shows that in the absence of policy correction, Botswana’s current fiscal policy is unlikely to be sustainable over the longer term, regardless of the evolution or impact of the global crisis. Large fiscal deficits will be accumulated in the near term, requiring financing by a combination of borrowing and the drawing down of fiscal savings. While small surpluses are projected for the outer years of NDP10, as government revenues continue to fall in 2016 and beyond, the situation would be quickly reversed leading to a growing structural fiscal deficit. Government savings would be depleted at about the same time as diamond production is projected to begin a rapid decline and the possibility that Botswana could fall into serious debt distress cannot be ruled out. However, Botswana can improve its fiscal prospect significantly by expanding domestic revenue sources to partially offset falling mineral revenues. Yet the extremely high profitability of diamonds simply cannot be match by other sources of revenue and an adjustment to an economy where government spending is a lower share of GDP is therefore needed. Reducing expenditures to a level commensurate with long-run projections of diamond revenues – i.e. between 25 and 30 percent of GDP -- is fundamental. Chapter 3 argues that the current planning and budgeting system may not be well suited to an environment of constrained resources or to the need to do more with less. Given the high remaining development needs in Botswana, a reorientation of public expenditures focusing on improving the quality of spending is needed. The chapter examines whether a planning and budget system that has been very successful in converting aid and mineral revenues into public infrastructure and social services over many years has the attributes the government requires in the very different circumstances facing Botswana in the years ahead. The conclusion is that Botswana’s current NDP/dual budget planning and budget model scores poorly as a toll of aggregate fiscal management and at achieving efficiency and effectiveness in spending. A fundamental budget reform, towards program based budgeting accompanied by a Medium Term Expenditure Framework is the recommended option to addressing the main weaknesses of the present system. This should not be seen as a completely new mechanism in Botswana, which for year has built its National Development Plans around a proto-MTEF. To ensure success, however, a strategy for managing the change of the planning and budget system needs to be developed, understood and owned by the political and bureaucratic leadership. Chapter 4 reviews public investment planning in Botswana, a particularly crucial component of development in mineral-based economies given the emphasis on transforming mineral wealth into productive assets. It describes some of the challenges faced in public investment planning and management as a result of the separation of the current and development budgets and the focus on a government-wide public investment program (PIP). The incentives and spending dynamics inherent in such a project-led (vs. policy-led) system militate against effective and efficient public investment by encouraging an expansion of expenditures and the gradual deterioration of public investment into wish lists that are not anchored in fiscal reality or the imperatives of service delivery. Chapter 5 focuses on social safety net expenditures (SSNs). Strengthening and rationalizing social protection services is relevant in the context of declining diamond revenues and the need for structural transformation of the economy. Botswana spends a significant amount on SSNs, many of which are universal/categorical and not means-tested or targeted. SSNs are also fragmented and hidden in various indirect social expenditures by the state. While the impacts of the SSN have been in many ways positive in terms of reducing poverty and supporting improved nutrition, most programs suffer from low coverage and high leakages, and they do not adequately cover the poor. Very significant informational challenges, however, make a thorough analysis of SSNs impossible. In addition to improving information on the poor and on the impacts of SSN programs, the chapter recommends enhanced targeting of some of the larger programs. Chapter 1: Fiscal Policy in an Evolving Reality 1. Botswana’s management of its mineral revenues has been justly acclaimed. Commentators have lauded Botswana’s ability to avoid the prudently manage its mineral wealth and to transform one of the world’s most impoverished countries into an upper middle income one via four decades of virtually uninterrupted, rapid growth. But Botswana is facing the long term decline of its diamond revenues, which have provided the basis of growth and development over the past four decades. While diamond revenues have enabled rapid development, they have also allowed the role of government to swell, and the public sector is now, with diamonds, a mainstay of the economy. With the decline of diamonds on the horizon, it has become increasingly urgent for Botswana to review and question its management of public finances and the role and objectives of the public sector in the economy, particularly given the large remaining human and social development needs, and the requirements of diversification. The current global economic crisis has helped to bring these issues to the fore, but these are long-term challenges that are independent of today’s temporary economic difficulties. 2. The unifying theme of this Public Expenditure Review (PER) is the impending need for Botswana to incorporate into its policy making and planning framework a series of measures to help it transition into an era of diminished diamond revenues. The report is organized as follows: - This chapter sets the stage for the following sections of the PER. It reviews the pillars of Botswana’s economic management that have supported its macroeconomic, fiscal and developmental success. The emergence of important challenges to this success is discussed, and a path forward is outlined. - Chapter 2 analyzes Botswana’s medium-term fiscal sustainability in view of the expected sharp decline in diamond production. Revenue and expenditure projections are developed for the public sector over the medium and longer term. Sensitivity analyses are carried out to examine the robustness of the fiscal paths projected in the baseline scenario, and alternative fiscal strategies are then examined for bringing Botswana’s fiscal stance back to a sustainable path. - Chapter 3 reviews the budget and planning process and outlines the features of potential alternative paths to reform. The central question motivating this chapter is the identification of a system of planning and budgeting that is most likely to help the Government address the challenges of diminishing diamond revenues of the medium term, as well as the need to diversify the economy and seek growth in new directions. This is partly about shifting resources to new spending priorities, but also about achieving better value for money in existing resource use, by changing the incentives structure for budget actors. - Chapter 4 reviews public investment planning and management in Botswana, a particularly crucial component of development in mineral-based economies. It focuses on an assessment of the institutional process within which project planning and management occurs, and the challenges posed by public investment in the current dual budget system. - Chapter 5 focuses on social protection expenditures, with a view to provide policy options for the government in its quest to improve the outputs and outcomes of social protection programs. Strengthening the social protection services also becomes very relevant in the context of declining diamond revenues in the medium term and the need for structural transformation of the economy. To help secure the buy-in of the population for what may be disruptive reforms, a coherent, well-targeted system of social programs is needed. I. Introduction 3. At independence in 1966, Botswana inherited modest administrative machinery whose primary function was the maintenance of law and order, not development. With meager human and capital resources before the discovery of diamonds in the early 1970s, Botswana in its early post- independence years was still almost totally reliant on grants-in-aid from Britain. These paid for a significant portion of recurrent spending; aid was also virtually the only source of development spending. An overriding objective for policy makers of the day was national financial self-sufficiency and the reduction of the degree of external control. Surpluses in the recurrent budget were achieved in 1972/73, through control of government spending and increased tax revenues; overall budget surpluses were attained starting in 1983/84 and have characterized 22 of the past 27 years. The premium placed on fiscal prudence during those early years set the tone for later fiscal management, and subsistence under harsh, drought-prone conditions taught the value of effective use of limited resources, and of saving when surpluses occurred. (Harvey 1992; Lange and Wright, 2002; Somolekae 1998). 4. Today’s National Development Plans (NDPs) have their roots in those early years when medium term planning was quickly developed as a means of utilizing limited resources in the most effective way possible. The planning system, a key part of Botswana’s early success in navigating the dangers of its diamond windfall, was geared towards channeling donor funding and mineral rents to development projects, and to maintaining a high rate of expenditures, provided these were reasonably productive. An important initial motivation for carefully prepared development plans was to attract foreign financing (Nordås et al., 1998). 5. The closing of the physical and social infrastructure gap in the country was a cornerstone of government spending policy in the first decades following the coming on-stream of diamond revenues. Rapid accumulation of capital stock was financed initially by a combination of donor financing and by FDI; later, government savings financed the bulk of public investment, as they do today. Policy-makers succeeded at creating the appropriate framework for each of these financing sources. Sound macroeconomic policies and careful cost-benefit analysis of major development projects attracted donor resources, while policies ensured reasonable returns to both private FDI and the nation (Leith, 1997). 6. Along with a strong planning system to channel mineral rents, success also rested on complementary pillars related to fiscal policy. Government fiscal policy has traditionally been built on keeping current spending in line with expected long run revenues, not short-run receipts. Botswana was able to smooth its consumption over time: when diamond receipts rose unexpectedly, expenditures did not rise to any significant degree. As a result, foreign exchange reserves and government balances have been built up at the Bank of Botswana for use in lean times. Having conserved the boom income, government did not consume all of the boom revenues domestically. Savings were allocated between international reserves and domestic investment. Policy makers took into consideration the economy’s short-run absorptive capacity for new domestic investment in setting investment ceilings, and put into place a system for vetting investment proposals that was fairly successful in excluding low-benefit projects, at least through the mid-1980s (Hill, 1991). 7. Fiscal policy was successful in avoiding external debt problems and maintaining a stable rate of growth over time. The strategy of smoothing consumption and building up foreign exchange reserves, combined with proper management of the exchange rate, has also meant that real exchange rate appreciation remained under control. There is little evidence of a consistent long-run overvaluation of the currency. While the manufacturing and agriculture sectors have, respectively, stagnated and shrunk, this is not thought to be the result of exchange rate appreciation ((Pegg 2010; Harvey xxxx; IMF, xxxx). 8. In addition to enlightened government policies and strong institutions, the management of Botswana’s diamond booms has been facilitated by a number of other factors. The direct links between the diamond mining sector and the rest of the economy are weak; were they stronger, the volatility of the sector would have created more widespread economic shocks affecting employment and incomes. Instead, the large financial reserves accumulated by government over the years have been used to cushion the economy from the volatility of diamond prices and demand. Diamond revenues are also not quite as volatile as those of some other commodities, particularly oil.1 The stream of rents from diamonds expanded gradually, and only began to exceed domestic absorptive capacity from the mid-1980s. Another factor that may have facilitated managing resource rents is that real GDP, real government recurrent spending and real development spending have been growing at very high rates. Public spending doubled every 5-6 years in real terms between the mid-1970s and the early-1990s. It is easier to accumulate reserves and resist pressure to further increase spending under such benign circumstances (Auty, 2001; Bank of Botswana, xxxx; Hill 1991; Pegg, 2010). 9. Although diamond production in Botswana is expected to continue at high levels for another decade or so, the prospects of mineral revenues accruing to government are highly uncertain. In the absence of new diamond discoveries or other sources of revenue to offset declining income from minerals, fiscal revenues are set to shrink sharply sometime during NDP 11 or 12. Chapter 2 of this PER carries out an analysis of medium-term fiscal sustainability in Botswana in view of this forecast rapid decline in diamond reserves. In the absence of policy reforms in revenues and expenditures, the analysis concludes that Botswana’s current fiscal stance is unlikely to be sustainable over the longer term. This evolving economic reality presents policy makers with significant, but certainly not insurmountable, tests. The challenges of stagnating revenues, the expansive role of the public sector in the economy, ever-increasing pressures to spend, and a weakening investment planning and management system require a change in the culture of managing public finances. 10. This stakes are raised by the significant remaining development challenges facing Botswana and which will require substantial resources to address. Poverty remains high for a MIC at over 30 percent of the population living below the national poverty line, as does income inequality. Incidence rates for HIV/AIDS are some of the highest in the world and have dramatically reduced life expectancy and increased infant mortality, and pose a significant social, economic and health care challenge. Infrastructure services still generally lag behind those of comparator countries, and diminish Botswana’s 1 Nevertheless, significant movements in revenues are fairly typical – often shifts of 15% or more – and diamond output has frequently been reduced in order to maintain prices. Botswana has not been immune to periodic revenue volatility. global competitiveness. There are significant resource needs to spur diversification into new sectors in order to sustain the growth of the economy as revenues from diamonds ebb. II. Fiscal Policy in a New Reality 11. A crucial question for mineral dependent economies is whether mineral assets are being transformed into other productive sources of income rather than consumed by government and the current generation. Part of Botswana’s success in managing its resource windfall has been the understanding and acceptance by policymakers that mineral revenues should be treated as the proceeds of an asset sale rather than value-added in production, and that these proceeds should be used to finance investment, rather than recurrent spending, in order to sustain development for future generations (Jefferis and Kelly, 1999; Lange and Wright, 2002). As government spending is the main link between the mining sector and the rest of the economy, the role of fiscal policy is critical in Botswana. This section looks at how this principle has been translated into policy, the evolving nature of public finance in Botswana and the mounting challenges to putting such policies into practice. Stagnating revenues 12. Rapid mineral development has been by far the major source of government revenues in Botswana since the 1970s. During the 1980s, with the coming on-stream of production from the country’s second major diamond discovery, Jwaneng, real revenues grew at an average yearly rate of close to 16 percent, quadrupling over the decade. This was the hey-day of diamond revenues, which accounted for up to 30 percent of GDP in some years. In addition to the direct revenues from mineral taxes and royalties, there was a rapid accumulation of foreign exchange reserves during this period, and earnings on these reserves became another major mineral-related source of revenue for Botswana, peaking at 23 percent of total revenues in 1996/97, but dropping to about 4 percent today. Figure 1: Figure 2: Revenues have been historically buoyant But the rate of revenue growth has been declining T R G R G R à à 13. In addition to adroit negotiations with De Beers which have enabled Botswana to capture much of the diamond rents, policy makers were cautious to sterilize the rent stream.2 Financial 2 Government recognized that given the high uncertainty associated with the value of mineral deposits, a fixed royalty rate would not be appropriate. A sizeable share of the profits of mining operations was preferred, and obtained, by requiring that “in addition to a modest royalty, that government be provided, at no charge, a portion of reserves grew rapidly in the 1980s, growing ten-fold between 1984 and 1992. The saved mineral revenues continue to be allocated between domestic investment and international reserves. The ratio of foreign to domestic investment is important because excessively rapid domestic absorption may give rise to Dutch disease effects. Today financial reserves stand at 13% of GDP. Close to 40 percent of the rents have been sterilized in offshore investments by the BoB. The interest earned makes a significant annual contribution (4%) to government revenue in the form of transfers from the BoB. Nevertheless, and inevitably, the allocation of savings to international reserves began to be increasingly questioned, starting in the 1980s, and pressures to increase spending continue today (Hill, 1991). Figure 3: Figure 4: Mineral revenues -to-GDP are falling As are Bank of Botswana transfers M GDP BOB T à 14. Additional efforts were made to further shield the economy from the vagaries of volatile revenues. Even before diamond revenues began to flow, in 1972 and during the construction of Botswana’s first diamond mine, the government established two funds in recognition that diamond revenues would exceed its absorptive capacity, threatening Dutch disease. The Public Debt Service Fund (PDSF) was established to help avoid overheating the economy by saving a portion of the diamond revenues. In order to address another challenge raised by mineral revenue windfalls, namely volatility, and to buffer the revenues that are absorbed via public spending from price shocks, the Revenue Stabilization Fund (RSF) was also set up in the same year to accumulate reserves during booms that could later be used to cushion downturns (Pegg, 2010). 15. Today minerals remain Botswana’s primary revenue source, although as a proportion of GDP their contribution has declined to just below 20 percent (in 2007, prior to the global crisis). As a proportion of total revenues, income from mineral exports has also declined slightly from between 55 and 60 percent in the mid- and late 1980s, to an average of 47% in the 5 years prior to the recent crisis. Looking towards the longer term, government and Debswana projections, while subject to a great deal of the equity in the mining operation”. The government also purchased additional shares in both Debswana and De Beers (Leith, 2005). uncertainty,3 are for diamond revenues to begin declining in a decade or so as a result of gradually depleting reserves and higher production costs.4 16. Receipts from the renegotiated Southern Africa Customs Union (SACU) agreement constitute the public sector’s second source of revenue, contributing slightly less than 10 percent of GDP and 25 percent of total revenues. The combination of a renegotiated SACU formula and a South African import surge are largely responsible for the growing importance of this income source between 2003/04 and 2008. These revenues are declining dramatically as high-value imports by South Africa have plummeted. Over the medium term, with advances in globalization and pressures for more liberal trade practices, the tendency should be for these transfers to decline. There are also pressures within SACU to review and reform the character of the revenue-sharing pool. The robustness and magnitude of these resource flows to Botswana over the medium and long run are thus uncertain at best. 17. The dominance of mineral revenues, complemented by SACU resources, has slowed the development of other revenues sources. Revenues from minerals and the SACU pool constituted 80-85 percent of total revenues until recently. The non-mineral income tax and the VAT have historically accounted for a minor portion of total revenues. 5 During the past decade a positive shift has taken place, however, in public revenue sources with these domestic sources growing gradually in importance. Prior to the global crisis, in 2007/08, the VAT and the non-mineral income tax contributed 11 and 10 percent of total revenues, respectively.6 With the budget bottom line being less important in the construction of domestic revenue policies, however, these revenue sources still do not contribute a major share of total public sector resources.7 18. In sum, Botswana’s revenues are stagnating. Medium term fiscal projections – following NDP 10 assumptions – are developed in Chapter 2 of this PER and indicate that revenue performance is likely to remain weak during the NDP 10 period. Revenues are likely to remain at about 30 percent of GDP, compared to an average of 38 percent in NDP 9. While mineral revenues would gradually recover to pre- crisis levels, they would be offset by declines in SACU revenues. In the longer term, beyond NDP 10 and absent policy reforms, revenues would decline further. Buoyant revenues that characterized the decade between the mid-1980s and mid-1990s have been gradually falling and since the mid-1990s the rate of growth of mineral revenues has been in decline. While total government revenues between 1984 and 1994 averaged 46 percent of GDP, they declined by ten percentage points of GDP to an average 36 percent between 2001 and 2007 (and declined further, of course, over the past two years as a result of the global crisis). Spending patterns have not yet, however, caught up with this reality. 3 Debswana, the 50-50 joint venture between the Government of Botswana and De Beers, currently projects its typical “life of mine” to 2040, although there are ongoing investments to extend the lifetime of diamond production at e.g. Jwaneng. 4 Diamond revenues constitute about 90 percent of total mineral revenues. 5 In 2002 a value added tax replaced the existing sales tax. 6 In 2009/10 it is estimated that VAT and non-mineral income taxes will generate 15 and 23 percent of total revenues, respectively. Part of this reflects a significant improvement in collections in absolute terms, but it is also a result of the decline in the size of the total revenue pie with the fall in mineral revenues. 7 It should be noted, however, that the government has taken a wider view of taxes, restructuring income taxation in order use low taxes as a stimulus for economic growth, establishing Botswana as a low tax jurisdiction with the Southern Africa region. Increased Demand for Public Spending 19. Given Botswana’s historically strong revenue stream from diamonds and other minerals, there have always been equally strong pressures to spend. The planning system has been quite successful in maintaining aggregate fiscal discipline, resisting some of the compulsion to spend excessively, and avoiding overheating the economy and the degeneration into an inflationary spiral. 20. Policy-makers have been cognizant of the temptation to overspend since the early days and have, over time, tried to institute rules to control these pressures. One principle of fiscal policy in Botswana has been that of smoothing public spending, in line with the temporary character of mineral revenues where the appropriate policy response is to keep current spending in line with expected long run revenues, not short-run receipts. Another has been the saving of boom incomes. These spending and saving decisions have been made in the context of Botswana’s six year National Development Plans. 21. The government has generally succeeded in charting a relatively smooth path for public spending (Figures 6 and 7). It has also avoided increasing expenditures to any significant degree in the aftermath of an unexpected rise in diamond receipts, a very common event given the frequent underestimation of mineral revenues in budgets and NDPs. Figure 8 shows the percentage difference between actual revenues (expenditures) and original budget estimates. Revenues were underestimated by as much as 20 percent in some years, but this did not lead to a burst of spending – until the past two years. In years where revenues were over-estimated (2001/02 – 2004/05) the variation between actual and budgeted spending remained stable over the period. Figure 5: Figure 6: Savings are invested in forex reserves Expenditure smoothing F E R à ” h à à R E Figure 7: Figure 8 Expenditure smoothed despite volatile revenues Spending unresponsive to revenues rises R S G R A B R E T R G T E N L 22. In order to successfully smooth public consumption, savings were accumulated, largely in the form of foreign exchange reserves (Figure 5). As noted above, this was another important ingredient of Botswana’s success in managing the mineral windfall and the pressures to consume it domestically. There has been a conscious policy of not spending all mineral revenues, given the nature of diamonds as an exhaustible resource, and Botswana has successfully managed to save a large proportion of its boom incomes, treating them as temporary windfalls. Reserve accumulation was justified by MFDP and BoB because projections of future export earnings suggested that high levels of surpluses would not continue. Where investment projects yielding adequate returns could not be identified, mineral revenues were accumulated as financial assets held by the Central Bank, which themselves yielded substantial returns (Figure 4). The savings have provided a useful cushion during frequent periods of drought as well as weak diamond markets, such as occurred in the early 1980s, and of course, today. 23. Nevertheless, money has been available for most identified public expenditure needs, and today public spending accounts for 46 percent of GDP. In fact, often the constraint on public spending has not been the availability of resources, but absorptive capacity. Thus, although government savings in the form of financial reserves have grown considerably – including in recent years between 2005 and 2007 – this was a result of capacity constraints in spending the development budget, and not discipline associated with a rigorous appraisal and evaluation of projects, which appears to have continued to weaken considerably over the years. Original spending plans in 2005-07 were much higher, and budgets were under-spent by as much as 15 percent during those years (Figure 8). This trend has been reversed over the past three years, since 2007/08, and the development budget has expanded by an astonishing 84 percent (in constant prices) between 2007/08 and 2009/10. The spending boom began prior to the onset of the global crisis and cannot be attributed entirely to government efforts at fiscal stimulus. In the absence of any apparent efforts to improve the investment management system, it seems clear that the discipline behind project selection and evaluation is succumbing to spending pressures (see the next section and Chapter 4). 24. The combination of a decline in revenues and an edging upward of spending has resulted in more frequent fiscal deficits (Figure 10). Looking to the coming decades, projections by government and Debswana are for an accelerating decline in diamond production and associated revenues. Spending pressures, however, will remain high. The need to continue to address the costs of the HIV/AIDS epidemic, the backlog of unmet basic social needs, and investments still required to strengthen Botswana’s competitiveness and spur the diversification of the economy all ensure that the pressures on the public purse will continue. The diminishing gap between revenues and expenditures means that Botswana is unlikely to be able to increase its savings at a rate similar to the past. The concern with fiscal sustainability that naturally arises when faced with declining revenue prospects and significant spending needs and pressures is discussed in detail in Chapter 3. Figure 9: Figure 10: Spending growth Stagnating revenues, spending pressures, and deficits R E B à ‡ $ à 9 à à R D B R E The Quality of Public Investment 25. The quality of public investment is a critical ingredient for putting into practice the underlying principle of public finance in Botswana, namely offsetting the depletion of mineral assets by a compensating increase in other forms of capital. Given this principle, choosing to use mineral revenues to finance investment rather than increased saving must be premised on the assumption that physical and human capital investments will yield higher returns than saving mineral rents in the form of, e.g. foreign exchange reserves.8 Yet serious concerns have been raised regarding the quality of public investment in Botswana, as far back as the early-1990s. Most recently, the 2010/11 budget speech focused on the quality of public investment, highlighting shortcomings and stressing the need to more rigorously evaluate public investment projects(Bank of Botswana Annual Reports, 1993, 1995 etc; Salkin, 1996; Leith, 1997; Wright, 1997; Jefferis and Kelly, 1999; Lange and Wright, 2002; World Bank/IMF). 26. Botswana’s rapid growth in the 1970s, and perhaps through much of the 1980s, was largely the result of high investment in both human and physical capital, much of it financed by government. Growth was facilitated by a very substantial increase in both the quantity of capital and labor inputs, and public policy was instrumental in this by supporting education, health, basic infrastructure, and a generally open economy. In addition to a rapid accumulation of human and physical capital, increases in productivity made a significant contribution to economic growth in the 1970s and 1980s. Total factor productivity (TFP) calculations, although beset by data as well as conceptual difficulties, indicate that between 1975 and 1990, TFP growth was positive and high, particularly in the early years.9,10 Since 1990, however, TFP growth has been negative.11 There appears to have been a shift 8 The government publically adopted the principle that all mineral revenues should be reserved for productive investment in 1994 and introduced the Sustainable Budget Index (SBI), discussed in paragraph 52. 9 For a critique of the Botswana TFP calculations, see Wright, 1999. 10 The early 1980s saw a large increase in the mineral sector’s capital stock, but demand for diamonds fell at just about the same time as a result of the 1981/2 global recession. The result was excess capacity and a drop in TFP in the early 80s which recovered later in the decade (Leith, 2005). from high payoffs to infrastructure and human capital investment in the first decades following the discovery of diamonds, to declining productivity of these investments over the past twenty years (Leith, 1997 and 2005; World Bank, 2005). 27. Looking more specifically at government investment, the government’s share in fixed investment has been high in Botswana, above 30 percent; over the past five years it has averaged 40 percent. 12 The associated gains in terms of the building of basic infrastructure and the provision of health, education and other services – where virtually none existed in the mid-1960s - have been remarkable. When Botswana achieved independence in 1966 it had 12 km of paved roads; today it has 6,000 km (and 12,000 km of unpaved network). Tarred roads reach most parts of the country. Education enrollment in primary, secondary and university increased from 67,000 in 1965 to over 500,000 today. Virtually the entire population is within 15 kilometers of a primary care facility. Ninety five percent of the population has improved access to a water source and 40 percent to improved sanitation facilities. But it is also true that over time, the investments have become less efficient in promoting long-term growth. While public spending in Botswana is high for a middle income country, the country ranks below other MICs in social indicators. Nearly a quarter of Botswana’s public spending goes to educate her children, significantly more than in comparator countries, for similar, or poorer results (Figures 12-15) (Leith, 2005; WDI, 2009; World Bank, 2009). Figure 11: Public Spending on Education 30 @‘ ƒÃQr…ÃT‡ˆqr‡ÃQ…v€h…’ QˆiyvpÃTƒrqvtÃ@qˆph‡v‚ pˆ ……r‡ÃVTÇ 25 1400 20 1200 1000 15 800 10 600 5 400 200 0 h h v r y h h 8 † h v  h i v v u p v …  v ‡ D H ˆ v ‡ i 0  8 s  v € € V … † 6 r ‚ ‡ ‚ h I à u t … ˆ h y ‚ Argentina Chile Mauritius South Africa Colombia Botswana Namibia ‡ 7 ˆ 6 H 8 ‚ T % of GDP % of total public spending @‘ƒÃQr…ÃT‡ˆqr‡ÃTrp‚qh…’ @‘ƒÃQr…ÃT‡ˆqr‡ÃUr…‡vh…’à pˆ……r‡ÃVTÇ pˆ……r‡ÃVTÇ 2500 30000 25000 2000 20000 1500 15000 1000 10000 500 5000 0 0 Argentina Chile Mauritius South Africa Colombia Botswana Namibia Argentina Chile Mauritius South Africa Colombia Botswana Namibia 11 The Investment Climate Assessment found that TFP in Botswana’s manufacturing sector (i.e. private sector) was lower than in most comparator countries – SACU economies and Argentina, Chile, Malaysia and Mauritius – and less than 50 percent of TFP in South Africa (World Bank, 2005). 12 Gross Fixed Capital Formation (GFCF). 28. In the early years after independence, an expanding government was undoubtedly good for growth. There was no private sector capable of taking on the challenge of building the country’s human and physical capital and providing services and financing. High rates of public investment in physical and human capital generated rapid economic growth because of the high productivity of those investments. Yet today, a high and rising public share of total fixed investment, with a stable or declining GDP, supports the view that public capital spending has not been efficient in promoting growth in more recent years. 29. There are a number of reasons behind the decline in the returns to public investment: with the large and expanding role played by public sector investment, the marginal productivity of this spending is likely to decline over time, even for well chosen projects.13 Benefits decline as the range of public investment is extended outside the provision of core public goods and diminishing returns can set in as the stock of public infrastructure expands. The overall level of spending on investment may also be negatively correlated with quality and cost, particularly in an environment with constrained implementation capacity. As capacity is stretched, the likelihood of less beneficial projects being approved increases. Greater demand for investment may lead to shortages and price increases in inputs, resulting in reduced and delayed returns from all projects, and possibly in previously good projects becoming unproductive. This seems to have happened in the late 1980s and early 1990s when the level of capital spending rose rapidly and is a clear danger today with the very large expansion in the development budget (Lange and Wright 2002; Leith, 1997; World Bank 1994). 30. A strikingly soft budget constraint, the result of continuous budget surpluses, has also been damaging to the quality of public investment. Moreover, the dominance and ease of collection of a few major revenue sources and the rapid rate of growth of revenues over long periods, not only affects the development of other revenues sources, but also the rigorous control of spending (Wright 1997). This has been made worse by the chronic tendency to forecast budget deficits: official forecasts have been projecting an imminent emergence of budget deficits since the mid-1980s. The large surpluses and a “cry wolf” aspect to fiscal projections has encouraged complacency towards restraining public spending. There are signs that this may well continue during the current NDP 10, despite government projections for balanced budgets in the final three years of the Plan. The NDP 10 development budget has been “frontloaded”, i.e. development spending is concentrated in the early years of the Plan, and expected to decline during the latter years. This may be an optimistic assessment. In the past, when budgets were frontloaded, high rates of spending growth continued throughout the Plan period. Bunching of expenditures in the early years is also likely to create excess demand on the limited capacity of the economy, particularly in the construction sector. This has happened in the past (e.g. late 1980s to early 1990s, as well as in the late 1990s), and the result has been an acceleration of inflationary pressures. 31. The surfeit of resources appears to have weakened the attention paid to cost-benefit analysis of projects. The introduction of careful project analysis was in part a response to the requirements of donors who provided financing for these projects through the early-mid 1980s. As government became less dependent on foreign donor financing, the discipline was diluted. “Marginal projects were approved and when costs escalated, turning once high return projects into low or even negative return projects, the original approval was seldom re-evaluated” (Leith, 1997). By all accounts, this still appears to happen 13 Chapter 4 discusses in greater detail issues related to investment planning and budgeting. today. The 2010/11 Budget speech noted problems related to investment budgeting across most ministries. These include “cost-overruns caused by inadequate budgeting, … [and] change of scope of projects without authority for additional funds.” The speech also notes that project appraisal needs to be “tightened to make it more robust to ensure that only projects with the highest potential added value are approved… and that any subsequent project changes do not change the favourable net appraisal of the project.” 32. Abundant resources have encouraged a focus on new construction, while maintenance of existing infrastructure is given less priority. Large sums are periodically approved to help clear the accumulated backlog of maintenance of public infrastructure14. Generally, the economic benefits of infrastructure investments are significantly reduced by insufficient attention to costs and benefits in project choice, low operating efficiency, and inadequate maintenance. In Botswana, roads are the obvious example where it is fairly clear that not enough resources have been devoted to maintenance for the Roads Department, and that there is a backlog of roads to be maintained. Average annual maintenance expenditure is a mere one percent of the total road asset values, and is clearly inadequate. A significant proportion of roads, previously in excellent condition, today needs various forms of rehabilitation. Much the same can be said of government buildings and other publically financed infrastructure (World Bank, 1994, 2009). 33. An overly generous definition of what constitutes the development budget reduces transparency and makes it difficult to monitor the relationship between the recurrent and development budgets and the transformation of mineral revenues into other productive assets. The development budget includes a number of items that would not be classified as investments under most definitions. The inclusion of maintenance expenditures has been mentioned. Botswana also classifies 30 percent of the recurrent budget as investment spending, on the basis that recurrent spending in areas such as health and education represent investment in human capital. Military spending is treated as investment in Botswana. Drought relief expenditures have been classified as capital spending in order to get donor support. HIV/AIDS treatment programs and ARVs are regarded as a development budget item. There are other examples. This is fairly extreme practice and complicates efforts to gauge the overall quality of investment spending. It also makes it very difficult to estimate and monitor the recurrent budget implications that derive from the investment program, crucial information for the public sector to ensure that it is assigning adequate resources to maintain the investments it finances. 34. The declining attention to careful project evaluation and appraisal, the relegation of maintenance spending, a loose definition of investment are all reflections of a debilitated planning system that is losing the ability to control costs and guarantee quality. There is also evidence of weak procurement practices, including poor selection of suppliers resulting in delays, higher costs, poor quality, if not outright lack of delivery. As monitoring and evaluation capacity is weak, when projects fail, lessons are not learned.15 14 In November 2001 (and in 2007/08 BoB Annual report) a total of P670 million was approved (Lange and Wright, 2002) 15 There is growing awareness of these shortcomings, which are beginning to be addressed by the Government Implementation Coordination Office (GICO) in the Office of the President. A National Programme Management Office is currently being set up, also in the Office of the President, in an effort to institutionalize M&E and monitor the performance of government. 35. Capital spending is the main means through which government transfers mineral rents to the population, and this role is taken very seriously as reflected by the emphasis on development spending in the planning process. But the ability of the planning process to deliver high quality investment spending has weakened. In fact, it appears that the planning process is at times by-passed entirely. It is not clear, for instance, if the Hubs - special initiatives to address what are considered to be major bottlenecks to competitiveness and diversification in the country – have benefited from a full cost evaluation or incorporated in the NDP process.16 The allocation of savings to international reserves rather than domestic investment is also being increasingly questioned.17 While government still does not spend all boom revenues on investment, in the last three years the choice seems to have been made to save less and to spend more on domestic investment. 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 and will continue beyond it. This is worrisome because the increased development spending is not clearly associated with increased capacity to properly implement this spending, and it may be that increasingly constrained mineral revenues are being consumed without sufficient attention being paid to ensuring value-for- money. 36. Part of the problem is that government has chronically promised more development spending than its administrative capacity can deliver. This has generated a view that government should be carrying out more investment than it currently does, and combined with a weak budget constraint, has heightened pressures to deliver on a large development budget without a complementary focus on maintaining a strong investment management system. This inflated development budget is also the result of a decentralized budget process which accumulates spending requests as a series of uncoordinated wish-lists from ministries and departments rather than a rational assessment of what is both feasible and desirable at the macro level. It is a commons problem where, in a decentralized setting, those who shape spending programs have little concern for the associated costs. While the budget process in Botswana has not formally or explicitly become more decentralized, the balance of influence between different ministries (and various branches of government) appears to have been altered. There is a loosening of authority exercised by MFDP in coordinating the budget process, which has not been helped by large financial reserves that have made the setting of a hard budget constraint very difficult. The assumption increasingly appears to be that all development spending is productive, and this goes against one of the pillars of Botswana’s success to date, that the quality of investment is key in turning mineral wealth into other forms of wealth to ensure sustainability. Competitiveness and the Size of Government 37. While the mining sector contributes the largest share of national output, it is government that plays a more central role in the Botswana economy (Table 1). Government is the largest employer, saver, investor and the second largest economic sector. Government is also a major consumer of the goods and services produced by other sectors, and is the driver of the construction sector while 16 Six Hubs have been identified in the following areas: Agriculture, Health, Education, Diamonds, Transport, and an Innovation Hub. Some, (e.g. the Agriculture Hub) have larger fiscal implications than others (Transport) which focus on policy and regulatory reforms. 17 This has been the case since at least the mid-1980s. For instance, in 1987 Parliament approved a proposal to use a portion of the accumulated reserves on economic development initiatives in areas considered to be major bottlenecks to economic development (Hill, 1991). pushing domestic demand more generally. The government has also played a significant role in the domestic financial sector with budget surpluses being a major source of savings in the economy. Moreover, several publicly owned development finance institutions were established, and government was also engaged in substantial lending to parastatals and local authorities; from the early 1980s to the late 1990s, this lending was higher than the combined lending of all other financial institutions in Botswana. Given the characteristics of the economy at independence, reinforced by the discovery of diamonds shortly thereafter, it is difficult to imagine a development scenario without a large government. Beginning in the 1990s, the mining sector’s share of GDP began to decline slightly, while that of government grew, and the trend continues today. Table 1: Importance of Government in the Economy1 (%) 1980s 1990s 2000s2 Share of GDP 13.4 15.8 15.5 Share of Formal Sector 34.5 36.1 38.3 Employment Share of Investment3 38.6 46.5 37.6 Central Government Spending/GDP 34.1 37.6 36.0 Source: CSO, National Accounts; World Bank, World Development Indicators; IMF, International Financial Statistics 1 General Government, including Central and Local 2 Share of GDP and Investment is through 2008/09 3 Gross Fixed Capital Formation 38. Government expenditure in Botswana has grown to be one of the largest in Africa. Although aggregate spending levels never increased to a level that threatened macroeconomic stability, the ratio of government spending to GDP has grown over time. From an average of 34 percent of GDP in the 1980s, government spending rose to 38 percent between 1990/91 and 2003/04. Following three years of large reductions, spending in 2008/09 and 2009/10 reached 40.2 and 46.4 percent of GDP, respectively. For 2010/11, government budget estimates are for central government spending to reach 39.3 percent of GDP. Of the sixteen countries in Africa where recent data is available, only Lesotho has a higher ratio of central government spending to GDP than Botswana.18 In comparison with other upper middle income countries worldwide with similar per capita GDP, total public spending is also high (Figure 12).19 Figure 12: Central Government Spending, Figure 13: Central Government Employment % of GDP (2008) Wages 18 Data are for 2008. Lesotho’s government expenditures were 48 percent of GDP in 2008, vs. 40 percent in Botswana. But data for Lesotho is distorted by the fact that migrant remittances constitute 20 percent of GDP, making the GDP figure much lower than GNP. 19 Botswana’s government spending levels are more in line with those of European countries, although in the latter welfare payments frequently constitute over half of total spending. In Botswana in contrast, welfare payments to households represent a minor portion of total spending (around 3%). Of much greater importance is the direct absorption of resources by government through consumption, payment of employees and investment which together usually constitute a small portion of government spending in developed countries.  Ã Ä  à 9 b h  h  ” CG LG GDP Source: World Bank, Unified Surveys; IMF Article IV Consultation Source: CSO, National Accounts and Formal Employment publications Reports *Wage bill for South Africa is for the Non-financial Public Sector 39. Beyond the fiscal and effectiveness aspects of spending, the size of government has a bearing on competitiveness of the economy as a whole. The impact of government on the economy is perhaps greatest with respect to employment. The government wage bill is high, averaging 10 percent of GDP and 28 percent of spending over the past five years (Figure 13). Of 45 Sub-Saharan African countries in 2007/08, only ten had a higher central government wage-GDP ratio than Botswana’s.20 The actual wage bill has exceeded original budget estimates in seven out of the past ten years, sometimes by a significant margin of up to 15-20 percent. Central and local government employment constitutes about 40 percent of formal sector employment, and this proportion has edged upwards since the mid-1990s from levels closer to 35 percent, but has been a fairly stable ratio since then.21 Employment in parastatal companies is not part of this total, and would add another 7 percentage points if included. In terms of employment growth, the government sector has contributed 41 percent of all formal sector jobs created between 1985 and 2009, although this ratio has declined from 52 percent of total jobs in the 1990s to 36 percent over the past ten years. Moreover, a large part of the growth in private sector employment is stimulated by government spending, especially in construction. The size and growth of the wage bill may be affordable from a narrow budget perspective, but it has an important impact on the broader economy, crowding out private sector access to scarce resources. 40. Not only is government the largest single employer by far, it also pays generously. Government workers claim more than their share of total earnings: while they constitute 40 percent of the total formal workforce, they claim over 54 percent of total wages paid. Central government employees enjoy 40 percent higher wages than the rest of the private and parastatal sector, and the difference increases to 64 percent if the parastatal sector is excluded (Table 2). The public wage-per capita GDP ratio, which indicates whether government employees are under- or overpaid in comparison to the prevailing standard of living, is, at 2.7, very high relative to other MICs (IMF, 2007). 20 These were Burundi, Guinea-Bissau, Eritrea, Sao Tome & Principe, Cape Verde, Swaziland, Zimbabwe, Lesotho, Namibia and the Seychelles (IMF, 2008). 21 The increase in government employment as a proportion of total formal employment in the mid-late nineties coincides with the emergence of the phenomenon of youth unemployment. Table 2: Monthly Average Cash Earnings by Industry, 2002-20091 I A M Q M W E C W R H R T C F I R E E H S W O C S P P P P C G L G A S Source: CSO, Stats Brief 2009/10 1 Figures are for March of each year, except 2002 which is September. Local Government does not include the Ipelegeng public works program. 41. Government employment affects the labor markets at both the high and low skill ends of the spectrum. Table 3 show the share of government in total payments to various categories of labor. Unfortunately this information is available for 1993/94, but not more recently.22 The government’s very high share of payments to professional and technical employees is of particular interest. In all categories, the government’s share of factor payments is higher than its share in the economy, and is significantly larger than the share of other economic sectors. Other evidence indicates that government pay scales are more compressed than in the private sector – higher at the lower end and lower at the higher end. The government has therefore been a wage leader at the lower end and has contributed significantly towards making Botswana a relatively high cost economy. (Bank of Botswana, xxxx; Pegg, 2010). Table 3: General Government Demand for Labor, 1993/94 Category Government share of total payments (%) Professional and Technical 59.8 Administrative and Managerial 30.9 Clerical 57.8 Skilled Manual 22.6 Unskilled 37.8 Total 41.9 Source: Social Accounting Matrix, 1993/94 42. The government has been outbidding the private sector for available labor. Government wages and salaries have grown rapidly, reducing the ratio of average monthly earnings in the private 22 A Social Accounting Matrix exists for 1996/97, but the figures for Government factor payments are problematic. sector relative to government (Figure 14). With four decades of high rates of economic growth, there have been shortages of skilled labor. These shortages have been addressed through efforts at on-the-job and other training in the public sector and state owned enterprises (although private sector spending on training has been very low according to the Botswana Investment Climate Assessment23). Expatriates fill the remaining shortages. An interesting point to note is that crowding out of the private sector has taken place through formal sector employment rather than through traditional crowding out channels associate with Dutch disease, namely the exchange rate and high real interest rates. Figure 14: Manufacturing and Central Government Earnings à à à Source: CSO. 43. The lack of skilled labor therefore does not seem to have constrained the growth of the economy, but has materialized as higher costs in the form of significantly higher wage bills. Shortages are made worse by the lack of signals from the labor market regarding which skills command a premium, and the incentives to acquire the needed skills have been absent. This is a particular problem in the case of youth unemployment, which emerged in the 1990s. The number of graduates is growing at a much faster rate than is the demand for its labor. Unemployment for the 15-24 age group stands at 32.6 percent compared to total unemployment rate for the labor force of 17.5 percent (2005/06). With wages heavily influenced by the public sector, and with poor market signals, the youth labor market has not adjusted to changing circumstances (Leith 2005). 44. It is fair to say that diamond rents have led to an over-expansion of the role of government in Botswana. This is particularly apparent in terms of the labor market. Diamond revenues have allowed the growth of a large government workforce with overly generous wages. The reservation wage that workers demand before accepting employment has risen, as workers expect salaries that are significantly higher that the market clearing levels in the non-government sectors of the economy. Concerns about a rising reservation wage despite high unemployment have been expressed by government on several occasions, including the Mid-Term Review of NDP 9 and the 2006 Budget Speech. The latter notes that “…Botswana workers have become selective in the jobs they are willing to accept, choosing instead to rely on relatives or social safety nets”. University graduates desire well-paying white-collar jobs, which are relatively scarce outside government due to limited private sector-led economic diversification. 45. Social programs in Botswana target a large population, and this poor targeting exacerbates the problem of a high reservation wage (Chapter 5 has a more in-depth discussion of social programs.) About a third of the population benefits from social programs. While beneficiaries are mostly people too 23 World Bank, 2007. old or too young to be in the labor force, the broad scope of the programs may indirectly provide incentives to dependency. This can be inferred by the rising number of welfare beneficiaries and the high unemployment despite strong economic growth over many years. 46. How government spends its money is of critical importance to the economy as a whole. There is a tight link between public spending, competitiveness and the potential for diversification and sustained growth.24 A large government sector and high wages exert upward pressure on costs in the entire economy, particularly given a constrained skills base. High costs mean that employers create fewer jobs than they otherwise would, leading to higher unemployment. In addition to high labor costs, Botswana also has high utility and transport costs because of its geographic location. High costs and a small market size, made effectively smaller by high rates of poverty and inequality, reduces Botswana’s attractiveness as an investment destination as it depresses anticipated profits. As a result FDI inflows have been low and concentrated largely in the capital-intensive minerals and mining sector, limiting investment’s potential positive impacts on poverty and inequality. This is despite Botswana’s exemplary policy reforms to attract foreign investment, and generally high scores on transparency and business climate surveys. III. Planning and Budgeting A. The Planning System 47. The previous sections have argued that fiscal policy-making in Botswana has been disciplined, and has helped the country avoid many of the pitfalls of dependence on mineral revenues seen elsewhere. Policy has minimized balance of payments crises and has provided a cushion for crises, including the current one. GDP growth has been high and less variable than income from mining. The rate of growth of spending has been smoothed, further contributing to the stability of growth. Overall, Botswana has tended to assume that mineral booms are temporary and has avoided the costs of having to cut expenditures and declines in non-boom output. 48. Yet over time the discipline in policy-making appears to have softened, be it in the costing of projects, in the definition of what constitutes investment, in hiring government employees and setting public wages, or in the commitment to fiscal rules. Investment budgets are bloated with projects of questionable value, and the wage bill has been overly generous and is crowding out the private sector. It is not only the quality of spending that has been affected. Even aggregate fiscal discipline, which Botswana has successfully ensured over decades, is now in the balance with more frequent and larger deficits and an outlook of sharply lower revenues in the coming decades. The pressure to spend has grown and the saving of mineral rents at the expense of domestic investment is increasingly being questioned, despite a consensus that the quality of investment spending has been in decline. The ramifications affect not only fiscal sustainability, but have a bearing on competitiveness and successful 24 Experiences from emerging market countries show that resource-poor countries tend to grow through internationally competitive industrialization. Since commodity exports are limited, labor intensive manufactured exports expand rapidly and soon absorb excess labor. This allows the economy to diversify within the skill-intensive manufacturing sector and later to enter into competitive capital-intensive industries, along the way making the economy more resilient to external shocks. Unemployment is reduced and saving and investment are boosted. Low unemployment leads to improvements in income equality. In Botswana, on the contrary, both unemployment income inequality remains stubbornly high, some of the highest rates among middle-income countries. diversification and on the assets and welfare of future generations. One of the underlying reasons for this diminished discipline is a very soft budget constraint. One way to address it is through the Planning System. 49. The bedrock of Botswana’s economic management is its Planning System. To a great extent, the disciplined, growth-promoting fiscal policy is attributable to the institution of the National Development Plans, a consensual process reinforced by effective financial controls. The system of multi- year planning of expenditures successfully kept total spending within bounds by keeping expenditures from growing as rapidly as revenues. From its roots as an exercise to attract donor funds for infrastructure, the planning process has evolved into a fairly sophisticated system. Much care is taken in the production of the NDPs. Spending ministries (and SOEs) develop their desired recurrent and capital expenditures for the projected six year Plan period. Major policy issues facing the ministries are discussed and resolved in planning documents. Separately, MFDP constructs feasible alternative spending paths for the Plan period. Priorities are then agreed in an extensive consultative process which includes all ministers and senior officials, and culminates in the adoption of the Plan by Parliament. This is supplemented by annual reviews of ongoing projects and an extensive mid-term review of the overall Plan during its third year. The national planning process is not only important for fiscal discipline, but also for setting out basic policy objectives. The current NDP 10, for instance, focuses on fulfilling the Vision 2016 pillars: an educated and informed nation; a prosperous, productive and innovative nation; a compassionate, just 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 (Leith 2005). 50. Include something on the annual budget? From CABRI report – NDP is annualized through a budget prep. Process that covers the five major categories: manpower budget, recurrent budget, development budget, consolidated fund revenue estimates and development fund revenue estimates…. Box 1: The Mechanics of Botswana’s Planning System Botswana uses a dual annual budget structure guided by a medium term national development plan. National Development Plans are a blueprint of the country’s six-year development priorities and forecast. Botswana is now in its tenth Plan. From the onset, the Botswana government was clear as to why it wanted to plan. The approach is clearly outlined in NDP 6: “Government is committed to careful planning to ensure that finance, manpower and national resources are used effectively and in accordance with national priorities. The progress that has been made since independence is the outcome of carefully planned government actions. However, the government’s commitment to planning is not intended to stifle private initiative but to create favourable conditions in which the private sector can contribute to national development.” Preparation of the development plan in Botswana takes place along the following stages: 1. The Keynote Policy Paper (KPP). MFDP starts the process by preparing a keynote paper, which serves two main purposes. First, it identifies the theme of the next development plan, and second, it provides an overview of the resource position within which the Plan will have to operate. 2. The Macroeconomic Outline. MFDP prepares and distributes to all ministries macroeconomic forecasts for the coming Plan period. The ceilings incorporated into the Outline assist individual ministries in drafting their respective sectoral chapters, and is a preliminary version of the macroeconomic chapter of the Plan. Ministries comment on the Outline. The Economic Committee of Cabinet discusses the macroeconomic outline which is then approved by Cabinet. Manpower ceilings are also prepared at this time to guide ministries in their staffing plans. 3. Proposed Sector Programs. Ministries and local authorities submit their proposed programs and projects, accompanied by cost estimates, for inclusion in the NDP, based on the KPP. Ministries must operate within the Manpower, Recurrent and Development expenditure ceilings approved by Cabinet. For each individual project, proposed project costs and associated recurrent costs must be phased by year. Additionally, each project included for NDP10 consideration must be supported by a project summary of Thumbnail sketch (TNS). The TNS sets out the basic features of the project: financial, technical, environmental impact, social and economic, and should be supported by tables indicating how the figures have been derived. 4. Sectoral chapters. Ministries prepare draft sectoral chapters on the basis of the macroeconomic outline and drafting guidelines. The chapters are distributed to all ministries for comment and then revised on the basis of comments. Permanent Secretaries in each ministry are responsible for ensuring that the cumulative total of all Departments does not exceed the ministerial targets for expenditure and manpower ceilings approved by Cabinet. 5. Draft NDP. MFDP consolidates the draft macroeconomic chapters and the revised sector chapters to produce a draft NDP. 6. The National District Development Conference (NDDC). This Conference is called to discuss the draft NDP. Representatives of all local authorities, NGOs, the private sector as well as line ministries all attend. Following the conference, MFDP consolidates all resolutions and recommendations into a Major Issues Paper which is presented to the Economic Committee of Cabinet (ECC) for discussion. The ECC makes final resolution regarding issues from the NDDC. The draft NDP is then discussed by Cabinet. 7. Parliamentary debate. The Draft NDP is distributed to Parliament for consideration. Parliament debates the Plan adopts it. The Plan, once approved, is financed through annual budgets. Botswana follows a dual budget system where recurrent and capital budgets are developed separately, with separate ceilings defined for each. Nevertheless, there are drafting guidelines that set out the procedure for calculating the recurrent costs arising from each project. A Mid-Term Review (MTR) is prepared mid-way during the Plan, where projections and policies are revised in light of progress in the first half of the Plan period and changing economic circumstances. 51. This planning model may be losing its ability to deliver on its raison d’etre, namely the judicious channeling and transformation of mineral wealth into development programs. The planning and budgeting system made sense for a country which needed to rapidly build up its physical and social infrastructure to enable development to take place. It contributed greatly to Botswana’s economic development and the reputation it gained for effective use of both external and domestic resources. But there has been a growing perception for some time now – including within government – that the system is not as effective as it once was and a number of criticisms have been leveled at the traditional planning and budget system. Overall, budgeting has evolved into an increasingly incremental process, with little regard to efficiency and with limited incentive to critically evaluate spending plans. This is particularly true of the recurrent budget, as the previous sections have tried to illustrate. The in- house macroeconomic forecasting capacity of MFDP has weakened substantially; this debilitates an already rather tenuous link between the objectives of the NDP and its quantitative framework. On the investment side, the concern for costs and development value of projects has been eroded, if not abandoned. Project planning is beset by an increasing lack of realism, including a failure to realistically estimate and meet staffing needs or other recurrent cost implications of investments. As already noted, the development budget now includes a mix of capital and recurrent expenditures making it virtually impossible to properly evaluate Botswana’s success in transforming mineral assets into other types of assets. There is a bias towards new investment while existing infrastructure goes unmaintained. Ambitious development proposals of national importance appear to be decided outside of the normal competitive NDP process of project screening – contrary to the provisions of the Finance and Audit Act.25 B. Complementary Fiscal Planning Tools 52. Complementing the planning and budget system proper, an array of tools and guidelines has been implemented at different junctures to promote a disciplined approach to budget management and to maintain control over expenditure. The existence of these tools shows a clear concern with sound financial management. But have they been successful at the practical level? Undermined by a soft budget constraint, the effectiveness of these instruments in practice is questionable. When put under pressure, the status and usefulness of this array of fiscal tools has become ambiguous, as discussed below. 53. While the first large diamond mine was still under construction, the government created two special funds in the 1970s to help manage mineral revenues and to enable the effective domestic absorption of diamond rents to avoid Dutch disease effects. With much foresight, recognizing that diamond revenues would exceed absorptive capacity, and following textbook recommendations, the government established a Public Debt Service Fund (PDSF) in 1972 to enable it to save money, to meet future debt obligations, rather than contribute to economic overheating by spending it. To address the volatility of mineral revenues, Botswana established the Revenue Stabilization Fund (RSF), also in 1972. The RSF was established in order to accumulate reserves during booms that could later be used to cushion downturns and cover temporary shortfalls of revenues in order to ensure the expansion and maintenance 25 According to the Finance and Audit Act, no expenditure can be incurred on any project prior to its inclusion in the Plan. This means that any new capital spending projects, or new recurrent expenditure activities linked to new policies, cannot be undertaken unless included in the Plan. In this way, the Plan is designed to offer a vehicle for disciplined implementation within a medium-term planning framework. of public services. These two funds have not, however, been used for the purposes originally intended. The amounts allocated to PDSF were not driven by a strategy for future debt management. Over the years the PDSF became principally a tool for lending of public funds to parastatals and local authorities at subsidized rates. The amounts deposited in the RSF were also not based on its role as a buffer stock. Government cash balances were not assigned to these funds. Substantial budget deficits could be supported via discretionary drawing down of cash balances, rather than being subject to the expenditure discipline envisaged by the establishment of the PDSF and RSF. Today, these funds are inactive, with no resourcess or loans being appropriated or issued from either the PDSF or the RSF (Wright, 1997; Pegg, 2010). 54. In 1994, the Sustainable Budget Index (SBI) was introduced, originally as a ceiling which must not be breached to monitor the commitment to ensuring that mineral revenues were directed towards investment. The SBI is defined as the ratio of non-investment spending to current revenues. An SBI value of 1.0 or less is interpreted to mean that current government consumption is being financed entirely out of non-mineral revenues, and is thus sustainable. Another ratio, the Accumulated Investment Surplus (AIS) further argues that asset revenues that remain unspent on productive investment cannot be used at a later date to finance consumption spending, although there is little evidence that this index was put to use. By around 2000, the SBI began to be binding (the ratio of 1 was breached between 2000/01 and 2003/2004) despite the generous definition of what comprised investment spending and its status was reduced to that of a measure to be monitored. Today it does not feature in budget discussions (Lange and Wright, 2002; Wright, 1997). 55. In more recent years other informal rules have been introduced to backstop efforts to control spending and its components. As part of the MTR of NDP 9 in March 2006, a spending limit of 40 percent of GDP was established for the central government. 26 The limit was reached in 2008/09, and is expected to be handily breached in 2009/10. The same MTR mandated a ratio of development spending to total spending of 30 percent. The objective was to ensure balance between the recurrent and capital expenditures, to avoid salaries and wages absorbing an ever increasing share of spending and to guarantee a larger portion of the budget to social and physical infrastructure investment on which long run growth depended. During the first two years of implementation, the 30 percent target was not reached, due to the perennial problems with capacity to implement the development budget. In 2008/09 and 2009/10, however, this was reversed, and the target was achieved and exceeded. It is not yet clear if this has been the result of the resolution of capacity constraints or a further easing of controls. 56. An important element of budget planning is the sustainability of recurrent spending. As Botswana follows a dual budget system, with separate recurrent and development budgets, establishing a link between development budget expenditures and their recurrent spending implication is critical to ensuring that government can pay for the maintenance of the infrastructure it builds. The relationship between these two sections of the budget has long been recognized in Botswana. At the macro planning level, the coefficient linking the development to the recurrent budget was used to design government spending strategies that were consistent over the planning period. At the sector level, the coefficient was 26 The basis for setting the figure at 40 percent was not clearly justified beyond the fact that this had been the level for the previous two or three years, and that this was a decline from extraordinarily high spending levels reached during 1999-2001, the final years of NDP 8. NDP 10 reduces the spending ceiling to 35% of GDP and states that given expectations for the long-term decline in diamond revenues, total government spending will need to decline to 30 percent of GDP by the end of NDP 11 (2021/22). used, along with development budget ceilings, to derive recurrent budget targets by ministry and department. This recurrent budget coefficient, in conjunction with rules such as the SBI that monitored the ability of government to finance its recurrent budget out of domestic revenues, along with targets for the ratio of government cash balances to annual expenditure and net lending, allowed policy makers to set aggregate development and recurrent budget ceiling and targets which were in accordance with anticipated government revenues. In practice, however, a long-standing complaint in the planning process has been the inadequacies and inaccuracies in the recurrent expenditure estimates included in the project summaries in the NDPs and project memoranda submitted to the MFDP for funding. This shortcoming is also evident given the predilection with new investment and the frequent disregard of maintenance spending (Salkin, 1996). C. Reforming the Current System 57. Despite this apparent erosion of the planning and budget system, Botswana has continued to perform well, with strong economic growth rates and a stable fiscal environment. This is perhaps largely because it continues to enjoy strong public financial management basics, and of course, plentiful revenues. The legal framework for the budget is well established and complied with. Sector strategies are prepared in the NDPs. There is a functioning Integrated Financial Management Information System, with timely financial statements and accounts. There are also functioning payroll controls, and a well- paid civil service. Internal and external audit systems are adequate, as is the legal framework for procurement (PEFA, 2008). There is also a genuine political commitment to improve public sector performance. 58. The current model of dual budgets and a medium term development plan worked well in the early days of independence when government was small and basic services needed to be created for poverty reduction and development. Being capital budget-led, dual budgets are best suited to an expanding government. The system is less well-suited for a middle income country like Botswana, which has a large public sector and an already extensive reach of public services. In terms of the three levels of effectiveness (see chapter 3) used to assess budget processes, the existing system comes up short. Because of its expansionary tendencies, the current model is not very good at setting and achieving sound fiscal objectives. With respect to the allocation of resources in line with government priorities, the rigidity of the system is increasing over time, as budgets follow Plan policies and programs, which are changed infrequently. And in terms of promoting efficient and effective spending at the level of service delivery, performance is poor given that the focus in on implementing projects, not on delivering services. 59. Large fiscal needs make it important for Botswana to develop the capacity to reprioritize spending more effectively, particularly as revenues plateau. These needs include the requirements of HIV/AIDS prevention, treatment and support, poverty alleviation, growing demands for spending on social protection, not to mention the human and capital investment needs related to competitiveness and diversification. Improved prioritization requires an improved capacity to identify program-related expenditures. This is not possible under the current budget system. As Chapter 5 notes for instance, Botswana does not have a system to track and monitor social safety net (SSN) spending by program classification. It is therefore impossible to effectively track this spending or what it accomplishes. Moreover, SSN expenditures are splintered into a number of initiatives across various public sector entities, making it difficult, if not impossible to monitor and evaluate the effectiveness of public spending. 60. There is understandable trepidation and reluctance to change the NDP system that has brought so many benefits to the country. Botswana’s planning and budget system has been extraordinarily useful in managing mineral revenues and directing their use towards developmental goals; but it is obvious that the system is flawed. Increasing stresses on fiscal sustainability are the result of weakening links between budgeting and planning. The failure of public investment to ensure the quality projects is also fundamentally a failure of the planning system. With a changing economic outlook, the quality of spending, service delivery, value for money, accountability in the use of resources, and incentives to monitor and evaluate expenditures become central. The traditional dual budget system is hard-pressed to deliver on these and reforms are required to make the annual budget system more strategic and medium-term in its outlook. 61. A shifting economic outlook requires different fiscal outcomes, and these require changing budget behavior through altering budget incentives, and through a culture change in how public finances are managed. As circumstances change, the outcomes associated with a given set of policies are altered. Policies which are initially growth inducing gradually become less effective in stimulating further growth. This has happened with the expanding role of government and with high rates of public investment. The future growth of the Botswana economy depends on how the policies and institutions pursued in the past are altered to meet changing circumstances. 62. With these challenges in mind, a move towards a system that integrates recurrent and capital spending into programs, and which is driven by policy rather than by projects is warranted. Debate has been unfolding in Botswana on whether and how far the planning and budget system should change to accommodate new economic and social circumstances. Botswana can no longer look forward to ever growing mineral revenues, as diamond production peaks, costs rise and profits fall. This transition will start during NDP 10. The central challenge is therefore to move to a system of planning and budgeting that is most likely to help the government address the challenges of diminishing diamond revenues over the medium term and the need to diversify the economy and seek growth in new directions. Chapter 2: Diamonds and Medium-term Fiscal Sustainability 1. This chapter analyzes Botswana’s medium-term fiscal sustainability in view of the expected declining revenues from diamonds in the future. The analysis shows that in the absence of policy correction, Botswana’s current fiscal policy strategy is unlikely to be sustainable over the longer term, regardless of the evolution or impact of the global crisis. There is a danger that the status quo fiscal policy would lead to outcomes that could threaten Botswana’s macroeconomic stability. However, Botswana can improve its fiscal prospects significantly by expanding domestic revenue sources to partially offset the falling mining revenue. It should be noted that while revenue measures are needed, it will be virtually impossible for Botswana to attain the level of revenues associated with diamonds through other means. The extremely high profitability of diamonds (estimated at 80 percent of income in Botswana) simply cannot be matched by other sources of revenue. An adjustment to an economy where government spending is a lower share of GDP is therefore needed. Reducing expenditures to a level commensurate with long-run projections of diamond revenues, whilst adopting a strategy that facilitates non-diamond export growth through a restructuring of planning, budgeting and expenditures are the other components of a strategy to maintain the economy on a sustainable fiscal path. While the details of this strategy are beyond the scope of this chapter, the growth of a dynamic non-mining sector is crucial for Botswana not only from the fiscal sustainability point of view, but from the point of view of achieving balanced development that will create jobs and deliver a durable reduction in poverty and inequality. Fiscal policy will have to play a central role in this process. I. Introduction 2. Since its independence from Britain in 1966, Botswana has achieved impressive economic growth. Over the past four decades, Botswana’s growth rate averaged 9 percent a year, while per capita GDP rose from US$250 in 1960 to US$4,800 in 2008 (in constant 2000 US dollar). Botswana has transformed itself from one of the poorest countries in the world to an upper middle income country. The major driver of Botswana’s remarkable GDP growth has been the abundance of mineral resources, in particular diamonds. Since the mid-1970s, Botswana’s economy has been heavily dependent on diamonds. In recent years diamond exports have accounted for 70 percent of total merchandise exports, while nearly half of fiscal revenue has come from the diamond sector. This has allowed large external and fiscal surpluses. As of end-2008, Botswana’s international reserves stood at over US$9 billion (or 20 months of imports). About 40 percent of this amount represents the accumulation of past fiscal surpluses. 3. Although diamond production in Botswana is expected to continue at high levels for another decade or so, the prospects of mineral revenue accruing to the government are very uncertain. Extraction costs are expected to rise, squeezing profits and government revenues. In the absence of significant new diamond discovery (or alternative source of revenue that offsets falling mining revenues), Botswana’s fiscal revenues are set to shrink sharply from 2021 onwards. The revenue outlook is further shadowed by the likely decline in customs revenues, as SACU revenues fall in line with expected continued trade liberalization as well as expectations of more subdued import consumption in South Africa. Against this backdrop, fiscal policy has become a central policy issue in Botswana. Policy discussions have intensified particularly at the onset of the global crisis in late 2008, when global demand for diamonds plummeted, triggering a sharp fall in its exports, an output contraction, and a large revenue loss from the diamond sector. 4. This paper analyzes Botswana’s medium-term fiscal sustainability in view of the expected depletion of diamond reserves in the next two decades. Four questions are addressed in this chapter: (i) Is Botswana’s current fiscal strategy sustainable in the longer term? (ii) If not, what could be an alternative fiscal strategy for Botswana? (iii) Is the existing fiscal rule appropriate to ensure fiscal sustainability? If not, what alternative fiscal rule should Botswana adopt? (iv) how can Botswana best use the substantial fiscal savings that it has accumulated over time? 5. The structure of the chapter is as follows. Section 2 reviews the main features of Botswana’s diamond mining and discusses the main impacts on fiscal operations to date. Section 3 empirically examines fiscal sustainability in Botswana through medium-term projections based on the financial programming framework used by the IMF; sensitivity analysis is also carried out to examine the robustness of the fiscal paths projected in the baseline scenario by varying key assumptions. Section 4 considers alternative fiscal strategies for Botswana with a view to identifying appropriate policy measures that would bring Botswana’s fiscal stance to a sustainable path. Section 5 concludes with a discussion of some policy options available to government. II. Botswana and Diamonds 6. Diamonds were discovered in Botswana in 1967, a year after its independence from Britain. Since the mid-1970s, mineral revenues have been the backbone of Botswana’s economy, with diamonds being the main contributor to exports, GDP and government revenue.27 This section provides some background on the role of diamond mining in the Botswana economy. Diamonds and Botswana Economy 7. Botswana is the world’s leading diamond producer in terms of output value. In recent years, Botswana has constantly exported over 30 million carats of rough diamonds per year for about US$3 billion. Diamond mining in Botswana takes place through Debswana Diamond Company, a 50-50 joint partnership of De Beers and the Government of Botswana. All rough diamonds produced in Botswana are sold to the Diamond Trading Company (DTC)—the sales and distribution arm of the De Beers Group – which sorts, values and sells approximately 40 percent of the world’s rough diamonds by value.28 8. Diamonds are characteristically different from other commodities. The most distinctive feature of diamonds is their relative price stability (see Figure 1).29 Much of this price stability rests on the fact that a single company, De Beers, controls a significant proportion of the world’s rough diamond 27 The mining sector began to contribute significantly to the economy with the building of the Selebi-Phikwe copper-nickel mine. 28 Since 2009 a small proportion, about 10 percent, of diamonds produced in Botswana is supplied domestically to facilitate the growth of local diamond cutting and polishing industries. 29 Other differences include the absence of universal world price per carat/gram of diamonds since prices vary widely depending on the stone’s carat, color, clarity and cut; and the low liquidity of diamonds as an investment vehicle due to the lack of liquidity and homogeneity as well as difficulties in grading and hence in pricing the stones. mines and distribution channels for gem-quality diamonds (Keretschmer, 1998). De Beers has used its dominant position to control the international diamond market. Nevertheless, with the emergence of new diamond producers such as Russia, Canada and Australia, which chose to distribute diamonds outside De Beers’ channel, the company’s dominant market position has weakened considerably in recent years (Nocera, 2008). Since the turn of the century, the diamond industry is said to have gone through a major transformation from a supply-controlled industry to that which is increasingly driven by demand. 9. The relative absence of the extreme boom and bust cycles in diamond mining has helped Botswana achieve rapid economic growth.30 Since production in Jwaneng mine reached its full capacity in 1985, Botswana’s real output has more than tripled, with per capita income growing more than five times in Pula terms.31 The mining sector—where the bulk of value added is generated by diamonds—has been the largest contributor to the country’s GDP, although recently its share has fallen to 35-40 percent (Table 1). Although from a very low base, output of other minerals, such as copper/nickel, soda ash and gold, has grown considerably over the past 5 years, their contribution to total exports vary significantly year by year (10-20 percent), owing to the swings in international commodity prices. Table 2. Botswana: Diamonds’ Contribution to Economy 1982 1990 2000 2006 Mining GDP in percent of total GDP (%) 28.7 42.8 44.3 40.8 Diamond exports (in millions of US$) 246.0 1,412.1 2,228.0 3,349.4 In percent of total exports (%) 53.0 78.2 83.1 63.0 Gross international reserves (in millions of US$) 276.1 3,331.5 6,318.2 7,992.4 In months of imports (months) 6.0 20.5 30.7 31.4 Sources: IMF IFS and Department of Mines. Figure 1. Commodity Price Indices (1990 = 100) /1 30 Although Botswana has experienced some output fluctuations over time (1981/82 and early 1990s), the magnitude of these fluctuations were smaller compared with those experienced by other resource dependent countries. 31 It was not only diamond mining but the sense of corporate social responsibility that De Beers brought to Botswana. It has built roads, hospitals and schools in Botswana, worked to help the country deal with HIV/AIDS and been involved in and paid for many other things that have helped make Botswana an African success story (Nocera, 2008). 400 350 Aluminum ($/t) Coal, Australia ($/mt) Copper ($/mt) Crude oil avg spot ($/b) Gold ($/toz) Iron ore (¢/dmtu) 300 Lead (¢/kg) Nickel ($/mt) Steel cr coilsheet ($/mt) Zinc (¢/kg) BW rough diamond ($/ct) 250 200 150 100 50 0 Sources: World Bank GEP database and Botswana Department of Mines Annual Reports (various issues). 1/ The price of rough diamonds is derived from the value of Botswana’s diamond exports and volume of production (carats recovered). See footnote Error! Bookmark not defined.. 10. While diamond mining has contributed immensely to Botswana’s economic growth, the large revenue stream from the mining sector has also allowed the creation and maintenance of a large government. Today the government sector accounts for 15 percent of GDP (over 25 percent of non-mining GDP) and hires 40 percent of formally employed labor force. The economy is directly and indirectly dependent on government spending. Therefore fiscal operations have a significant bearing on Botswana’s domestic demand. 32 2.2 Diamonds and Botswana’s Fiscal Policy 2.2.1 Revenue 11. Revenue from the mining sector makes up about 45-50 percent of total government revenue. The Agreement governing revenue-sharing between Debswana and the government is confidential. Using diamond exports data as a rough approximation of Debswana’s profit it can be surmised that diamond revenues accruing to government coffers have been on the order of 60-65 percent of diamond exports since 2000.33 Revenue collection from other minerals is negligible, with nearly 99 percent of mineral revenue coming from the diamond sector, according to the Department of Mines. As such, throughout this chapter mineral revenue refers to the government’s share of diamond mining. 32 Note, however, the government sector’s contribution to national income needs to be interpreted carefully. As directed by the UN System of National Accounts (SNA), a government’s output is valued at cost. Therefore, expanding the cost of government activities, and in particular, the cost of its employment will show up uncritically as increased output in the national accounts, regardless of whether this output is of any real value. 33 When the original agreements over Orapa and Jwaneng were negotiated in the 1970s, the split, on paper, was 70/30 in Botswana’s favor; nevertheless, operational factors mean that De Beers would recoup some of this. 12. The government also receives transfers from the Bank of Botswana (BoB), which derive from the investment return on the government’s savings (past fiscal surpluses). These are held in the BoB and invested in long-term assets, as well as dividends from the BoB.34 Since the bulk of the BoB transfer is accounted for by the investment return on the fiscal reserves, made possible by large mineral revenue, the IMF defines the BoB transfer as part of the government’s total mineral revenue. To avoid confusion, this report follows the government’s definition of mineral revenue, which comprises mineral taxes, royalties and dividends (see Figure 2). 13. In line with the decline in mineral revenues, total government revenue has fallen over time (Figure 2). Domestic revenue—revenue excluding mineral, SACU, BoB transfer and foreign grants—has grown recently following the introduction of VAT (2002) and improvements in tax compliance and collection (2003/04).35 However, its contribution to overall revenue is still low, accounting for only about 30 percent (10 percent of GDP) of total revenues. With 70 percent of revenue outside the control of the government, Botswana’s revenue structure is vulnerable to external conditions. For example, in 1998/99, when the Asian crisis led to a drop in diamond exports, government revenue plunged by 10 percent of GDP, turning Botswana’s fiscal position into a deficit for the first time since 1982/83. 34 According to the BoB Annual Report (1996), the BoB transfer to the government is calculated on the basis of the expected returns from the government’s portion of the Pula Fund (i.e., Government Investment Account, see Box 1), estimated in SDR terms over a five to seven year period. The actual payment to the government is made on every quarter during the fiscal year, and recorded as government revenue. In addition, the BoB’s net income for its financial year, after accounting for administrative and interest expenses, as well as any appropriation to the BoB’s General Reserve, are paid annually to the government at the end of the fiscal year (March 31). 35 The VAT rate was increased from 10 to 12 percent, effective April 2010. Figure 2. Government Revenue Excluding Grants (in percent of GDP) 70 Non-mineral, non- 60 SACU revenue (domestic revenue) 50 40 SACU revenue 30 BoB revenue 20 Mineral revenue Mineral tax Mineral revenue 10 by government royalties & by IMF definition definition dividends 0 Source: Botswana authorities. Figure 3. Botswana Diamonds: Production Volume and Average Price 600 40 Volume of production Moving average real price (right scale) 35 500 per carat (left scale) /1 Volume of production (million carats) 30 Price per carat (in 2005 Pula) 400 25 300 20 15 200 10 100 5 0 0 Source: Department of Mines Annual Reports (various issues) and Bank Staff calculations. 1/ 2-year moving averages of the price of rough diamonds, estimated from the volume of production and value of diamond exports. See footnote Error! Bookmark not defined.. Table 3. Central Government Operations (in percent of GDP, unless otherwise indicated) 2003/04 2004/05 2005/06 2006/07 2007/08 2008/09 /1 Revenue and Grants 38.4 36.5 40.9 40.2 35.9 34.8 Revenue 38.3 35.8 40.7 39.5 35.2 34.1 Tax Revenue 19.8 20.5 22.0 23.3 21.7 23.4 Mineral Tax 5.6 5.1 5.3 5.6 4.7 3.9 Non-Mineral Income Tax 4.9 4.2 5.5 4.5 3.2 5.3 Customs Pool 5.3 6.6 7.2 9.7 9.8 8.9 Export Duties 0.0 0.0 0.0 0.0 0.0 0.0 Sales Tax/VAT 3.7 4.3 3.6 3.3 3.6 5.0 Other taxes 0.2 0.3 0.3 0.3 0.3 0.3 Non Tax Revenue 18.4 15.3 18.7 16.2 13.5 10.7 Mineral Royalties and Dividends 13.7 12.6 15.0 13.6 10.7 7.7 Interest receipts 0.5 -0.2 0.2 0.1 0.1 0.1 Other Property Income, of which 2.3 0.9 1.7 0.9 1.3 1.6 BOB revenue 1.8 0.8 1.4 0.8 1.1 1.5 Fees and charges 1.9 2.1 1.9 1.5 1.4 1.3 Grants 0.1 0.7 0.2 0.7 0.7 0.7 Expenditures and Net Lending 38.6 35.3 32.4 28.9 31.1 40.2 Recurrent 30.7 28.0 26.0 23.4 23.3 27.3 Wage bill 9.8 10.4 9.6 8.5 8.6 9.9 Interest 0.5 0.6 0.6 0.3 0.3 0.3 Domestic debt 0.4 0.6 0.3 0.3 0.2 0.2 External debt 0.1 0.0 0.3 0.1 0.1 0.1 Other recurrent expenditure 20.4 16.9 15.8 14.5 14.4 17.0 Development Expenditure 10.1 8.0 6.9 5.9 8.2 13.1 Net lending -2.2 -0.6 -0.6 -0.4 -0.4 -0.2 Overall Surplus/Deficit (-) -0.2 1.2 8.5 11.2 4.8 -5.4 Primary surplus/deficit 0.3 1.8 9.1 11.6 5.1 -5.0 Financing 0.2 -1.2 -8.5 -11.2 -4.8 5.4 External borrowing, net -0.3 -0.3 -0.3 -0.3 0.0 -0.1 New borrowing 0.1 0.0 0.0 0.0 0.2 0.2 Amortisation (- entry) -0.4 -0.3 -0.3 -0.3 -0.3 -0.3 Domestic borrowing net 4.6 0.0 -1.4 0.0 0.6 1.5 New borrowing 4.6 0.0 -1.4 0.0 1.6 2.2 Amortisation 0.0 0.0 0.0 0.0 -1.1 -0.7 IMF Transactions 0.0 0.0 0.0 -0.1 -0.1 -0.1 Pension Liability Service Fund -8.8 -3.2 -2.1 0.0 0.0 0.0 Other Financing -4.8 -1.0 1.4 -0.5 -0.2 -1.2 Change in Cash Balances 9.4 3.4 -6.2 -10.4 -5.0 5.3 Memorandum items Mineral revenue 19.4 17.7 20.3 19.2 15.5 11.6 Mineral corporate tax 5.6 5.1 5.3 5.6 4.7 3.9 Royalties, dividends 13.7 12.6 15.0 13.6 10.7 7.7 Domestic revenue /2 11.8 10.8 11.8 9.8 8.7 12.1 Non-mineral balance (% of non-mining GDP) -34.1 -28.0 -21.3 -15.4 -20.0 -28.7 Non-min primary balance (% of non-mining GDP) -34.2 -26.7 -20.6 -15.0 -19.6 -28.3 GDP (in millions of Pula) 42,178 49,176 54,457 68,221 79,754 87,458 Non-mining GDP (in millions of Pula) 15,790 18,887 20,866 29,136 32,597 31,154 Sources: Botswana authorities and Bank Staff calculation. 1/ Estimates. 2/ Revenue excluding mineral revenue, SACU revenue, BoB transfers and foreign grants. 2.2.2 Fiscal Rules 14. To prevent excessive spending and ensure fiscal sustainability in anticipation of future diamond depletion, the government has, over time, set various fiscal rules, both formal and informal. First, the “Principle of Sustainable Budgeting” was introduced in 1994 with the intention of ensuring all mineral revenue is invested productively or saved, and not used for consumption.36 The 36 1994/95 is the year when government revenue fell sharply due to low diamond sales and BoB transfers associated with adverse conditions in the international capital markets. adoption of the Sustainable Budget Index (SBI) was an attempt to maintain that non-investment expenditure—defined as all government expenditure excluding development spending and recurrent spending on health and education—should be financed by non-mineral revenue, so as to conserve the country’s wealth and generate other sources of income when diamonds are depleted. Another fiscal “rule”, introduced in 2006 as part of the Mid-term Review of National Development Plan 9 (MTR NDP9), sets the maximum government expenditure at 40 percent of GDP for the Plan period, to be consistent with the projected medium-term government revenue.37,38 The MTR NDP9 also targeted an increase in the share of development spending in the budget to 30 percent by 2008/09. Furthermore, under Section 20 of the Stock, Bonds and Treasury Bills Act (Chapter 56:07, 2005), the government’s total domestic debt and guarantees cannot exceed 20 percent of GDP. Similarly, the stock of government foreign debt and guaranteed is limited to 20 percent of GDP. 15. Adherence to the fiscal rules has enjoyed mixed success. Although the SBI fell below unity immediately after its introduction in 1994/95, the effect has not been long-lasting (Figure 4). The SBI exceeded unity in 2001/02, when exceptionally high SACU revenue during 1999/00-2000/01 came to an end.39 In 2003/04, however, measures to raise domestic revenue, coupled with across the board cuts in all expenditures, resulted in a dramatic fall in the SBI. The more recent 40 percent rule has also helped discipline overall expenditure somewhat. Yet again, in 2008/09, barely three years after the introduction, total government expenditure exceeded 40 percent of GDP marginally, caused by the sharp rise in development expenditure and the negative GDP growth affected by the global financial crisis (see below). 16. On balance, however, Botswana has run a prudent fiscal policy, avoiding many pitfalls experienced by other natural resource-rich countries. It has avoided an excessive accumulation of external debt. Rent-seeking behavior has been kept to a minimum, as reflected in Botswana’s excellent governance indicators. Apart from the authorities’ prudent approach, the relative stability of diamond prices has also played a role in the maintenance of stable fiscal policy. In addition, the government’s capacity constraints to implement development expenditure in full may have also worked to prevent rapid increases in expenditure.40 Botswana maintained a large fiscal surplus for most of the 1980s and 1990s ( 17. 18. Figure 5). These surpluses have been saved in the Government Investment Account (GIA) with the Bank of Botswana as the government’s portion of the Pula Fund (see Box 1), and invested in long- term assets, mostly abroad. 37 The 40 percent of GDP rule also allowed for a breach of the expenditure limit, in case of emergencies where a Presidential declaration proposes expenditure measures to alleviate the hardships. The MTR NDP9 (para. 122) also notes the possibility of revenue side emergencies. Should a serious sustained shortfall on any major revenue source appear imminent, an emergency declaration by the President would be warranted, specifying the foreseeable picture and propose a new expenditure program to bring government expenditure as a share of GDP into line. 38 NDP 10 states that the average level of public spending needs to be reduced to less than 30 by NDP12 (2022- 2029) percent of GDP, although because of the decision to implement a fiscal stimulus to counteract the effects of the global crisis on the economy, this cannot be implemented for each year of the Plan, but is the average for the 7- year Plan period. 39 SACU revenue was exceptionally high during 1999/00-2000/01 because of the rapid depreciation of the South African Rand. 40 In Botswana, actual development spending has been consistently below budgeted development expenditures due to capacity constraints to implement projects. Figure 4. Botswana: Sustainable Budget Index (SBI) /1 1.2 Introduction of SBI 1.1 1.0 0.9 0.8 0.7 0.6 0.5 Source: Bank Staff calculation based on data provided by the authorities. 1/ Ratio of recurrent expenditure (excluding education and health) to non-mining revenue excluding grants and sales of assets. 19. Nonetheless, with the declining mineral revenue, the occurrence of fiscal deficits has become more frequent in the last decade. Government financed the deficits primarily by drawing down on the GIA. Foreign and domestic borrowing has played a very limited role to close the fiscal gap. Issuance of government securities, which began in 2002/03, was intended largely for the development of local capital markets. At end-2007/08, external debt amounted to 2.9 percent of GDP, whereas domestic debt was 2.8 percent. Both of them were well below the statutory limits. Figure 6 shows Botswana’s fiscal operations as measured by non-mining GDP, a conventional measure for resource dependent countries. A combination of overall fiscal surpluses, non-resource deficits and a large government characterizes the budgets of major resource dependent countries.41 Figure 5. Botswana: Fiscal Operation (in percent of GDP) 41 Resource abundance is also often associated with poor development of human capital, as a low degree of economic diversification provides few opportunities for leaning and doing (Söderling, 2005). 70 60 Fiscal balance Primary balance Revenue and Grants 50 Expenditures and Net Lending 40 30 20 10 0 -10 Source: Botswana authorities. Figure 6. Botswana: Non-mining Fiscal Operation (in percent of non-mining GDP) 160 Non-mineral fiscal balance 140 Non-mineral primary balance 120 Non-mineral revenue Expenditure and net lending 100 80 60 40 20 0 -20 -40 -60 -80 -100 Source: Botswana authorities. Box 1. Pula Fund The Pula Fund was re-established in its current form under the new Bank of Botswana Act (1996) to provide greater flexibility in the management of international reserves, and greater certainty in the forecasting of annual “dividend” payments to the government from the BoB.42 The Act came into operation on January 1, 1997. Under the Act, Botswana’s international reserves are split into two portfolios: (i) the Liquidity Portfolio, to provide the foreign exchange needed for normal day to day international 42 The Pula Fund was originally established in November 1993 under the Bank of Botswana Act (1994). transactions; and (ii) the Pula Fund, to be invested in long-term assets to achieve higher returns.43 The Pula Fund, managed by the BoB, is composed of the Government Investment Account (GIA), which reflects savings from accumulated fiscal surpluses, and the BoB’s reserve accumulation above the target for liquid reserves (see the chart below). When import cover falls below six month, resources are transferred into the liquidity portfolio, and vice versa. Pula Fund assets are invested in long-term instruments overseas. The Act charges the BoB with managing and determining investment policy of, and the payment of dividends accruing from, the Pula Fund, in consultation with the Minister of Finance and Development Planning. Together with the Liquidity Portfolio, the Pula Fund has been held with a global custodian since 1993. While the Act provides a legal framework for the establishment, management, and auditing of the Fund, it does not specify the objective of the Fund in the context of overall fiscal policy and rules on the operation of the Pula Fund, particularly concerning payments into the Fund, withdrawals from it, and their uses. Although the original idea behind the establishment of the Pula Fund was to invest in long-term offshore assets those financial resources that could not be absorbed domestically for productive purposes (Mohohlo, 1997), there are no other laws, constitutions, regulations nor guidelines that explicitly specify the link between the Pula Fund and fiscal policy. By and large, the Pula Fund has served as a revenue stabilization fund, rather than an investment fund, but also has been regarded by many as a future generations fund, despite the absence of clearly defined objectives as such.44 International Reserves Liquidity Assets at Pula Fund Portfolio IMF Government Investment BoB Portion Account Pula Fund: Stock and Investment Return 43 The Pula Fund is officially referred to as “long term investment funds” in the Act. 44 For example, “2009 Budget Review” by Econsult views the accumulated fiscal surpluses as part of the inheritance of future generations (Econsult, 2009). 60 30.0 GIA (left) 25.0 BoB portion (left) 50 Implied rate of return (in percent p.a.) Pula Fund (in billions of nominal Pula) Nominal rate of return (right) 20.0 Real rate of return (right) 40 15.0 10.0 30 5.0 20 0.0 -5.0 10 -10.0 0 -15.0 Sources: Bank of Botswana Annual Reports (various issues) and Bank Staff calculation. 1/ The implied rate of return is derived from the balance of the Pula Fund and gross investment income (before administration charges). 2.3 Diamonds and Future Challenges 20. The recent global economic crisis painfully exposed Botswana’s heavy dependence on diamonds as a source of growth, exports and fiscal revenue. With the advent of the crisis in September 2008, global demand for commodities collapsed almost immediately, prompting a sharp fall in their trading. Global demand for luxury goods such as diamonds was hit even harder. In November 2008 Botswana’s diamond exports came to an abrupt halt. Diamond exports fell to virtually zero in the last quarter of 2008, forcing Debswana to temporarily close of all four of its diamond mines for nearly four months. Botswana’s mining GDP plunged by 68 percent in the first quarter of 2009 year on year. 21. Government revenue suffered as a result. In 2008/09, mineral revenue declined by 27 percent in real terms from the previous year. In contrast, public expenditure recorded a sharp increase. This was due in part to 2008/09 being an election year. Current expenditure rose by 13.5 percent from the previous fiscal year, reflecting a further increase in the public sector wage bill, whereas development expenditure grew by more than 50 percent as the government accelerated the construction of airports, roads, dams, and schools, and embarked on the Morupule B power project. The fiscal outcome of 2008/09 was a deficit of 5.4 percent of GDP, the bulk of which was financed by drawing on the fiscal reserves.45 22. Following on signs of market improvements, Debswana resumed operation in two of four mines in April 2009, and another was re-opened later in the year. Diamond production in 2009 is estimated to have reached 20 million carats, about 60 percent of a pre-crisis production level of 33 million 45 Based on the data provided by the MFDP in November 2009. carats. Diamond exports are also estimated to have been reduced by 35-40 percent in 2009 to US$2.5 billion. The corresponding losses in output and government revenue will be more acutely felt in 2009/10 than in 2008/09. The latest estimates suggest that Botswana’s output could experience zero growth this fiscal year (2009/10), whereas the fiscal deficit is expected to exceed 15 percent of GDP.46 In May 2009, Botswana secured budget support in the amount of US$1.5 billion from the African Development Bank. In some way, the global crisis has served as a reminder of some future reality when Botswana runs out of diamonds. 23. The outlook for diamond exports is highly uncertain, both in the short and medium terms. While the global economy is pulling out of a recession, stabilization is uneven and the recovery is likely to be sluggish. Albeit slightly better than the early predictions, global economic activity is still estimated to have contracted by 1.4 percent in 2009 (IMF, 2009). A sustained pick-up in advanced economies, where the bulk of demand for diamonds comes from, is unlikely until the second half of 2010 at the earliest. The prospects for the medium term are even more uncertain. 24. Although diamond production in Botswana is expected to continue at fairly high levels for another decade – subject to a recovery in demand – revenues accruing to the government are projected to stop growing beginning in 2015/16. This will happen before the actual depletion of diamond starts and mining moves from low-cost surface mining to underground. Policy makers in Botswana have focused on the projected dates for the beginning of the decline in diamond revenues and the depletion of diamonds. Yet the key date is closer at hand: the time when the growth of diamond revenues peaks; this is currently projected to happen within five years, i.e. starting in 2015/16. At that time government spending will come under pressure and real spending cuts will be required to maintain fiscal balance. Production may continue until the anticipated depletion of reserves around 2030, but profitability and economic feasibility of underground mining are uncertain. The revenue outlook is further shadowed by the likely decline in customs revenues from the SACU pool.47 The time to worry about fiscal sustainability is now. 3. Fiscal Sustainability Analysis 25. Concerned about uncertainties regarding the future of diamonds, and especially, in the advent of the global crisis, policy discussions have intensified whether the government needs to change its fiscal policy course. Is Botswana’s current policy stance sustainable in long run? If not, what 46 The 2009 Budget presented in February 2009 proposed a fiscal expansion despite a significantly lower revenue projection. Both Moody’s and Standard & Poor’s cut Botswana’s foreign currency rating on account of a possible sharp rise in public debt and lower revenue. In response to the rating companies’ negative reaction, the government quickly announced an across the board cut of 7 percent for current expenditure and 5 percent for development expenditure. 47 On the upside, Debswana has begun works on the Jwaneng “Cut 8” project which refers to the current open pit mine being expanded to facilitate access to the kimberlite ore at a greater depth, thus postponing the need to go underground. This expansion includes relocating and rebuilding portions of existing surface infrastructure that are impacted by the pit expansion. Debswana hopes that this project would extend the life of the mine by ten years although this is uncertain given lack of complete information regarding the shape of the kimberlite pipe and the quantity and quality of diamonds. A factor that may lead to higher diamond prices is the fact that there have been no large diamond finds in nearly two decades. Perceptions of limited supply may increase the buoyancy of future prices. policy actions are needed? This section carries out medium-term fiscal sustainability analysis for Botswana. With the prospects of declining mineral revenue as diamond reserves are depleted, it is important that Botswana adopts a forward looking approach, and formulates a good strategy. Determining whether the current fiscal policy is sustainable is important for policymakers because the answer may indicate the need for policy correction. 3.1 Operational Definition of Fiscal Sustainability 26. Before analyzing Botswana’s medium term fiscal sustainability, it would be useful to define what is meant by “sustainable fiscal policy”. For this purpose this report draws on the standard intertemporal sustainability framework, but takes a very operational approach of assessing medium-term fiscal sustainability. Specifically, fiscal policies are sustainable if:48 • If a country is expected to be able to continue servicing its debt without an unrealistically large future correction to the balance of revenue and expenditure, or without defaulting on its debt.49 • A reasonable level of external shocks does not bring a country into debt distress. 3.2 Fiscal Sustainability Analysis for Botswana 27. Fiscal sustainability analysis for Botswana is carried out in the following manner. First, using a forward-looking accounting framework, a baseline fiscal scenario based on a set of plausible macroeconomic assumptions is constructed, assuming that the policy strategy described in NDP10 is maintained.50 The projected medium-term fiscal scenario – the path of government revenue, expenditure, fiscal balance, debt-to-GDP ratio, debt service profile, etc – is then be assessed against the operational sustainability conditions described above. An increasing debt ratio is generally regarded as a cause for concern, because it will require policy adjustment sometime in the future. 28. Second, in order to assess the robustness of the baseline fiscal paths, sensitivity analysis is performed by varying the underlying assumptions, such as the GDP growth rate, exchange rate and 48 Technically, a country’s fiscal policy is defined sustainable if ex post it satisfies the solvency condition expressed as: ∞ PBt +τ D0 = ∑ τ =0 (1 + r )τ where D denotes the stock of public debt, PB the primary balance, and r the interest rate on public debt (assumed constant), all expressed in real terms. Monetization of the deficit is not considered in this framework in that all fiscal adjustments take place through the level and composition of government revenue and primary expenditure. 49 See IMF (2002). 50 The accounting framework employed in this paper is analogous to the “Fiscal” module, which comprises one of the five modules in the financial program used by IMF economists. The financial programming framework is deterministic. It is not a fully-specified macroeconomic model, and thus does not directly capture the behavioral relationships (e.g., secondary responses to policies, such as the impact of fiscal policy on growth and inflation). While such behavioral relationships can be estimated by a computable general equilibrium (CGE) model, the operational application of this approach is extremely problematic for Botswana given the relatively limited availability of macro- and microeconomic data for the identification and estimation of parameters with reasonable precision. cost of borrowing. The sensitivity analysis can be interpreted as introducing uncertainties into the deterministic framework. Alternative scenarios that incorporate different policy strategies will be examined in the following section. 29. Fiscal sustainability analysis for Botswana considers only central government operations.51 In the absence of information, it does not include contingent liabilities of the government. For the same reason, quasi-fiscal operations other than the revenue impact of the central bank’s sterilization operations are not incorporated. The interest expenses associated with sterilization are captured in the budget as transfers from the BoB, which derive from the investment return on the fiscal savings as well as the BoB’s net income (which includes the cost of sterilization).52 30. The projection period adopted in this paper is 20 years, with 2009/10 as a starting year (Period 1). While medium-term scenarios generally consider a time horizon of 5-10 years, the prospect of a significant change in Botswana’s economic conditions argues for an extended projection period. This will allow policymakers to identify potential future problems and prepare an appropriate policy strategy well in advance, although projections of more distant years are less precise and subject to larger uncertainties. 3.2.1 Medium-term Fiscal Projections: Baseline Scenario Macroeconomic Assumptions 31. Our Baseline Scenario takes the NDP10 macroeconomic assumptions and policy strategies, and adopts the 2009/10 Budget as the starting point of the medium-term fiscal projections.53 For Botswana, the backbone of the macro-fiscal framework is the diamond sector’s performance. Following NDP10, the diamond sector is assumed to make a small recovery in 2010/11, after a sharp contraction in 2009/10. A full recovery of diamond production is then assumed to take place over 2011/12—2013/14. However, in 2015/16, mining sector growth slows significantly. The production profile and price of diamonds are then calibrated to be consistent with the mining growth rates and information obtained from Debswana. The non-mining sector is conservatively assumed to grow gradually from 3 percent in 2009/10 to 5.5 percent in 2015/16.54 The slow growth of the non-mining sector is due in part to the government’s plan to reduce its level of spending in real terms during the Plan period. The baseline assumptions on inflation and the exchange rate are taken from the IMF World Economic Outlook (October 2009).55 The 7-year 51 The government has several special funds. The consolidated budget table (Table V) reports changes in their cash balance as a financing item, and thus they are incorporated in our fiscal framework. 52 The stock of Bank of Botswana Certificates (BoBCs) is not included in the projection of medium-/long-term government domestic debt, due to difficulties in projecting the magnitude of sterilization. 53 At the time of writing the Bank had access to the draft NDP 10 macroeconomic chapter. The final report should not see significantly different assumptions. 54 It is not clear if the non-mining growth rates used in the Plan incorporate the impact of the ongoing power project (Molupule B). 55 NDP10 does not provide the exchange rate and inflation assumptions. Plan period is then extended till 2028/29 (Period 20) taking the 2015/16 assumptions, except for the growth rate of mining and non-mining sectors. The medium-term mining growth rates are calibrated assuming a certain path of diamond output and profit per carat, taking into account projected increases in mining costs and production profile as described in NDP10. As regards non-mining GDP, it is assumed that its average annual growth rate during NDP10, about 4 percent, will continue during the rest of the projection period. This is a customary practice adopted in medium-term projections. Table 4 provides the summary of baseline macroeconomic assumptions. Table 4. Baseline Macroeconomic Assumptions 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 Real GDP growth (annual percentage change) -10.7 3.2 10.6 12.2 14.5 6.0 -2.6 Real mining GDP growth (annual percentage -35.5 4.7 31.0 29.5 31.1 6.8 -14.2 change) Real non-mining GDP growth (annual 3.0 2.7 3.4 4.5 5.3 5.4 5.5 percentage change) Inflation (period average, in percent p.a.) 7.9 6.3 5.8 5.3 5.2 6.0 6.0 Exchange rate (Pula/US$) 7.4 7.7 8.2 8.6 9.0 9.5 9.5 Source: NDP10 and IMF WEO (October 2009). Fiscal Policy Strategy 32. The Baseline Scenario assumes no major policy changes throughout the 20-year projection period, except for the following. In line with the 2009/10 Budget statement, the Baseline Scenario assumes that in 2009/10 the government creates 1,610 additional positions and makes no cost of living adjustment for public service wages. For the remainder of the Plan, the government will allow no net increase in its establishment so as to limit the growth of the government wage bill.56 Other recurrent expenditures, i.e., recurrent expenditure excluding the wage bill and interest payments, are assumed to grow for the maintenance and replacement of existing infrastructure, at the rates indicated in NDP 10.57 Development expenditure is front-loaded for critical investment projects such as the Morupule B power project, and is assumed to decline according to the rates indicated in the Plan.58 From 2016/17 onwards, both other recurrent and development expenditures are assumed to grow at the rate of inflation. 33. As per NDP10, it is assumed that no specific revenue measures will be introduced during the Plan and in the medium term.59 Where fiscal deficits emerge, the fiscal gap will be filled by a combination of domestic and external borrowing as well as running down of the GIA balance. It is assumed that, until the statutory debt limits are reached, the government will prefer cheaper external borrowing to more expensive domestic borrowing for deficit financing, and will also prefer domestic borrowing to drawing down of fiscal savings so as to maintain the balance of the GIA as far as possible. For 2009/10, the Baseline Scenario adopts the borrowing plan provided by the Ministry of Finance and Development Planning (MFDP) and assumes that the government will source US$2.1 billion externally 56 NDP10, para. 6.51. 57 Table 6.6. Note that for recurrent spending this analysis deviates from NDP10 assumptions by projecting the wage bill, interest payments and other recurrent expenditure separately. This will allow us to monitor how each variable would evolve in response to changes in key assumptions and policies. 58 Table 6.7 NDP 10. 59 Since the preparation of the sustainability analysis, the 2010/11 budget speech introduced an increase in the VAT rate from 10 to 12 percent, beginning in April 2010. and P 1.5 billion from domestic sources to finance the budget deficit (Table 5). See Appendix 1 for further details on the assumptions and calibration. Table 5. Anticipated New External Borrowing in 2009/10 Loan amount (in millions Creditor Purpose of US dollar) AfDB Budget support 1,500 OFID Kasane Airport 10 BADEA Kasane Airport 8 OFID International Transport 40 IBRD International Transport 186 AfDB n.a. 244 WD n.a. 153 Total 2,141 Source: MFDP Projection Results 34. The Baseline projection results are presented in Table 6. For immediate policymaking purposes, the focus will first be on the details of the projection over the 7-year Plan period. Later, medium-term sustainability issues will be considered. 35. Revenue performance is projected to remain weak during the Plan. Overall government revenue is likely to stay at around 30 percent of GDP, compared with the NDP9 average of 38 percent. Although mineral revenue would return to the pre-crisis level as the diamond sector recovers from the effect of the global crisis, the recovery in mineral revenue is likely to be offset by declines in SACU revenue. SACU revenue is projected to decline not only because of temporary factors – the negative impact of the global recession – but also because of a likely permanent shifts such as further trade liberalization.60 36. Government spending would be high in the first two years, because of the front-loading of development expenditure and O&M spending, with overall expenditure projected to exceed 40 percent of GDP, breaching the fiscal rule. The higher wage bill, due to the hiring of additional 1,610 personnel, would also contribute to the high spending in the early years of the Plan. With the strict enforcement of the no net personnel increase policy, however, the government wage bill could fall sharply to below 7 percent of GDP, provided that wage increases are limited to cost of living adjustments at the rate of inflation. As large infrastructure projects phase out, development expenditure would fall in outer years of the Plan. Total expenditure is projected to decline sharply to below 30 percent of GDP during 2013/14—2015/16. 37. Reflecting the combination of low revenue and very high spending, the fiscal outcome in the first two years would be a significant deficit, of over 10 percent of GDP. Botswana has never run a fiscal deficit of this magnitude. The deficit in 2009/10 (13.1 percent of GDP) would be more than financed by external and domestic borrowing, allowing the government to save the remainder (5 percent 60 For example, the South African authorities have revised the SACU revenue pool forecast downwards by 20 percent for 2010/11 after the global recession, considering the adverse impacts on South Africa’s imports, especially automobiles, which account for about a third of total SACU revenue. of GDP) in the GIA. However, the new external borrowing in 2009/10 is likely to bring public external debt to over the legal limit of 20 percent of GDP. Hence, the projected large fiscal deficit in following year (12.6 percent of GDP) would need to be financed by a combination of domestic borrowing and drawing down of the fiscal savings. 38. The fiscal position would improve only later in the Plan, if indeed expenditure falls as planned. Fiscal surpluses are projected in the last three years, but they are not large enough to offset the earlier deficits. The cumulative fiscal deficit during the Plan would exceed P 22 billion (equivalent to 25.5 percent of 2008/09 GDP), in sharp contrast to the cumulative surplus of P 14 billion in the NDP9 period, both expressed in 2008/09 Pula.61 The balance of the GIA would fall from about 31.6 percent of GDP at end-2008/09 to 25 percent by the end of the Plan. Next the longer term implications of the Baseline Scenario are considered. Medium-term Sustainability Assessment 39. The projection results suggest that in the absence of decisive policy actions Botswana’s current fiscal stance is unsustainable over the longer term (see Panel 1). There is a real danger that the status quo fiscal policy would lead to unsustainable outcomes that may threaten Botswana’s macroeconomic stability. Unless alternative sources of revenue are found to offset the declining mineral revenue, overall revenue would continue to fall, to a level far below what underpins the 40 percent fiscal rule. As government revenue falls, the fiscal improvement projected later in the NDP10 period would be reversed in 2016/17 (Period 8). Thereafter, the fiscal deficit – a structural deficit – would continue to grow. At first, the government could finance the deficit by borrowing externally and domestically. However, once the statutory debt limits—20 percent of GDP for each of domestic and external debt – are reached, the government would have no option but to draw heavily on the fiscal reserves to finance the deficits. As the fiscal deficit grows and the government continues to run down the fiscal reserves, eventually the GIA would be depleted around 2024/25 (Period 16), at about the same time as what is currently projected to be the beginning of the rapid decline in diamond production. 40. At this point, the government would be left with virtually no means of deficit financing under the existing legal framework, since the Bank of Botswana Act (1996) limits central bank financing of the deficit to 5 percent of government average revenue for the preceding three fiscal years (up to 10 percent when the limit is temporarily waived). While this may be a remote possibility in Botswana, governments in a similar situation have in the past resorted to (i) extensive use of central bank advances by effectively rolling over the short-term facility; (ii) creative interpretation of legislation concerning the government’s borrowing limits; and/or (iii) accumulation of arrears (salary, pension, supplier arrears as well as repudiation of contractual obligations) to finance the deficit.62 61 The cumulative deficit projected here is lower than that in NDP10 (P 33 billion). This is because we project the path of the wage bill, interest and other recurrent expenditures separately. 62 Section 20 of the Stock, Bonds and Treasury Act (1995) is subject to a number of different interpretations with respect to “government debt”. For example, it does not specify the level of government (central government only or inclusive of local authorities, statutory or wholly-owned government corporation, government agency, etc) or the composition of “debt” (inclusive of concessional and/or short-term debt, or gross or net terms). 41. According to this Baseline Scenario, the government would fall into a net debtor position in the second year of the Plan (2010/11), that is, the value of its assets (GIA balance and outstanding balance of on-lending through PDF, RSF) is smaller than the value of its liabilities (public external and domestic debts). Thereafter, Botswana is unlikely to recover the net creditor position. By the end of the projection period, the government’s net debt would exceed 70 percent of GDP. The standard debt indicators, the ratios of net present value (NPV) of net external debt service to GDP, revenue and exports, are all rising rapidly, albeit from a low base.63 This means that the possibility that Botswana could fall into a serious debt distress sometime in the future cannot be ruled out. 42. It is important emphasize here that Botswana’s fiscal vulnerability identified above is not brought by the global crisis that started in late 2008. Rather, Botswana’s fiscal vulnerability rests on its dependence on a non-renewable natural resource, which is expected to be depleted in the near future. 43. It should be noted that the size of the projected results is for indicative purposes, especially in the distant future because of significant uncertainties surrounding the macroeconomic assumptions. The uncertainties are most significant in terms of the prospects for diamonds, but also because the accounting framework employed here cannot directly capture various behavioral relationships. In terms of the latter, for instance, the large investment spending in the initial years of the Plan may boost private sector growth in the NDP11 period, raising non-mineral tax revenue and improving fiscal prospects. Alternatively, as Botswana continues to run large fiscal deficits, external creditors may revise Botswana’s country risks downwards. External borrowing costs may rise, putting upward pressure on the budget. In the case of domestic borrowing, private sector growth may be severely hampered if government’s borrowing led to crowding out of private investment. The next subsection addresses these drawbacks by performing sensitivity analysis. 63 The Debt Sustainability Framework (DSF) used in the World Bank and IMF relies on a set of empirically derived policy-dependent debt thresholds for assessing a low income country’s debt sustainability. For countries with strong policies and institutions, the relevant indicative thresholds are: 50 percent for the NPV of external debt-to-GDP ratio; 300 percent of the NPV of external debt-to-revenue ratio; and 200 percent for the NPV of debt-to-exports ratio. These thresholds are applicable to public and publicly guaranteed gross external debt. Although this framework is primarily for debt sustainability analysis in low income countries, the indicative thresholds can serve as a useful guide to assess debt sustainability of a middle income country like Botswana. As noted earlier, given Botswana’s large fiscal savings, the debt indicators for Botswana are calculated in net terms (gross external debt minus the GIA balance). Table 6. Medium-term Fiscal Projection: Baseline Scenario Act Act Proj Proj Proj Proj Proj Proj Proj 2007/08 2008/09 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 Revenue and Grants 35.9 34.8 31.2 29.8 30.8 30.6 30.8 30.6 29.6 Revenue 35.2 34.1 30.8 29.5 30.5 30.4 30.6 30.4 29.5 Tax Revenue 21.7 23.4 21.9 20.4 19.7 18.7 17.8 17.5 17.5 Mineral Income Tax 4.7 3.9 2.3 2.3 2.8 3.2 3.7 3.7 3.3 Non-Mineral Income Tax 3.2 5.3 5.2 5.2 4.8 4.5 4.1 4.1 4.5 Customs Pool (SACU) 9.8 8.9 10.5 9.0 8.2 7.3 6.3 5.9 6.1 Sales Tax/VAT 3.6 5.0 3.5 3.5 3.5 3.5 3.5 3.5 3.5 Other taxes, incl. export duties 0.3 0.3 0.3 0.3 0.3 0.3 0.2 0.2 0.2 Non Tax Revenue 13.5 10.7 8.9 9.1 10.9 11.7 12.7 13.0 11.9 Mineral Royalties and Dividends 10.7 7.7 6.9 7.0 8.3 9.6 11.0 11.1 9.8 Interest receipts 0.1 0.1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Other Property Income, of which 1.3 1.6 0.5 0.6 1.2 0.9 0.7 0.9 1.2 o/w BOB revenue 1.1 1.5 0.4 0.5 1.1 0.8 0.6 0.8 1.0 Fees and charges 1.4 1.3 1.5 1.4 1.3 1.2 1.0 1.0 1.0 Grants 0.7 0.7 0.4 0.3 0.3 0.2 0.2 0.2 0.2 Expenditures and Net Lending 31.1 40.2 44.3 42.4 37.8 33.6 29.3 27.5 28.0 Recurrent 23.3 27.3 31.9 31.1 28.4 25.8 22.9 21.7 22.3 Wages and Salaries 8.6 9.9 10.6 10.4 9.6 8.7 7.7 7.2 7.4 Interest 0.3 0.3 0.5 0.5 0.5 0.5 0.5 0.6 0.7 Domestic Interest 0.2 0.2 0.4 0.3 0.4 0.4 0.4 0.5 0.6 External Interest 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 0.1 Other 14.4 17.0 20.8 20.2 18.3 16.6 14.7 13.8 14.2 Development Expenditure 8.2 13.1 12.5 11.4 9.6 8.0 6.5 5.8 5.7 Net lending -0.4 -0.2 -0.2 -0.1 -0.1 -0.1 -0.1 0.0 0.0 Overall Surplus/Deficit 4.8 -5.4 -13.1 -12.6 -7.0 -3.0 1.5 3.1 1.6 Primary surplus/deficit 5.1 -5.0 -12.6 -12.1 -6.6 -2.5 2.0 3.7 2.3 Financing -4.8 5.4 13.1 12.6 7.0 3.0 -1.5 -3.1 -1.6 External borrowing, net 0.0 -0.1 17.9 -0.3 -0.2 0.3 2.4 1.1 1.1 New borrowing 0.2 0.2 18.2 0.0 0.0 0.5 2.6 1.9 1.8 Amortisation (- entry) -0.3 -0.3 -0.3 -0.3 -0.2 -0.2 -0.2 -0.8 -0.7 Domestic borrowing, net 0.6 1.5 0.2 1.1 2.3 1.6 1.3 1.2 1.1 New borrowing 1.6 2.2 1.8 3.2 3.7 3.9 3.9 4.0 4.5 Amortisation (- entry) -1.1 -0.7 -1.5 -2.1 -1.4 -2.3 -2.6 -2.9 -3.4 IMF Transactions -0.1 -0.1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Pension Liability Service Fund 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Other Financing -0.2 -1.2 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Change in Cash Balances -5.0 5.3 -5.0 11.8 5.0 1.2 -5.2 -5.3 -3.8 Memorandum items Mineral revenue (% of GDP) 15.5 11.6 9.2 9.4 11.1 12.8 14.7 14.8 13.0 Domestic revenue (% of GDP) 8.7 12.1 11.0 11.1 11.2 10.3 9.6 9.7 10.3 NDP10 Cumulative deficit (in billions of 2008/09 Pula) -10.2 -20.4 -26.6 -29.6 -27.9 -24.2 -22.3 Non-mineral fiscal balance (% of non-mining GDP) -20.0 -28.7 -30.7 -30.4 -27.8 -25.8 -23.4 -21.3 -19.6 Non-mineral primary balance (% of non-mining GDP) -19.6 -28.3 -30.0 -29.8 -27.2 -25.0 -22.5 -20.3 -18.5 Government Investment Account (in millions of current Pula) 31,822 27,615 33,006 23,497 19,833 19,322 28,283 39,021 45,746 Government Investment Account (% of GDP) 39.9 31.6 39.2 25.4 18.3 15.1 18.4 22.6 25.6 External debt (in millions of US dollar) 2,962 2,922 2,836 2,972 3,412 3,641 3,761 External debt (% of GDP) 25.9 24.2 21.5 20.0 20.0 20.0 20.0 Domestic debt (in millions of Pula) 2,200 3,600 3,809 4,855 7,355 9,355 11,355 13,355 15,355 Domestic debt (% of GDP) 2.8 4.1 4.5 5.3 6.8 7.3 7.4 7.7 8.6 Total debt (% of GDP) 30.4 29.5 28.3 27.3 27.4 27.7 28.6 SBI /1 0.8 0.8 1.0 1.1 1.0 1.0 1.0 1.0 1.0 Source: Bank Staff calculation.. 1/ Assumes that education and health current expenditure accounts for 30 percent of total recurrent expenditure. Panel 1. Medium-term Fiscal Projections: Baseline Real GDP Diamond Production Volume and Real Profit 140 20 1000 40 GDP (in billions of 2000/01 Pula) Profit (ml. 2000/01 Pula perct.) Real GDP growth (in percent p.a.) 120 15 900 35 Production vol. ( ml. of ct.) 800 100 10 30 700 80 5 600 25 60 0 500 20 40 -5 400 15 300 20 -10 10 200 0 -15 100 5 0 0 Mineral GDP (right) Non-mineral GDP (right) GDP growth (left) Production vol. (right) Real profit per carat (left) Government Operation (% of GDP) Government Revenue (% of GDP) 50 30 40 25 30 20 20 15 10 10 0 5 -10 0 -20 -5 Fiscal balance Primary balance Revenue and Grants Expenditures and Net Lending Mineral SACU BoB transfer Others Deficit Financing (% of GDP) Government Debt (% of GDP) 25 45 40 20 35 15 30 25 10 20 5 15 10 0 5 0 -5 -10 -15 External debt Domestic debt GIA External Domestic GIA Balance (% of GDP) /1 Government Net Asset Position (% of GDP) 50 20 40 10 30 0 -10 20 -20 10 -30 0 -40 -10 -50 -20 -60 -30 -70 -40 -80 Scenairo 1/ Negative GIA balance suggests a portion of fiscal deficits that the government cannot finance under the existing legal framework. See text. Interest Payments/Receipts (% of GDP) Interest Payment (% of GDP) 4 4 3 3 2 2 1 1 0 0 -1 -1 -2 -2 -3 -3 -4 -4 Interest receipts -5 -5 Domestic debt Externa l debt BoB Other (PDSF, RSF, etc.) Interest on public debt Net interest pa yment Amortization Payments/Receipts (% of GDP) Net Debt Service (% of GDP) 12 14 10 12 10 8 8 6 6 4 4 2 2 0 0 -2 -2 -4 -4 -6 Domestic debt Externa l debt Other (PDSF, RSF, etc.) Net interest payments Net a mortiza tion Debt Service (% of GDP) NPV(Net External Debt) /GDP (%) 14 30 12 25 10 20 8 6 15 4 10 2 5 0 -2 0 -4 -5 -6 -10 Debt service on public debt Net debt service -15 NPV(Net External Debt)/Revenue (%) NPV(Net External Debt)/Exports (%) 200 500 400 150 300 100 200 50 100 0 0 -100 -50 3.2.2 Sensitivity Analyses 42. How would the Baseline fiscal paths respond to changes in key assumptions? Sensitivity analysis is carried out by simulating the following six assumptions ceteris paribus first, and then in combination (see Table 7). The results are discussed below. The results indicate that diamond sector performance and the size of the government establishment have the largest impact on Botswana’s fiscal outcome. Non-mining GDP growth and the domestic interest rate have a smaller effect. The domestic interest rate and the cost of new external borrowing are important mainly in the longer term as the need for financing grows. For immediate policymaking purposes, the focus in this section is primarily on the details of fiscal projection over the 7-year Plan period, since changes in assumption do not fundamentally alter the result of Botswana’s fiscal sustainability assessment discussed above. Table 7. Summary of Sensitivity Analysis /1 Assumptions Changes introduced to Baseline Asumptions Diamond sector performance • Baseline • 5% lower diamond exports • 10% lower diamond exports • 15% lower diamond exports Non-mining GDP growth • Baseline • 1 percentage point lower growth • 1 percentage point higher growth • 2 percentage point higher growth Pula/US dollar nominal exchange rate • Baseline • 10% appreciation in 2010/11 • 30% appreciation in 2010/11 • 10% depreciation in 2010/11 • 30% depreciation in 2010/11 Domestic interest rate • Baseline • 1 percentage point higher • 2 percentage points higher • 3 percentage points higher Cost of new external borrowing • Baseline • 1 percentage point higher • 2 percentage points higher • 3 percentage points higher Size of government establishment • Baseline • 1% larger establishment • 2% larger establishment • 3% larger establishment Combination • Baseline • Combination of (i) slower recovery of diamond exports; (ii) 4.4% growth of government personnel every year; and (iii) nominal Pula appreciation of 15% against the US dollar. 1/ Unless otherwise indicated, changes in assumptions are for the entire NDP10 period. Diamond Sector Performance 43. Diamond sector performance has a critical implication for Botswana’s fiscal outcomes. Sensitivity analysis is carried out by assuming a decline in diamond exports by 5 percent, 10 percent and 15 percent during the Plan period. Lower diamond exports occur through combinations of the production volume and prices. For example, a 15 percent decline in diamond exports could represent a combined effect of a 10 percent fall in diamond production and a 5 percent decline in the export price of diamonds (in nominal terms) in a given year. While an assumption of a further decline of 15 percent may appear too pessimistic, such a situation could potentially occur if global demand for diamonds did not recover as quickly as assumed in NDP10. Other things equal, reductions in diamond exports during the Plan period would cause a large drop in mineral revenue, raising the fiscal deficit. The results, presented in 44. 45. Table 9, suggest that different assumptions of non-mining GDP growth would have little impact on domestic revenue (i.e., revenue excluding mineral, SACU revenues and grants) when measured by GDP. This suggests that effective taxation on the non-mining sector is low. The average effective VAT rate during 2002/03-2008/09—derived by dividing VAT collection by GDP—is 3.6 percent, as opposed to the statutory rate of 10 percent (Table 10). As regards non-mineral income tax, the comparison is more difficult since the breakdown of personal income tax and non-mining company tax is not available. The average effective tax rate for non-mineral income tax is low, at 7.5 percent over the same period, compared with a 45.5 percent effective tax on the mining sector. This illustrates how unlikely it will be that other forms of private sector activity will generate as high a proportion of their output in the form of government revenue as diamond mining has. 8 suggests that a 5 percent decline in diamond exports would raise the cumulative fiscal deficit from P 22 billion to almost P 28 billion (in the 2008/09 Pula). Should diamond exports dip by a further 15 percent during the Plan, the cumulative deficit could exceed P 38 billion (in the 2008/09 Pula), equivalent to 45 percent of 2008/09 GDP. The GIA balance would fall below 12 percent of GDP at the end of the Plan, compared with 26 percent in the Baseline Scenario. Non-mining GDP growth 46. The second sensitivity test examines the impact of the non-mining sector’s performance on the projected fiscal path. Since the Baseline assumption on non-mining growth is on the conservative side, both lower and higher cases are examinded. The lower case considers the situation where the non- mining sector growth rate is one percentage point lower than assumed in Baseline throughout the 7-year period, whereas the higher cases refer to the situations where the sector grows one and two percentage points higher. 47. The results, presented in 48. 49. Table 9, suggest that different assumptions of non-mining GDP growth would have little impact on domestic revenue (i.e., revenue excluding mineral, SACU revenues and grants) when measured by GDP. This suggests that effective taxation on the non-mining sector is low. The average effective VAT rate during 2002/03-2008/09—derived by dividing VAT collection by GDP—is 3.6 percent, as opposed to the statutory rate of 10 percent (Table 10). As regards non-mineral income tax, the comparison is more difficult since the breakdown of personal income tax and non-mining company tax is not available. The average effective tax rate for non-mineral income tax is low, at 7.5 percent over the same period, compared with a 45.5 percent effective tax on the mining sector. This illustrates how unlikely it will be that other forms of private sector activity will generate as high a proportion of their output in the form of government revenue as diamond mining has. 8. Sensitivity Analysis: Diamond Exports 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 Mineral revenue (% of GDP) Baseline 9.2 9.4 11.1 12.8 14.7 14.8 13.0 5% lower diamond exports 8.9 9.0 10.7 12.4 14.2 14.3 12.6 10% lower diamond exports 8.5 8.7 10.3 12.0 13.8 13.9 12.2 15% lower diamond exports 8.2 8.3 9.9 11.5 13.3 13.4 11.7 Fiscal balance (% of GDP) Baseline -13.1 -12.6 -7.0 -3.0 1.5 3.1 1.6 5% lower diamond exports -13.8 -13.3 -7.8 -3.9 0.6 2.2 0.7 10% lower diamond exports -14.5 -14.0 -8.6 -4.7 -0.4 1.2 -0.3 15% lower diamond exports -15.2 -14.7 -9.4 -5.7 -1.4 0.2 -1.2 Cumulative fiscal balance (in billions of 2008/09 Pula) Baseline -10.2 -20.4 -26.6 -29.6 -27.9 -24.2 -22.3 5% lower diamond exports -10.6 -21.2 -28.0 -31.8 -31.2 -28.6 -27.8 10% lower diamond exports -11.0 -22.0 -29.4 -34.0 -34.4 -33.0 -33.3 15% lower diamond exports -11.4 -22.8 -30.8 -36.2 -37.6 -37.4 -38.8 Source: Bank Staff calculation.. Table 9. Sensitivity Analysis: Non-mining Growth Rates /1 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 Domestic revenue (% of GDP) 1 p.p. lower growth 11.0 11.1 11.2 10.3 9.5 9.6 10.2 Baseline 11.0 11.1 11.2 10.3 9.6 9.7 10.3 1 p.p. higher growth 11.0 11.0 11.2 10.3 9.7 9.8 10.4 2 p.p. higher growth 11.0 11.0 11.2 10.4 9.7 9.9 10.5 Fiscal balance (% of GDP) 1 p.p. lower growth -13.3 -12.9 -7.4 -3.4 1.2 2.7 1.1 Baseline -13.1 -12.6 -7.0 -3.0 1.5 3.1 1.6 1 p.p. higher growth -13.0 -12.3 -6.6 -2.7 1.8 3.4 2.1 2 p.p. higher growth -12.8 -11.9 -6.3 -2.3 2.1 3.7 2.5 Cumulative fiscal balance (in billions of 2008/09 Pula) 1 p.p. lower growth -10.3 -20.6 -27.0 -30.3 -29.0 -25.8 -24.5 Baseline -10.2 -20.4 -26.6 -29.6 -27.9 -24.2 -22.3 1 p.p. higher growth -10.2 -20.2 -26.2 -29.0 -26.9 -22.6 -20.1 2 p.p. higher growth -10.1 -20.0 -25.9 -28.3 -25.8 -21.0 -17.7 Source: Bank Staff calculation. 1/ Lower growth refers to one percentage point higher growth rates of non-mining GDP throughout the Plan than assumed in Baseline. Higher growth 1 and 2 are one percentage point and two percentage points higher growth of non-mining GDP, respectively. Table 10. Effective Tax Rates (2002/03-2008/09 Average) /1 Mineral revenue Non-mining income tax VAT Effective tax (in percent) 45.5 7.5 3.6 Proxy of tax base Mining GDP Non-mining GDP GDP Source: Bank Staff calculation. 1/ Since a major change was introduced in the tax structure (rates, coverage, and exemptions) in 2002/03, effective tax rates prior to 2002/03 are incomparable. Pula/US dollar Nominal Exchange Rate 50. The impact of different exchange rate assumptions on the Baseline fiscal paths are considered next. Changes in the Pula/US dollar exchange rate will impact Botswana’s fiscal position through multiple channels.64 First, nominal depreciation of the Pula vis-à-vis the US dollar raises Botswana’s external interest payments in Pula terms, putting upward pressures on overall government expenditure. However, depreciation also will raise, in Pula terms, the return on the fiscal reserves that are invested overseas and foreign grants (assumed to be provided in the US dollar), raising government revenue. The net impact on the budget depends on the net external position of the government, the differential between the interest rate on external debt and the rate of return on the fiscal savings, and the amount of foreign grants offered to Botswana. 64 One can interpret nominal appreciation/depreciation of the Pula against the US dollar as the movement of the Rand vis-à-vis the US dollar, since the Pula is pegged to a basket of currencies with a heavy weight on the South African Rand. 51. Three scenarios are considered: 10 percent and 30 percent nominal appreciation/ depreciation of the Pula against the US dollar in 2010/11 (Period 2). After the initial appreciation/ depreciation, it is assumed that the Pula will follow the trend assumed in the Baseline Scenario (see Figure 7). The results, presented Table 11, show that, during the Plan, a nominal depreciation of the Pula against the US dollar would lead to a slight improvement in the fiscal position through higher net external interest receipts, that is, net of the interest payment on public external debt and investment return on the fiscal savings. With a 30 percent nominal depreciation of the Pula, for example, the cumulative fiscal deficit would decline from P 22 billion to P 20 billion (both expressed in the 2008/09 Pula). Conversely, nominal appreciation of the Pula would lower the government’s the net external interest receipts, thereby increasing the fiscal deficit. But the impacts are small overall. It should be remembered in this context that exchange rate policy has limited value in a country like Botswana, which is a price-taker for its main exports, and the policy recommendation would therefore not necessary be to rush into an active exchange rate policy for fiscal purposes. Figure 7. Sensitivity Analysis: Pula/US dollar Nominal Exchange Rate 13.0 12.0 11.0 10.0 9.0 8.0 7.0 6.0 5.0 30% apprecia tion 10% a pprecia tion Ba seline 10% depreciation 30% depreciation Source: Bank Staff calculation. Domestic Interest Rate 52. Changes in domestic interest rates would raise the government’s domestic borrowing cost.65 Three situations are examined, where the domestic interest rate is higher than the baseline assumption by 1 percentage point, 2 percentage points and 3 percentage points. As in the Baseline Scenario, the government’s domestic borrowing cost is assumed to be 2 percentage points above the domestic interest rate. The analysis shows that, as far as the Plan period is concerned, higher domestic interest rates would cause a relatively small impact on the overall fiscal position, reflecting the relatively small domestic debt stock. However, in the longer term, when domestic public debt grows towards the prescribed debt ceiling, the impact of the higher domestic interest would be felt more significantly. 65 Although it is not modeled specifically in this framework, higher domestic interest rates may affect the private sector’s incentive for investment negatively, and could compromise future growth. Non-mining tax revenue is might suffer as a result. Table 11. Sensitivity Analysis: Pula/US dollar Nominal Exchange Rate 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 Net external interest payment (% of GDP) /1 30% appreciation -0.3 -0.3 -0.7 -0.5 -0.4 -0.6 -0.8 10% appreciation -0.3 -0.4 -0.9 -0.6 -0.5 -0.7 -0.9 Baseline -0.3 -0.4 -1.0 -0.7 -0.6 -0.7 -1.0 10% depreciation -0.3 -0.4 -1.2 -0.8 -0.7 -0.8 -1.0 30% depreciation -0.3 -0.5 -1.5 -1.1 -0.9 -1.0 -1.2 Fiscal balance (% of GDP) 30% appreciation -13.1 -12.7 -7.3 -3.2 1.3 2.9 1.4 10% appreciation -13.1 -12.6 -7.2 -3.1 1.4 3.0 1.5 Baseline -13.1 -12.6 -7.0 -3.0 1.5 3.1 1.6 10% depreciation -13.1 -12.5 -6.8 -2.9 1.6 3.1 1.7 30% depreciation -13.1 -12.4 -6.5 -2.5 1.9 3.4 1.8 Cumulative fiscal balance (% of GDP) 30% appreciation -10.2 -20.5 -27.0 -30.3 -28.7 -25.2 -23.5 10% appreciation -10.2 -20.4 -26.8 -30.0 -28.3 -24.7 -22.9 Baseline -10.2 -20.4 -26.6 -29.6 -27.9 -24.2 -22.3 10% depreciation -10.2 -20.3 -26.4 -29.3 -27.5 -23.6 -21.7 30% depreciation -10.2 -20.3 -26.1 -28.6 -26.4 -22.3 -20.2 Source: Bank Staff calculation. 1/ Negative net external interest payments mean net external interest receipts. Table 12. Sensitivity Analysis: Domestic Interest Rate 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 Net domestic interest payment/receipt (in percent of GDP) Baseline 0.3 0.2 0.2 0.3 0.4 0.5 0.6 1 p. p. higher 0.3 0.3 0.3 0.4 0.4 0.6 0.7 2 p. p. higher 0.4 0.3 0.3 0.4 0.5 0.7 0.7 3 p. p. higher 0.4 0.3 0.4 0.5 0.5 0.7 0.8 Fiscal deficit (in percent of GDP) Baseline -13.1 -12.6 -7.0 -3.0 1.5 3.1 1.6 1 p. p. higher -13.2 -12.6 -7.1 -3.1 1.4 3.0 1.5 2 p. p. higher -13.2 -12.7 -7.1 -3.1 1.3 2.9 1.4 3 p. p. higher -13.3 -12.7 -7.2 -3.2 1.3 2.8 1.3 Cumulative fiscal deficit (in billions of 2008/09 Pula) Baseline -10.2 -20.4 -26.6 -29.6 -27.9 -24.2 -22.3 1 p. p. higher -10.3 -20.4 -26.7 -29.8 -28.2 -24.6 -22.8 2 p. p. higher -10.3 -20.5 -26.8 -30.0 -28.5 -24.9 -23.2 3 p. p. higher -10.3 -20.6 -27.0 -30.2 -28.7 -25.3 -23.7 Source: Bank staff calculation. Cost of New External Borrowing 53. Next, the government’s external borrowing cost is considered. In view of the growing fiscal deficit, external creditors may impose a higher interest rate on the government’s new borrowing. Sensitivity results suggest that, again as far as the NDP10 period is concerned, a higher cost of new external borrowing would have virtually no impact on Botswana’s fiscal position. This is simply because of the fact that, faced with the external debt ceiling, the government could engage in only a very limited external borrowing during the Plan. As its borrowing of US$2.1 billion in 2009/10 would breach the statutory limit of 20 percent of GDP, the government would be unable to borrow externally until 2012/13 (Period 4), when higher GDP allows for additional external borrowing. 54. Higher external borrowing costs would, however, pose a significant negative impact on Botswana’s fiscal position in the long term. As GDP grows and existing external loans are repaid (especially the amortization of the US$1.5 AfDB loan that will start in 2014/15), the government would be able to borrow externally, or put it another way, the government would have to seek roll-over financing for the amortization. Then, the pain of higher external borrowing costs on the budget would be felt more acutely. Size of Government Establishment 55. Finally, possible slippage in the government employment policy is considered. Despite its commitment, the government could be placed under increasing pressure to hire additional personnel. Alternatively, the government may give in the demand from the public service union to raise wages above the inflation rate. What would happen to the projected path of the wage bill? A situation where the size of the government establishment grows by 1 percent, 2 percent and 3 percent with respect to the baseline assumptions is contemplated. 56. The projection results suggest that the impact of a larger government is significantly detrimental to the overall fiscal position (Table 13). A one percent increase in the size of government establishment every year—a very modest expansion considering the average annual increase of 4.4 percent during NDP9—would raise the NDP10 cumulative fiscal deficit by P 2.5 billion to P 28.5 billion (expressed in the 2008/09 Pula). Moreover, within the Planning system, the dynamics tend to encourage growth of public employment, which is unlikely to be constrained by a stated commitment to zero growth. Botswana’s wage bill is already at a very high level, with government employees accounting for over 40 percent of the labor force in the formal sector. The high wage bill is worrisome not only from the fiscal perspective but also from the growth perspective, as an upward movement in public sector wages tends to impede the private sector’s ability to compete effectively.66 66 Atkinson and Hamilton (2003) show evidence that growth rates have suffered in resource abundant countries, whose government has directed resource revenues towards public sector wages. Their cross-country regression analysis shows that a one percent increase in the share of government wages and salaries in total government expenditure has, on average, led to lower growth by 0.6 percent in resource rich countries. Table 13. Sensitivity Analysis: Size of Government Establishment (in percent of GDP) 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 Wage bill (in percent of GDP) Baseline 10.6 10.4 9.6 8.7 7.7 7.2 7.4 1% higher 10.6 10.4 9.7 8.8 7.9 7.5 7.8 2% higher 10.6 10.6 10.0 9.2 8.3 8.0 8.4 3% higher 10.6 10.7 10.2 9.5 8.6 8.4 8.8 Fiscal deficit (in percent of GDP) Baseline -13.1 -12.6 -7.0 -3.0 1.5 3.1 1.6 1% higher -13.1 -12.6 -7.1 -3.2 1.2 2.8 1.2 2% higher -13.1 -12.8 -7.4 -3.6 0.8 2.3 0.6 3% higher -13.1 -12.9 -7.6 -3.8 0.5 1.8 0.0 Cumulative deficit (in billions of 2008/09 Pula) Baseline -10.2 -20.4 -26.6 -29.6 -27.9 -24.2 -22.3 1% higher -10.2 -20.4 -26.7 -29.9 -28.5 -25.1 -23.7 2% higher -10.2 -20.5 -27.1 -30.7 -29.8 -27.0 -26.4 3% higher -10.2 -20.6 -27.4 -31.2 -30.7 -28.5 -28.5 Source: Bank staff calculation. Combination 57. The above results suggest that Botswana’s fiscal position is vulnerable to a combined shock of lower diamond exports, a larger government establishment and, to a small extent, nominal appreciation of the Pula. The final part of the sensitivity analysis thus considers a simultaneous change in all these three elements, and examines how the Baseline fiscal paths would respond. It is assumed that: (i) diamond exports will not recover to the pre-crisis until 2014/15 (Period 6); (ii) the number of the government established posts will grow at 4.4 percent a year, the average growth rate during NDP9, in addition to the new hiring of 1,610 in 2009/10; and (iii) the Pula will appreciate against the US dollar by 15 percent in nominal terms in 2010/11 (Period 2). As before, the Pula/US nominal dollar exchange rate is assumed to follow the trend assumed in the Baseline Scenario. 58. Botswana’s fiscal position could deteriorate significantly by the simultaneous changes in the three key assumptions (Table 14). With the rapid growth of the establishment, the government wage bill would rise beyond 10 percent of GDP in the first few years of the Plan, absorbing a third of total government revenue. With the higher wage bill, coupled with lower mineral revenue, the fiscal deficit would grow considerably. The projected surpluses in later years under the Baseline Scenario would virtually disappear. The cumulative deficit in the Plan would more than double that of the Baseline case, from P 23 billion to P 46 billion (in the 2008/09 Pula), equivalent to 53 percent of 2008/09 GDP. To finance the deficit, the government would resort to the GIA, using up the bulk of its fiscal savings during the Plan. As a result, the GIA balance would fall from 40 percent of GDP in 2009/10 to 6 percent in 2015/16, compared with 26 percent under the Baseline Scenario. Table 14. Sensitivity Analysis: Combination 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 Wage bill (in percent of GDP) Baseline 10.6 10.4 9.6 8.7 7.7 7.2 7.4 Combination 10.6 11.1 11.3 11.0 9.9 9.4 9.2 Fiscal balance (in percent of GDP) Baseline -13.1 -12.6 -7.0 -3.0 1.5 3.1 1.6 Combination -13.1 -13.9 -12.2 -9.9 -4.2 -1.5 0.1 Cumulative fiscal balance (in billions of 2008/09 Pula) Baseline -10.2 -20.4 -26.6 -29.6 -27.9 -24.2 -22.3 Combination -10.2 -21.2 -31.2 -40.1 -44.5 -46.2 -46.0 GIA balance (in percent of GDP) Baseline 39.2 25.4 18.3 15.1 18.4 22.6 25.6 Combination 39.2 19.1 9.0 1.1 1.0 2.4 6.1 Source: Bank staff calculation.. 59. The above sensitivity analysis confirms that the Baseline results of the medium-term fiscal projection could look dramatically more pessimistic depending on some key assumptions, in particular, diamonds. In the presence of uncertainties about the future of diamonds, precise medium- term projections are extremely challenging. As such, it is critically important that the government conduct, regularly, scenarios and sensitivity analysis of the medium-term fiscal position, especially as new information becomes available. Projections can be performed at various levels of complexity, but they should allow the government to identify emerging risks and vulnerabilities, and thereby internalizing them in fiscal policy formulations. 4. Alternative Fiscal Strategy for Botswana 60. The fiscal sustainability analysis above indicates that under current assumptions for diamond production and revenues, and in the absence of policy correction, Botswana’s fiscal position is unlikely be sustainable in the long run. As soon as 2015/16, government revenues will peak. In the following years, revenues will cease to grow, despite the fact that diamond revenues will not yet have initiated their rapid decline. To preserve fiscal balance and long run sustainability, public spending cuts would be required. Botswana has little time left to start the deep seated structural changes in its economy, starting with the public sector. These changes must be well underway before the long term decline in diamond exports sets it. The present global crisis is an opportunity to set in motion these needed changes. 61. Although the fiscal position is likely to return to a surplus in the last three years of the Plan, these surpluses are temporary and not be large enough to offset the deficits in the early years. As mineral revenue fall, the fiscal position would begin to worsen. By the beginning of NDP11 fiscal deficits would re-emerge, as mineral revenue begins to decline and interest obligations on now higher public debt start putting upward pressure on government expenditure. From this point onwards, Botswana’s fiscal position would continue to deteriorate. Initially the deficit could be financed by borrowing, but later the government would have no option but run down the fiscal reserves, until eventually the reserves are depleted in 2024/25 (Period 16). Botswana needs to adjust its fiscal strategy before it accumulates a considerable level of debt and runs down the fiscal reserves. 62. What are some of the options available to policy makers? This section will consider alternative fiscal policy strategies to bring Botswana’s fiscal stance to a sustainable path. 4.1 Alternative Scenario 1 : Higher Tax Rates 63. The first alternative scenario examines the potential for revenue enhancing measures to bring Botswana’s fiscal stance closer to a sustainable path. Botswana today has limited control over government revenue since about 70 percent of revenue is sourced outside the country. The levels of mineral revenue, SACU receipts and BoB transfers each year are exogenously determined. While good progress has been made on this front, especially since 2002, domestic tax revenue is limited as shown earlier in Table 10. 64. The assumptions underlying this scenario are that the government will embark on tax administration reform as per the IMF’s technical advice (IMF, 2004), and that these efforts will result in: (i) an increase in the effective non-mining tax rate by 0.5 percentage points every year, from 7 percent in 2009/10 to 10 percent in 2015/16; and (ii) an increase in the effective VAT rate by 0.25 percentage points every year, from 3.5 percent in 2009/10 to 5 percent in 2015/17 (see Table 15).67 All other baseline assumptions are maintained, including mining and non-mining growth projections, and the government’s adherence to a policy with a zero net increase of its establishment. How would this affect Botswana’s fiscal prospects? Table 15. Alternative Scenario 1: Higher Effective Tax Rates (in percent) 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 Alternative Scenario 1 Non-mining tax 7.0 7.5 8.0 8.5 9.0 9.5 10.0 VAT 3.5 3.75 4.0 4.25 4.5 4.75 5.0 Baseline Scenario Non-mineral tax 7.0 7.0 7.0 7.0 7.0 7.0 7.0 VAT 3.5 3.5 3.5 3.5 3.5 3.5 3.5 Source: Bank staff calculation. 65. With the revenue enhancing effort, Botswana would be able to improve the fiscal stance significantly from the Baseline Scenario. The projection results are presented in Table 16. The government would be able to run fiscal surpluses later in the Plan, albeit not large enough to offset earlier deficits. The NDP 10 cumulative fiscal deficit would fall sharply from P 22 billion to P 8 billion (both in the 2008/09 Pula). As the fiscal deficits decline, there would be less need to draw on fiscal savings. The balance of the GIA in 2015/16, at 37 percent of GDP, would be much closer to where it started (39 percent), compared with 25 percent under the Baseline Scenario. 67 This assumption is based on cross country experience that when designed well and implemented properly, VAT yields an effective tax rate which is about half of the statutory rate. For example, the effective VAT rate is about 7.5 percent in South Africa, where the statutory VAT rate is 14 percent. Table 16. Alternative Scenario 1: Projection Results (in percent of GDP, unless otherwise indicated) 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 Alternative Scenario 1 Government revenue 31.2 30.4 32.0 32.4 33.1 33.6 33.5 Fiscal deficit -13.1 -12.0 -5.8 -1.2 3.8 6.1 5.4 Cumulative fiscal deficit -10.2 -19.9 -25.0 -26.2 -21.9 -14.5 -8.1 (in billions of 2008/09 Pula) GIA balance 39.2 26.1 20.1 18.5 23.7 30.5 37.1 Baseline Scenario Government revenue 31.2 29.8 30.8 30.6 30.8 30.6 29.6 Fiscal deficit -13.1 -12.6 -7.0 -3.0 1.5 3.1 1.6 Cumulative fiscal deficit -10.2 -20.4 -26.6 -29.6 -27.9 -24.2 -22.3 (in billions of 2008/09 Pula) GIA balance 39.2 25.4 18.3 15.1 18.4 22.6 25.6 Source: Bank staff calculation. 66. In the longer term, however, more action would be required to ensure Botswana’s fiscal sustainability. As Panel 2 shows, improvements in the effective tax rates could only partially offset the declining revenue from diamond mining. Overall government revenue would continue to fall, and the fiscal deficit would widen. As in the Baseline Scenario, the government’s financing needs would be initially met by foreign and domestic borrowing. While debt would accumulate more slowly than in the Baseline Scenario, by 2021/22 (Period 13) both external and domestic government debt stock would reach the statutory limits, and fiscal savings would become the main source of deficit financing. Although fiscal savings are unlikely to be depleted within the projection period, the government would fall into a net debtor position in 2025/26 (Period 17), after a brief period as a net creditor. By this time, external creditors would likely perceive Botswana as a risky destination for investment. The government would face a very high interest rate on its borrowing. The story line of the Baseline Scenario would be repeated under the Alternative Scenario 1, albeit with a lag. 67. These findings suggest that ensuring medium-term fiscal sustainability requires more policy action in order to offset the expected decline in mineral and SACU revenues. Botswana would also need to be to stabilize the fiscal position through expenditure measures, while promoting the long-term development of non-diamond exports, which would not only generate foreign exchange to offset the declining diamond exports, but also expand the domestic tax base. The next alternative scenario will examine the impacts of such a strategy. Panel 2. Medium-term Fiscal Projection: Baseline Scenario and Alternative Scenario Baseline: Government Operation (% of GDP) Alternative Scenario1: 50 Government Operation (% of GDP) 50 40 40 30 30 20 20 10 10 0 0 -10 -10 -20 -20 Fiscal balance Primary balance Fiscal bala nce Primary balance Revenue and Grants Expenditures and Net Lending Revenue and Grants Expenditures a nd Net Lending Baseline: Government Debt (% of GDP) Alternative Scenario 1 45 Government Debt (% of GDP) 40 45 35 40 30 35 25 30 20 25 20 15 15 10 10 5 5 0 0 External Domestic Externa l Domestic Baseline: GIA Balance (% of GDP) /1 Alternative Scenario 1: GIA Balance (% of GDP) 60 60 50 50 40 40 30 30 20 20 10 10 0 0 -10 -10 -20 -20 -30 -30 -40 -40 Baseline: Government Net Asset Position (% of GDP) Alternative Scenario 1: 20 Government Net Asset Position (% of GDP) 20 0 0 -20 -20 -40 -40 -60 -60 -80 -80 1 Source: Bank Staff calculations. 1/ Negative GIA balance suggests a portion of fiscal deficits that the government cannot finance under the existing legal framework. 4.2 Alternative Scenario 2: Revenue and Expenditure Measures 68. Alternative Scenario 2 explores a strategy in which the government tries to maximize the effectiveness of public spending to boost the development of non-diamond sectors, through improvements in recurrent and investment expenditure efficiency. It is assumed that these efforts will, with some lag, raise the annual growth rate of non-mining GDP by one percentage point from 2016/17 onwards. The 20-year average growth rate of non-mining GDP is assumed to rise from the Baseline assumption of 4 percent to 4.6 percent. 69. As the non-mining sector grows and the economy becomes more diversified, it would be necessary to revise the tax regime so that the tax burden is borne more fairly across sectors. The assumption is that in 2016/17 (Period 8), the government will raise the VAT rate from 10 percent to 14 percent, to be in line with South Africa. Let’s say that this will boost the effective VAT rate from 5 percent in 2015/16 (Period 7) to 7 percent in 2019/20 (Period 11). In addition, with the more thriving non-mining economic activity, the tax base would expand gradually, resulting in an increased effective non-mining income tax rate of 12 percent in the medium term. All other assumptions and policy strategy, including the tax administration reform introduced in Alternative Scenario 1 and the government’s hiring policy assumed in the Baseline Scenario, are kept unchanged. 70. Compared with the two earlier scenarios, the longer-term fiscal prospects would be considerably improved under Alternative Scenario 2. The projection results are presented in Panel 3. Higher non-mining growth and the restructuring of the tax policy, facilitated by the government’s effort during NDP10, would boost revenue collection in the long term. Expenditure, on the other hand, is projected to fall to 30 percent of GDP by 2016/17 and to 25 percent of GDP by the end of the projection period, 2028/29, as the productivity increase in government spending would require fewer resources for public services, without compromising the quality. 71. The long-term fiscal position would be stabilized at a comfortable level, at around 2 percent of GDP. A structural surplus of this size could accommodate the impact of an unexpected external shock, should it occur, through automatic stabilizers and a temporary fiscal stimulus, without jeopardizing long- term fiscal sustainability. 72. Under this scenario, the government could choose to repay its debt as they mature without the need for roll-over financing. This is an extreme case assumption. Doing so would result in an initial decline in the GIA balance, as the government draws on the savings to finance the fiscal deficits in the first four years of the Plan. Nevertheless, the projected surpluses in later years would allow the government to gradually re-build the fiscal reserves. The government would remain as a net external creditor and a net debt service recipient, meaning that its investment return on the GIA (BoB receipt) would exceed the debt service obligation on public debt. The NPV debt ratios would show a significant improvement. All ratios exhibit a declining path, indicating that Botswana’s fiscal position would be sustainable under this scenario. Panel 3. Medium-term Fiscal Projection: Alternative Scenarios 1 and 2 Alternative Scenario1: Alternative Scenario2: Government Operation (% of GDP) Government Operation (% of GDP) 50 50 40 40 30 30 20 20 10 10 0 0 -10 -10 -20 -20 Fisca l bala nce Primary bala nce Fisca l ba lance Primary ba la nce Revenue and Grants Expenditures a nd Net Lending Revenue a nd Gra nts Expenditures and Net Lending Alternative Scenario 1 Alternative Scenario 2 Government Debt (% of GDP) Government Debt (% of GDP) 45 50 40 35 40 30 25 30 20 20 15 10 10 5 0 0 External Domestic External Domestic Alternative Scenario 1: GIA Balance (% of GDP) Alternative Scenario 2: GIA Balance (% of GDP) 60 60 50 40 40 30 20 20 10 0 0 Alternative Scenario 1: Alternative Scenario 2: Government Net Asset Position (% of GDP) Government Net Asset Position (% of GDP) 30 30 20 20 10 10 0 0 -10 -10 -20 -20 Alternative Scenario 1 Alternative Scenario 2 NPV (next external debt)/GDP NP V(net external debt)/GDP 0 0 -10 -10 -20 -20 -30 -30 Alternative Scenario 1 Alternative Scenario 2 NPV (net external debt)/Revenue NPV (net external debt)/Revenue 0 0 -20 -20 -40 -40 -60 -60 -80 -80 -100 -100 Alternative Scenario 1 Alternative Scenario 2 NPV (net external debt)/Exports NPV (net external debt)/Exports 0 0 -50 -100 -100 -200 -150 -300 Source: Bank Staff calculation. 5. Conclusions and Policy Options 73. The chapter has argued that Botswana’s fiscal prospects would improve considerably if Botswana were to take action to: (i) raise revenue, particularly from domestic sources, through improved tax administration and reform in the tax regime; and (ii) reorient public expenditure focusing on improving the quality of spending through enhanced public expenditure management, achieving improved results and service delivery with more constrained resources. As has already been noted, however, in the absence of buoyant diamond revenues, it will be almost impossible to fill the revenue gap through other sources due to the extremely high profitability of diamond mining in Botswana and the possibility of taxing it at high rates. This makes taking action to improve planning and budgeting in government of critical importance. Planning and budgeting reform is the subject tackled in Chapter 3 of this PER. 74. These policies will take time to yield results and the analysis of this chapter argues that according to current projections, time is fairly limited before diamond revenues, and consequently total government revenues, will peak and begin to decline. In preparation for the eventual large reductions in revenues accruing to government from diamonds, actions are necessary not only from the fiscal sustainability point of view, but also from the point of view of facilitating private sector activities and accelerating growth of non-diamond exports, that is, maintaining Botswana’s living standard. A timely response to Botswana’s emerging fiscal challenges would make a transition less painful, before the current fiscal policy stance leads to unsustainable outcomes that require significantly large adjustments. 75. Botswana policy makers do have access to well-established fiscal institutions to assist in the transition to a new fiscal reality. One of these is the substantial fiscal savings it has accumulated over time in the Pula Fund. For greater transparency and accountability, it would be preferable for the fiscal savings to be separated from international reserves, and formally established as a real fund under a specific law, regulation or constitutional amendment. This would cover its objectives in the context of overall fiscal policy, rules determining the payments into and out of the fund, arrangements for the financial management of the fund, the nature of the oversight, and so on. The size and investment of the fund would then need to be determined on the basis of the overall fiscal policy. 68,69 A number of options regarding the objective of the Pula Fund obviously exist, including the following. Fiscal savings can serve as a stabilization fund, with the primary objective of insulating the budget against revenue swings, in the same manner that the government has used its fiscal savings in the past. Another option is to use the savings as a future generation fund, which aims to convert mineral rents into a diversified portfolio of financial assets. This is an option in line with the IMF’s recommendation based on the permanent income hypothesis.70 76. Another institution that Botswana has made use of is the concept of a fiscal “rule”, but these rules would require some re-engineering. The current rule, limiting public expenditure to 40 percent of GDP has the advantage that it is simple to understand and it encourages policy makers to think about the long term balance between public spending and available resources and between the government sector and the rest of the economy. The key problem with this rule is that it is set too high. The analysis in this chapter indicates that the aim should be to stabilize spending in the short term at about 30 percent of GDP, bringing it down to about 25 percent over the long run in order to ensure a more balanced economy. NDP10 notes that total expenditures need to be brought down to below 30 percent by 2029, and indeed brings spending down to just under 30 percent by the end of the Plan period, but the concept of a fiscal rule appears to have been laid aside. 77. While fiscal rules are useful to focus policy-makers attentions, they have had mixed results in Botswana. As Chapter 1 argued, they have been at times abandoned when they became binding, perhaps because of the lack of a real budget constraint and due to the buoyant revenues experienced over a long period. Another issue has been the challenge of applying a fiscal rule – such as the 40 percent rule - because monitoring of fiscal activity is hampered by the poor availability of fiscal and macroeconomic information on a timely basis, i.e. within the period to which the rule is being applied. As a result, one can learn if the rule has, in fact, been adhered to only ex post, a few months after the fiscal year is over. 68 For example, where countries have created a resource fund as an instrument of saving for future generations, the desired size of the fund is usually determined by reference to a policy of attempting to keep the real wealth of the fund constant by spending on average only the permanent income from the resource wealth as part of its overall fiscal stance (e.g., Timor-Leste). Where a fund is created for stabilization purposes, the size of the fund depends on expected revenue collection and the average spending needs of the government. 69 See Bacon and Tordo (2006), who provide extensive information about the operation and performance of various sovereign funds, including the legislation governing these funds. 70 See Basdevant (2008), Clausen (2008) and IMF (2008). Similarly, rules on statutory debt limits can be a useful instrument in restraining the size of overall fiscal deficits, as long as fiscal and macroeconomic data are available in a timely manner.71 A further problem has been that the interpretation of the fiscal rule has been somewhat inconsistent. Increasingly, the 40 percent expenditure rule was being seen as a target, rather than a ceiling, for total government spending. In the budget process, 40 percent is almost automatically applied to “target” GDP (rather than projected GDP) to derive a budget envelope, with virtually no consideration to revenue prospects or the economy’s absorptive capacity. As important as the fiscal rule itself is the ability to monitor it, and Botswana would do well to ensure that reliable fiscal and macroeconomic data are available to policy makers in a timely manner. 78. Then how can Botswana know if its fiscal policy is sustainable? Given its dependence on mineral revenues, a multiple indicator approach may be most useful. In addition to the overall balance, Botswana has used the “non-investment” balance, defined as the difference between non-mineral revenues and current spending (excluding expenditures on health and education, which are regarded by government as investments in human capital) to gauge whether mineral revenues are being used for consumption, rather than investment purposes and thus not contributing to wealth creation. For this to be a useful measure, however, a much better and more transparent definition of investment would be required. Additionally, if, as recommended in Chapter 3, Botswana were to move away from a dual budget, the differentiation between recurrent and investment expenditures would cease to be useful. 79. For resource rich countries, the literature shows that monitoring the non-resource balance measured by non-resource GDP is a good benchmark for evaluating the fiscal stance. As overall balances do not tell the whole story – they could improve due to higher mineral revenues and then be misinterpreted as reflecting fiscal adjustment. When mineral revenues rise, a fiscal expansion could be temporarily masked by an improving overall balance. To get a better sense of the fiscal stance, a multiple-indicator approach is useful. In this sense, the non-mineral (primary) balance is an additional fiscal indicator for a multiple indicator approach. By excluding mineral-related revenues and expenditures and net interest from the overall fiscal balance it provides a picture of the fiscal situation abstracting from the mineral sector. The non-mineral balance can provide an analytically important tool in that it provides a useful measure of the fiscal effort and underlying fiscal policy stance. Especially from a short-term stabilization perspective, the non-mineral balance is a better guide by delinking the fluctuations associated with the developments in mining. Indeed, in countries such as Norway, budget documents and fiscal policy discussions focus primarily on the non-mineral balance and its impact on the domestic economy. The non-diamond fiscal stance might also be of better use for assessing the soundness of fiscal policy relative to the business cycle. It can constitute an important input into a fiscal sustainability and intertemporal analysis. 80. This paper has analyzed Botswana’s medium-term fiscal sustainability in view of declining mineral revenues. It has been shown that, in the absence of a serious policy correction, Botswana’s current policy strategy is unlikely to be sustainable over the longer term, regardless of the ongoing global crisis. Sensitivity analysis has shown that Botswana’s fiscal policy is extremely sensitive to the prospects of diamond exports, and is also strongly impacted by the size of the government establishment. In view of the great uncertainty surrounding diamonds, a cautious fiscal management is warranted. In particular, 71 In addition to data availability, the Bank’s Medium-term Debt Strategy mission identified Botswana’s weakness in debt management skills. a further growth in the government wage bill should be resisted. Looking forward, growth of a dynamic non-mining sector is crucial for Botswana not only from the fiscal sustainability point of view, but from the point of view of achieving balanced development that will create jobs, and delivery durable reduction in poverty and inequality. Fiscal policy will have to play a central role in this process. Chapter 3: Planning and Budgeting I. Introduction 1. Uncertainty over future diamond revenues and the need to do more with less resources present an opportunity to review how Botswana’s long-standing planning and budgeting system stands in the face of new challenges. This chapter looks at planning and budget reform in Botswana. The central question it addresses is which system of planning and budgeting, and associated financial management, is most likely to help the Government address the challenges of diminishing diamond revenues over the medium term, while at the same time improving the efficiency and effectiveness of government spending. Botswana’s planning and budget system, based upon medium term national development plans and annual budgets comprising separate recurrent and development estimates, has served the country well since independence. In recent years it has come under criticism for being repetitive, insufficiently responsive to new priorities, and not focused enough on results. It may also not adapt well to a more constrained resources environment. This chapter examines whether a planning and budget system which has been very successful in converting aid and mineral revenues into public infrastructure and social services over many years has the attributes the Government requires in the very different circumstances facing Botswana in the years ahead, and suggests options for reform. Ultimately, it is a matter of whether the focus should continue to be on projects or, more broadly, programs. For over four decades Botswana has operated a project-led budget system, with the greatest attention paid to designing and implementing new development projects. Much less attention has been paid to the considerably larger share of the budget allocated to recurrent spending, or to whether departments are obtaining the best value out of the totality of their budget allocations. 2. The focus of the chapter is the inherent nature of different planning and budget systems, how they work in practice and whether they suit the changing conditions in Botswana. It is less about the technical details of budget reform, though these will have to be addressed if the government decides to change the planning and budgeting model. Fortunately, the implementation intricacies of a move to program based budgeting accompanied by a Medium Term Expenditure Framework (MTEF) have already been set out in an IMF Fiscal Affairs Department report on strengthening budgeting in Botswana, dated May 2005. If the government decides to embrace budget reform, this report should be re-examined, and at the same time a look should be taken at comparable middle income countries which have embarked upon comparable reforms.72 3. Budget systems cannot be changed in isolation of how the public service is run, and governance Essentials generally. Changing a government’s planning and budgeting system is Box 1: MTEF conditions not essential characteristic of a MTEF is the reconciliation of policy priorities and budget realities through the The combination of a top-down assessment of total resource availability and a bottom-up determination of sector spending and funding needs. This entails multi-year budgeting, since most public programs require funding over several years to yield results. MTEFs 72 can be applied to a variety of different budget structures: unitary or dual, organizational or program based, Botswana: which can retain a traditional the Strengthening Budget focus and reporting on inputs or add anSystem to Track emphasis Key Policy on outputs priorities, and results. They including HIV/AIDS work with both cash and poverty reduction measures: An Agenda for Reform. IMF Fiscal Affairs dept, May 2005, and accrual accounting systems and variants in between. MTEFs in middle income and advanced countries by Peter Heller, are Pokar Khemani, typically Marc associated Robinson, with budgets and that Blandina Nyoni. The allocate resources to immediate programs andfocus of this report report is oninformation. the budget But a performance classification system, but this leads to a discussion of broader budget reform, such as a MTEF and MTEF can be applied also to traditional line item dual budgets. Indeed, the fiscal projections in Botswana’sprogram based NDPs budgets. are a sort of proto-MTEF. The starting point for a MTEF is a Medium Term Fiscal Framework (MTFF), prepared by the finance ministry in conjunction with other economic ministries and in consultation with the central bank. The MTFF focuses on public expenditure aggregates, taking into account tax and non-tax revenues, borrowing capacity and aid commitments, if applicable, setting an aggregate ceiling for total spending by all ministries, after taking into account debt servicing and the desirable trajectory for the budget balance. Calculation of the MTFF may take into account fiscal rules, solely a technical matter. Reform needs to take account of the broader institutional context within the government. The institutional context has been undergoing continual change since the basic framework of Botswana’s planning and budget system was laid down over 40 years ago. The public sector in Botswana is vastly different to what it was in the first decade of independence, and recent years have seen numerous initiatives aimed at making the public sector more performance oriented. Furthermore, planning, budgeting, financial management and civil service management systems need to be compatible. They are not at present. 4. The macro-fiscal context has become challenging in the short run as well as the long term. Since the scope of the Public Expenditure Review (PER) was defined in late 2008, the macro-fiscal context has become more complex. Not only is Botswana facing a transition over the medium term to less dynamic diamond revenues due to the change from open cast to more costly underground mining, and eventual exhaustion of deposits, it is also having to deal with a sharp and immediate fall in government revenues. This is due to the impact of the global financial crisis on Southern Africa Customs Union (SACU) receipts and the temporary closure of Botswana’s principal diamond mines. Further, as these mines resume production, structural changes in the world diamond market portend price uncertainty in the years ahead. II. Evolution of Botswana’s Planning and Budgeting System 5. The planning and budgeting system of Botswana features periodic national development plans and dual annual budgets. Botswana has one of the longest running planning and budgeting systems in Sub-Saharan Africa. Technically, it is a dual annual budget structure guided by a medium term national development plan. The duration of the plan is 6 years, with a mid-term review (MTR) at the half way point.73 The annual budget is in two parts: recurrent estimates comprising ministry and departmental operating votes, together with 73 The latest plan, NDP10, spans an exceptional 7 years so that the end of the plan period coincides with the final year of the Vision 2016 timeframe. debt servicing and other statutory expenditure, to be met from the Consolidated Fund (CF), and development estimates of ministries and departments, comprising projects, to be met from the Development Fund (DF). All revenues are paid into the Consolidated Fund. Aid grants and credits are paid into the Development Fund, which also receives transfers from the CF. The recurrent estimates are organized by ministries and departments, and broken down by standard expenditure groups, and within these, items. Development estimates are arranged by ministry, and by project, with a single line for each project. Although the recurrent estimates show authorized and actual spending to date for the previous budget year, the amount put forward for appropriation is for a single year. The development estimates go further, and give information on the total cost of the project approved for the plan period, the amount so far spent, the amount for the forthcoming budget year, and any residual left for completion. In this respect, it is a multi-year budget. Different rules govern the implementation of the two budgets. 6. The National Development Plan (NDP) document itself comprises a review of implementation of the previous plan, the macro-fiscal framework for the new plan, and chapters of sector policies and major programs. A public investment program (PIP) lists the projects to be implemented during the plan period, together with short narrative summaries, which contain preliminary capital and incremental recurrent costs. 7. The system’s core features go back over four decades, to Botswana’s independence in 1966. The essential features of this system were in place at Independence, with the Transitional Development Plan 1966-68, followed by the first National Development Plan, 1968- 73. There was nothing particularly revolutionary in the planning and budgeting system Botswana adopted at Independence, versions of which could be found in almost all newly independent countries in the region.74 While the new plan, NDP10, is a longer and more sophisticated document, its structure and the way it is to be implemented is little changed since NDP1. Some changes, however, have taken place, and the main developments are worth noting. 8. The main features were developed and institutionalized in the decade that followed. The 1970s saw the merger of the Ministries of Finance and Planning, and the creation of the combined Ministry of Finance and Development Planning (MFDP) in 1970 which was to dominate economic policy making for several decades to come. The Ministry of Planning became the Division of Economic Affairs, and satellite ministerial planning units were created in the major line ministries, staffed from the economist cadre, under the professional control of the Director of Economic Affairs. The first national manpower plan was produced in 1973, and shortly afterwards manpower ceilings were introduced into the budget process, to buttress establishment control and guide ministries in their staffing plans. About the same time inter- year flexibility was increased by introducing the concept of control by Total Estimated Cost (TEC).75 The first set of district development plans was introduced to coincide with NDP3. As revenues rapidly expanded with mineral development, two special funds, the Revenue Stabilization Fund (RSF) and the Public Debt Service Fund (PDSF) were established to offset pressures to increase spending and provide a cushion in case of a revenue downturn. The first 74 The particular planning model adopted in Botswana was derived from the Tanzania national planning system, and developed in the Ministry of Planning by the Permanent Secretary, Quill Hermans and the Government Economist, Pierre Landell Mills, and strongly supported by the then Minister and Vice President, Quett Masire. Although a few other countries in Africa still produce national plans, most have fallen away, and none has been so influential and durable as Botswana’s. 75 Borrowed from Malawi. Provided the TEC of a project is not breached, a department may bring forward spending on a project if implementation is proceeding faster than expected. At the same time, the TEC allows a contract to be let that may exceed the expected provision for the project in a single year. three plans had been five year plans rolled over after 2-3 years. NDP4 became a six year plan broken into two parts with a mid-term review76. 9. More recent developments have been the adoption of perspective planning, and the introduction of fiscal rules. There were few changes during the 1980s, but there were some further modifications in the two subsequent decades. In 1994 the Government announced a sustainable budget policy, aimed at holding recurrent budget spending, excepting health and education, to what could be financed from non-mineral revenues. In 1996 the Government launched Vision 2016, a twenty year perspective plan, more aspirational than costed, with a secretariat outside MFDP. The MTR of NDP9 added a fiscal rule to the planning and budgeting framework, imposing a limit of 40% of GDP on total government spending, and at the same time mandating a ratio of development spending to total spending of 30%. 10. Unlike many other countries in the region, Botswana steadily modernized its accounting system, as technology improved. Throughout the period, accounting systems have been progressively modernized, taking advantage of technological developments in accounting automation. Currently, the Botswana Government operates an integrated financial management system, the Government Accounting and Budgeting System (GABS), based on Oracle software. A linked system, also Oracle based, handles payrolls. Compared with other countries in the Region, Botswana seems to have developed and rolled out its Integrated Financial Management Information System (IFMIS) at moderate cost and without serious delays, which speaks to the professionalism of the Office of the Accountant General and the willingness of the Ministry to invest in timely accounting capacity. 11. All NDPs have contained a fiscal framework to guide budget-making in the plan period. From a budgetary perspective, the heart of the NDP is the fiscal framework for future budgets, in turn derived from forecasts of GDP growth and government revenues. The fiscal framework determines the rate at which aggregate budgets may safely grow, and from this are derived the recurrent budget growth ceilings for individual ministries and departments, within which they are supposed to frame their annual estimates proposals. The goal is a balanced program of spending, comprising high return development projects whose incremental operating costs are affordable within the planned growth of ministerial recurrent budgets. 12. An insufficiently recognized component of the system is the discipline of the “Thumbnail Sketch” – the first step in a sound project cycle. A key element of plan preparation has been the requirement that each new project proposal should be accompanied by a brief narrative description (Thumbnail Sketch – TNS), including preliminary capital and incremental recurrent cost estimates over the plan period. The aim of the TNS is to provide a first check on the consistency of new project proposals with sector strategies, the affordability of their capital costs, and the sustainability of their long run operating outlays. 77 TNSs have also proved useful over the years in helping parliamentarians understand the projects that the government plans to implement through annual budgets, having featured in every NDP since NDP1. 76 The concept was copied from Malaysia. 77 The importance of the TNS discipline cannot be overstated. As a practical matter project proposals need to be included in the plan before sufficient information is available on costs and benefits to enable an economic rate of return (if feasible) to be calculated. TNSs are therefore the first screening gate projects must pass, and arguably the most important, given the limitations of cost-benefit analysis for many categories of projects and the fact that once sufficient data are assembled, preparation may be too advanced to cancel the project without political costs. 13. Recurrent and Development Budgets have always been viewed differently. The recurrent estimates of the annual budget represent the cost of maintaining existing government services, together with new services consequent upon the completion of development projects. Although there is nothing that prevents a department terminating an activity and redeploying staff and money, in practice the recurrent estimates process is relentlessly incremental, accommodating the costs of additional staff and supplies to operate what are typically expanded services. The development budget (DB) is best seen as a time slice of the PIP that is part of the NDP, and thereby provides for the funding of ongoing and new development projects. By law, new projects may be admitted in to the annual budget only if they are approved plan projects, and fit within the ceilings given ministries in the annual budget call circular. Authority to spend does not follow automatically with the passage of the appropriation bill in the case of development projects, and requires the additional step of approval of the project memorandum by MFDP.78 14. In its day, Botswana’s dual budget system was best practice and had many practical advantages. But these advantages are no longer as important as they once were. The dual budget/NDP system, when adopted by Botswana in the 1960s, reflected best practice planning and budgeting for developing countries of that period. The system made sense for a country which needed to rapidly build up its physical and social infrastructure to enable development to take place. It also matched the expectations of aid donors, whose grants and concessionary credits became increasingly available to Botswana in the 1970s and into the 1980s. Aid management, which Botswana practiced adroitly, focused on allocating projects to donors in ways that both matched their preferences and could be implemented in the manner dictated by their aid modalities. The aim was to minimize the friction of inflexible donor procedures, while maximizing the technical value a knowledgeable donor could add to a project. At the same time, the planning and budget system enabled Botswana to manage the explosion of mineral revenues, beginning with the opening of the Orapa diamond mine in 1972, followed a few years later by Jwaneng. Indeed, the strength of the planning and budget system and the confidence placed in it by the political leadership was one of the reasons why mineral wealth was not wasted in Botswana, but converted effectively into social and physical capital, thereby developing the economy and contributing to high GDP growth. 15. While the system always had to cope with revenue volatility, the latter was usually benign. In the 1970s the volatility was invariably on the up side, with actual revenues regularly exceeding forecasts, constantly posing problems of absorptive capacity (met by phasing aggregate public sector construction and supply side policies). The 1980s began with a global recession which required Debswana to stockpile diamonds, and thus for the Government to forego revenues (which it met by a combination of cutbacks and drawing down of reserves). Thereafter, through much of the 1980s revenues were buoyant. This stressed implementation capacity and delayed project completions (and caused costs to escalate). In the 1990s and beyond 2000, revenues became less predictable, with the budget swinging between surplus and deficit, resulting in further project completion delays. 16. Manpower ceilings, originally derived from manpower plans, reinforced recurrent budget financial ceilings. Then, as now, the budget envelopes for the annual budget estimates cycle were drawn from the macro-fiscal framework of the NDP, adjusted for the latest economic data, and converted into current prices. Traditionally, ministries with a larger volume of 78 Few projects are stopped at this point. More likely is a requirement to beef up weak components, adjust timing or revise costs. Some of these powers have now been delegated to line ministries. approved development projects and programs have been allowed to increase their recurrent budgets at a faster pace than more traditional ministries, though the distinction has been less marked in recent times. Management of the recurrent budget, which has always been primarily a salaries and wages budget, was reinforced by a system of manpower ceilings, originally derived from manpower planning. In the skills and labor constrained environment of the 1970s and 1980s this attempted to predict the demand and supply of skilled manpower, and ensure a fair distribution (but not always succeeding) between the public and the private sectors. 17. Monitoring and evaluation took place, but it was input oriented, except when the NDP was being updated, when there was a stock-taking of achievements. In the traditional system, there was no formal monitoring and evaluation of results attached to the budget. Rather, monitoring was input oriented, with a focus on ensuring that monies allocated to development projects were being spent, and that recurrent spending fell within legally approved limited. The locus of M&E was the NDP (and its MTR), which would review development policy implementation progress, sector-by-sector, against policy objectives, utilizing the growing coverage and reliability of Central Statistics Office statistics, augmented by periodic surveys. In the early days, the small size of government and the modest numbers of projects in the Development Budget, accompanied by regular field visits by HQ staff and radio net reporting meant that most departmental directors knew what was happening in their sectors. 18. Not surprisingly, Human Resource (HR) systems of staff appraisal were traditional and simple. There was no comparable monitoring of public service performance. Staff assessment reports were trait based. Morale of the civil service in the early days was high, staff numbers were moderate at that point, and departmental managers had a good grasp of the issues facing their policies, programs and projects across the country. It helped that government was much smaller than it is nowadays. 19. Fiscal rules were proposed in the MTR of NDP9, to check the revenue-induced ratcheting upwards of spending, especially on the wage bill. The introduction by the NDP9 MTR of a “fiscal rule” capping total spending to 40% of GDP was in part a response to the increased volatility of revenues and the need to anchor spending at a stable proportion of GDP, to avoid spending being revenue driven. The 70/30 rule was an attempt to ensure balance between the two main components of the budget, to avoid salaries and wages absorbing an ever increasing share of spending – even though the Development Budget had by then become much less of a capital budget, and typically comprised a mix of capital and current spending. The rule also aimed at ensuring more of the total budget was allocated to social and physical infrastructure investment on which long run growth depended. 20. Taking account of these modifications, the planning and budgeting system in Botswana’s fifth decade is not greatly different to the one that guided resource allocation in the country’s first decade. The budget remains split into separate Consolidated and Development Fund estimates, the format of the Recurrent Budget is administrative, not program based, and when NDP10 is published, it will be possible to trace a clear lineage back to NDP1 and the Transition Plan. 21. A balanced assessment would be that Botswana’s planning and budget system has contributed greatly to the country’s economic development and the reputation it gained with donors for effective use of both external and domestic resources. However, its continued effectiveness is being challenged and criticism of MFDP’s way of managing public revenues is increasing. In one sense, this consistency of process is an institutional strength. The system of planning and budgeting is firmly embedded in bureaucratic culture, its routines are familiar, and in no small measure it is responsible for both the success of macro- economic management in a mineral driven economy, and development of enlightened and responsive sector policies. At the same time, both within and outside MFDP, there is a growing sense that the system is not working as well as it once did, and that a choice needs to be made between addressing the weaknesses that have merged in an evolutionary way, and more radical change. 22. Among the criticisms levied at the traditional planning and budget system have been the following: ƒ The discipline of the project cycle has eroded, manuals have not been updated, new project proposals are not subject to rigorous scrutiny, and schemes are funded that have limited economic and social justification. Above all, a concern for costs and development value has been abandoned. ƒ Whereas the Development Estimates once comprised mainly public investment, they now contain a mix of capital and recurrent schemes, some with no discernable end in sight. ƒ There is a bias towards new investment while existing infrastructure goes unmaintained. Returns on timely maintenance are typically much higher than on new investment, since failure to maintain risks losing past investments. ƒ Projects have been included in the Development Estimates for which there is no hope of staffing, either because manpower ceilings would be breached or skilled staff is unavailable. ƒ Ambitious development proposals (e.g. “hubs”) are being decided outside of the normal competitive process of NDP project screening. ƒ Public Private Partnership (PPP) schemes have been promoted without regard to their ultimate impact on the government budget, in the erroneous belief that they have no fiscal impact or risk. ƒ The macroeconomic forecasting capacity of the Government has weakened. ƒ Budgeting has become an incremental process, with line ministries arguing for more positions and funding, with little regard to the efficiency with which existing resources are being used, and whether some long running programs should be terminated and resources re-deployed. 23. More broadly, MFDP no longer dominates the economic policy landscape as it once did. In one sense this is a positive development. Line ministries have gained greater confidence in their sectors, but at the same time, a central discipline of asking hard questions about policies, programs and projects has been lost. For many ministries, the role of MFDP is to find the resources to implement programs, not to question their merit. 24. With the expansion of the Public Service (over 40% of the formally employed labour force) has come citizen and business sector criticism of its performance, and it is seen as unresponsive and unhelpful to the business climate. Overall, the Public Service is seen as too large, inward looking, protective of its privileges, bureaucratic and unproductive. This view is widely held by the private sector. It is underscored by social outcomes, where, notwithstanding generous funding, health and education results fall short of that expected from a middle income country. There is a sense that the public sector has become a protected employment domain, a costly welfare entitlement program for its employees, and the softness in public financial management is simply one dimension of a wider malaise. Corruption exists, but not currently at a large scale and Botswana rates well in international comparisons, though this cannot be taken for granted. 25. Public service reform efforts so far have had mixed results – successive waves of reform have been implemented, and new mechanisms created, but performance overall continues to attract criticism. Attempts have been made by DPSM (Directorate of Public Service Management) to make the Public Service more performance oriented. A variety of different schemes have been launched - total quality management, citizen’s charters, new appraisal systems for individual staff performance, and the placing of all permanent secretaries on performance contracts. These reforms have achieved some changes, but there has been resistance, and what shows on the ground falls short of the prospectus. 26. The most recent initiative to introduce performance management principles into the Public Service is Results Based Management (RBM). This has been launched by the Public Service Reform Unit (PRSU) of the Office of the President, which seeks to cascade downwards from the Vision 2016’s goals, objectives, Key Results Areas and strategic plans, with Key Performance Indicators to monitor progress. This initiative is ongoing, and for a period last year, its consultants took over the coordination of NDP 10 preparation from MFDP, in an effort to make the plan an instrument of RBM, before responsibility reverted back to MFDP. 79 III. Assessment of Botswana’s Current Planning and Budgeting System 27. A framework is needed for assessing the performance of planning and budgeting systems. There are many ways of looking at government planning and budget systems. One way, widely used in the World Bank, is to assess budget processes against three levels of effectiveness. The first level is the budget’s ability to implement the government’s fiscal strategy, in aggregate terms. The second level is its effectiveness in allocating resources to government spending priorities. The third is whether the budget promotes efficient and effective spending at the level of service delivery. Sometimes a fourth level is added: the extent to which the budget process is transparent, and thereby strengthens accountability. There are also other ways of assessing budget systems. The Public Expenditure and Financial Accountability (PEFA) approach uses 28 high level indicators to assess the overall public financial management system, of which budgeting is a part.80 The PEFA approach measures the performance of the systems in place but does not evaluate its inherent attributes. A summary of the PEFA Assessment for Botswana is found in Appendix 2, as is a summary of prospects for PFM reform planning and implementation. 28. The first question to ask is whether to strengthen the existing system, or move to a new paradigm of planning and budgeting. The choice facing MFDP, as it goes about preparing plans and annual budgets in a greatly changed fiscal environment which is both constrained and more volatile, is whether to stick with the familiar and tested traditional planning and budgeting system, tightening its discipline and making small adaptations, or move to a new 79 The take-over floundered, it appears, partly because MFDP was able successfully to reclaim its territory, but also because the resource allocation envelopes at the heart of the plan prioritization process were abandoned, resulting in significant departmental overbidding, which threatened to turn the plan into a needs, not availabilities, based document, which it always had been. 80 A PEFA assessment was recently carried out in Botswana by consultants financed by the EU. This should enable the Government to see how its Public Financial Management (PFM) system performs alongside those of other countries, and is a useful thing to do. A PEFA assessment utilizes 28 high level indicators to measure six dimensions of a public financial management system: (i) Credibility of budget; (ii) Comprehensiveness and transparency; (iii) Policy based budgeting; (iv) Predictability and control in budget execution; (v) Accounting, recording and reporting ; and (vi) External scrutiny and audit. The PEFA also evaluates donor practices as they relate to planning and budgeting. system of planning and budgeting, less rooted in the past and more in line with practice elsewhere in middle income countries. Needless to say, whatever planning and budget system Botswana adopts for the future should be consistent with how the DPSM is running the Public Service, and with the governance realities of the country. 29. The traditional planning and budget system has served Botswana extraordinarily well. It has been a central factor in enabling Botswana to manage mineral wealth and become the country it is today. No other country in the region has maintained such an unbroken record of consistent medium term economic planning and annual budgets brought to the legislature on time, credibly delivered, regularly reported, accounted for and audited. From one plan to another, there is consistency in policy development that makes for clarity and coherence in the country’s social and economic development strategy.81 30. Dual budgets worked well when government was small, and basic services needed to be created for poverty reduction and development. There are drawbacks to the present system for a middle income country like Botswana, which has a large public sector and an already extensive reach of public services. Dual budgets made much sense in the early days of independence for countries in the region, which made the transition to independence with very limited social and physical infrastructure outside the main urban centres. There was a huge need to invest in schools, health facilities, water supplies, roads and telecommunications, as well as basic government services like law and order, both to support private sector development and to reduce poverty. National plans and dual budgets also met the requirements of donors, who wanted discrete projects to adopt and finance. Development needed to be public sector led in a country like Botswana, and the combination of NDPs and dual budgets ensured that mineral revenues were utilized effectively. Line item recurrent budgets, organized by departments and divisions, were simple for the civil service to comprehend and operate and for ministers to explain to the National Assembly. They also fulfilled the purpose of funding salaries of an expanding public service and the operating costs of projects. 31. Dual budgets are best suited to expanding government, because they are capital budget-led. Budget practitioners have long known that dual budgets linked to NDPs are intrinsically expansionary. Of the two budgets, the Development Budget has always been the politically attractive one, commanding the attention of legislators, donor agencies and the general public. The emphasis on expanding government infrastructure (whether social or physical), however, creates an upward dynamic for government recurrent spending. The discipline of having to identify the incremental recurrent costs of a capital project, and then relate these costs to the departmental budget frame contained in the NDP, can be side-stepped by underestimating recurrent costs. There are also recurrent costs associated with policy changes which are not project related (e.g.: abolishing school fees), these are not easily captured in the traditional system. Project implementation almost always slips, which should delay the onset of recurrent costs, but in reality the corresponding adjustment is never made to the recurrent budget envelope. In countries with weak domestic revenues, implementing dual budgets is more likely to lead to delayed pay adjustment rather than a cut in hiring, and, eventually, a spiral downwards of real public sector wages leading to breakdown of a merit based civil service, informality and multiple pay augmentation mechanisms as public employees struggle to survive. Where domestic revenues are strong (as has so far been the case in Botswana), the result is likely to be an ever 81 In maintaining planning through ten or more cycles of national plans, Botswana is in the company of countries like Brazil, India, Malaysia, Tunisia and Turkey. No other SSA country can demonstrate such an unbroken sequence. growing public service, well paid but costly and increasingly unproductive. One casualty in a dual budget system is the maintenance of existing infrastructure, crowded out by new capital spending and the wage bill. Evidence (see Chapter 1) suggests this has been a growing problem in Botswana. 32. Dual budgets are generally poor at coping with revenue volatility. A further characteristic of the NDP/dual budget system is its slowness to react to changes in macro-fiscal conditions. When preparing a new plan, a major effort is invested in the macro-fiscal framework for the plan, which then becomes the reference point for annual budgets. Each year, the macro- fiscal framework is adjusted on the revenue side (particularly if shortfalls loom), but a comparable effort is seldom made to adjust on the spending side. Cuts may be levied, typically across the board, but the trade-offs between old and new priorities are seldom confronted. And if political priorities shift, the tendency is to add the new requirements, not eliminate or scale back some old ones. 33. The new fiscal context means that more attention will have to be paid in the future to “availabilities” and less to “needs”. Unfortunately, the NDP/dual budget system is “needs” not “availabilities” based. Plans are set typically by working backwards from a target rate of growth to a desired level of public investment, and, in turn, a gap to be filled by external fund raising – certainly true in Botswana while aid flows were significant. This makes annual budgeting incremental, and often adversarial, with spending ministries ganging up on MFDP, to argue for additional resources justified by Plan (or Vision 2016) goals. The corollary is that departments seldom, if ever, look at the existing use of funds, or make the trade-offs between competing programs, or seek to optimize spending within the confines of existing resource availability. There is little incentive for budget actors to review the efficiency with which existing budget resources are being used. 34. The traditional planning and budgeting model scores poorly as a tool of aggregate fiscal management and at achieving efficiency and effectiveness in spending. Measuring the NDP/DB model against the criteria of the three levels, we could say that: ƒ The model, because of its expansionary tendencies, is not very good at setting and achieving, at Level 1, sound fiscal objectives. ƒ At Level 2, allocating resources in line with government priorities, the model is initially effective but increasingly rigid over time, as budgets are supposed to follow plan policies and programs, which are changed infrequently. ƒ Performance at Level 3, however, is distinctly poor. The focus is on implementing projects. It is not on delivering services, which depends on the effectiveness of both recurrent and development budgets, and the balance between them. And spending on projects has long had a tendency to be late and with higher costs. 35. Improving the workings of the existing system would show some useful results, but not as much as more fundamental budget reform. Faced with these challenges, one option would be to try to restore discipline to the traditional NDP/Development Budget system, by injecting discipline into new project scrutiny, holding firm on annual budget ceilings, and firmly controlling the submission of supplementary budget estimates during the fiscal year. If this were to happen, the quality of budgets would improve, but such an outcome runs against the trends of recent years, and would require substantial reinforcement of MFDP’s authority as a central resource allocation ministry and its capacity to scrutinize projects. But it would still leave the government with a planning and budget system more suited to managing an expanding public sector than one that needs to improve its performance within a much tighter fiscal environment, both in the short term and the medium run. 36. What shape, then, might a future planning and budgeting system take in Botswana that best meets the changed realities of short run revenue uncertainty, and a long run revenue/GDP adjustment, tighter control over the wage bill, more efficient use of staff, and better service delivery? The next section offers some ideas what that might be. IV. A Planning and Budgeting System for the Future 37. Reform should focus on changing the annual budget system so that it becomes more strategic and more medium term in its outlook. It should be clear that there is little point in changing the annual budget system unless it addresses the main weakness of the present system. Reform should improve budget performance at Level 1 (setting and achieving aggregate fiscal objectives) and at Level 3 (encouraging efficient and effective use of resources). The two budget reforms that address these weaknesses are (i) abandonment of the dual budget and adoption of an integrated program based budget, and the progressive development of performance information; and, (ii) creation of a MTFF/MTEF to guide the framing of annual budgets and to support consistent policy making and more transparent budget management. To support the latter, a Budget Strategy Paper (BSP) should be prepared at the start of the annual budget process, and discussed in cabinet, ahead of the issue of the call circular and the setting of departmental submission ceilings. In turn, budget reforms need to be harmonized with public service reforms, to ensure that both support a more performance orientation to government. Once suitable changes have been made to the budget and rolled out to all departments, thought can be given to what role future NDPs should play and what form they should take. 38. There is a need to change the way departments view recurrent and capital spending, up to now in a very compartmentalized way, a source of considerable waste. Most middle income countries have abandoned the dual budget structure, and sought to integrate the capital and recurrent parts in a unified structure. The main reason is to encourage departments to view the two types of spending (and transfers) of equal value, to be weighed against the other, from the perspective of meeting service delivery goals efficiently and effectively. In modern budgeting there is no inherent superiority of capital spending over recurrent spending. Meeting health care goals, for example, requires a blend of current and capital spending, and, possibly, transfers and policy changes which encourage private sector responses. Spending capital budget resources to build a clinic which, on completion, is staffed and run with the recurrent budget is the traditional DB-led approach to meeting health policy goals. But the marginal pula may be more effectively spent on recurrent funding (directly or though transfers) than by capital spending. The aim should be to create a budget system which is neutral between the two types of spending, and encourages efficient production of health outputs and effective achievement of health outcomes. This points not only to a unified budget, with a single departmental budget preparation ceiling, but also the introduction of a program structure, re-arranging the budget by major departmental programs, rather than by administrative divisions and by projects. Within each program (which can be further divided into sub-programs) there would be an item structure which encompasses both wages and salaries, other recurrent charges, transfers and capital spending. 39. A program structure for the budget would enable performance information to be included in the budget document and thus begin a process of making departments accountable for results. Having a program structure enables performance information to be developed and attached to each program, to allow both departmental managers and policy makers to monitor outputs and outcomes. There is an extensive literature on the experience of countries developing performance measures. Keeping them simple, few and measurable is a good place to start. They could be derived from the relevant sectoral chapters of NDP 10, or indirectly, via departmental strategic plans. Some countries have structured performance information in a hierarchy of Key Results Areas (KRAs), derived from an overarching policy document (like the PSRU is attempting with Results Based Management – RBM – using Vision 2016).82 Box 2: MTEF Experience in Sub-Saharan Africa At a recent count, some 28 countries in Sub Saharan Africa were reported as implementing MTEF based budgets. In reality, this is a misleading statistic. While many countries report providing multi-year expenditure estimates in their budgets, a much smaller number are applying the disciplines which are at the heart of MTEF. The most developed MTEF in the region is that of the Republic of South Africa (RSA), which has helped the government engineer a huge reallocation of spending, both between and within sectors in response to new political priorities after the end of apartheid, and also underpinned a more performance based approach to public management. The starting point for RSA, however, was a unified budget structure – it never adopted a dual budget, and some years earlier had converted to program budget structures. Also worth looking at, but at the other end of the scale, is Uganda, which developed its MTEF based budgeting a decade ago, supporting better aid management and improved allocations of spending between and within sectors. For a long period Uganda retained its traditional dual budget structure. 40. A more recent MTEF Changing implementer to a program is Mauritius, structure which. be should Like Botswana, done at thehad retained same timea dual budget structure. as developing a But when growth faltered a few years, and it ago, and it became necessary to take a more critical look at the MTEF. A unified, programmatically structured annual budget, with performance information, way public monies were allocated and spent, it began to develop a MTEF based approach to budgeting, needs together to hold the including within of abandonment overall capital an separate fiscal framework. This is and current budgets typically and doneof the adoption with a Medium program based Expenditure Termbudgets. Framework, The similarities comprising, between Botswana in andthe first place, Mauritius ofgrowth, – high a Medium Term middle Fiscal income countries with Framework of thebut stable politics main fiscal public aggregates expenditure (revenues, management expenditures, systems balance that were shown and up financing), not to the challenge of economic covering and crisisof a period changed 1+2 or 1+3 years.priorities spending – suggest A key feature ofMFDP policy makers the MTFF is that should look closelyon it is constructed at the the Mauritius experience. basis of the most realistic estimate of resource availability, not on a “needs” basis, as is often the case with national plans. The MTEF is derived from the MTFF, and is essentially a breakdown of total resources into departmental envelopes, initially on an existing distribution basis, and then, in the outer years of the MTEF, progressively shaped to reflect the areas where the government wants to increase funding, and where it wants to restrict. Eventually (as happens in more advanced, mainly OECD countries and others like RSA with well developed MTEF models) the outer years of the MTEF reflect the cost of existing policy, leaving the scope for “fiscal headroom” to be allocated politically. But to start a MTEF, the allocations should be “top-down”, migrating to a more “bottom-up” system with clearly identified “headroom” (i.e. fiscal space) as program cost information is built up. 41. The advantage of a MTEF is that the framework can be re-estimated and rolled over whenever there is a change in economic parameters or in overall fiscal policy. This should take place each budget cycle, or more frequently, if required. Stability is a desirable feature of budgets, but not if it comes through delayed adaptation to new fiscal circumstances or changed policy priorities. 42. A MTEF based budget process has the potential to bring back fiscal discipline to policy-making. Both the MTFF and the MTEF are presented to Cabinet at the start of the annual budget cycle, when the opportunity may be taken to debate policy priorities, and adjust 82 There was an attempt to convert the preparation of NDP10 to RBM principles, but this initiative failed when ceilings were ignored and sectoral submissions greatly exceeded what could be afforded. individual envelopes (but not the total MTFF). The ultimate allocation of resources to ministries is a political decision, guided by the finance ministry’s advice on what existing policies and programs are likely to cost in the coming three years. The finance ministry’s main task is to determine the overall fiscal envelope, and advise on the cost of existing and new policies and programs. Ministers who seek more resources for their departments must persuade colleagues to do with less. As with the current system, ministers taking new policy proposals to Cabinet would have to explain the cost implications. A MTEF system would show whether they were affordable, and instead of putting MFDP in the position of being obliged to “find the money”, the onus would first be placed on the originating department to come up with savings elsewhere in the sector portfolio. Failing that, a collective decision would have to be taken on priorities. 43. The annual budget preparation cycle should be initiated with a Budget Strategy Paper (BSP), which would be prepared by MFDP and discussed in Cabinet. Countries, both OECD and developing, with effective MTEF based budgeting accompany the annual budget process with some sort of overall fiscal strategy document, which is prepared by the finance ministry and discussed in Cabinet, usually ahead of the preparation of the detailed estimates. Typically, a version is also published, either at the time of presenting the budget estimates to the parliament or earlier, for information or to facilitate public consultations on the fiscal situation and budget priorities. Box 3 sets out what a typical BSP might cover. In the case of Botswana there is a precedent. MFDP has long made a practice of initiating the NDP preparation process with an issues paper which contains projections of the main macroeconomic aggregates, discusses strategic priorities and other development issues, and sets out options. MFDP has also had a practice of going to cabinet at the start of the budget process with a memorandum giving the fiscal outlook and proposed MDA ceilings. A BSP would, in effect, combine these two practices.83 44. For years, Botswana has built its NDPs around a proto-MTEF, so to extend the MTEF to annual budgets is not a large step. A MTFF/MTEF based approach to budgeting should not be seen as a completely new mechanism in Botswana. The fiscal framework embedded in the macro-economic chapters of the NDP is a proto-MTEF. The main difference is that a full MTEF is rolled over at least annually, and formally approved by Cabinet at the start of the annual budget cycle. In the traditional model, it was approved as part of the plan, but then applied administratively and not very transparently by MFDP in the context of the annual recurrent budget ceilings in the budget call circular. A primary concern of MFDP will be to adjust total spending over the medium term, to bring it into alignment with what will be fiscally sustainable. As a practical matter, this is likely to target total expenditure between 25% - 30% of GDP, not the 40% of the fiscal rule, which is no longer relevant. It is important that all central government spending is brought within the MTEF – there should be no exempt categories to be financed outside the budget. The MTEF is the crucible in which all competing spending requirements are sorted and prioritized, and collectively agreed. In this way it becomes the fiscal framework at the heart of the policy making process. No new policy or change in existing policy with acknowledged financial implications is ever discussed in Cabinet without reference to the impact on the MTEF and whether resources are likely to be available to finance the policy (or program or project) in the outer years. 83 Two countries that have made this a practice are South Africa and Uganda. Box 3 Budget Strategy Paper (BSP) Budget strategy papers (BSPs) typically follow a structure with four parts. The first part discusses the overall economy, including external developments, macroeconomic aggregates and the current fiscal situation. (Not unlike the economic overview the Minister presents in his Budget Speech). The second part usually talks about development challenges, and the policies the government has been pursuing. The third part deals with budget performance, chiefly the outturn expected with the current budget and issues arising therefrom. The fourth part will look at the implications for budget policy and choices. There will also be tables showing the budget aggregates looking forward. For the past decade South Africa has published a Medium Term Budget Policy Statement (readily accessible from the National Treasury’s website). The most recent MTBPS was released in October 2009. It is organized into four chapters, headed as follows: • Supporting the recovery, transforming the economy • Economic policy and outlook • Fiscal policy and trends • Medium term expenditure framework and division of revenues, The fourth chapter contains a table (4.1 Consolidated Government Expenditure 2008/09 – 2012/13) which is broken down by both type of public service (eg: public order and safety, economic sector, health, education etc) and by economic classification (current payments, transfers and subsidies, and capital payments). It is, in effect, a summary of the MTEF. 45. Budget reform should start with the adoption of a government-wide MTEF. Creating a government-wide MTFF and MTEF should be done at the outset. A MTEF which is not comprehensive of government spending is meaningless, since the trade-offs between competing spending allocations are not faced. On the other hand, development of ministerial program structures can be done over several years, starting with a few representative pilot ministries, to test program structures and develop relevant performance indicators. 46. Moving to a program structure for the budget and making programs the focus of budget management means shifting attention away from projects. Projects, insofar as they are time limited capital spending schemes, become sub-sets of programs, expected to last indefinitely, or until policy changes. The question is often asked how a MTEF based budget should deal with development projects, including discrete, large ones. With the dual budget model, projects large and small are individually approved and financed, and fitted into the department’s annual budget generally by expanding that department’s resources envelope for project spending. MTEF based budgeting calls for a different approach. Routine capital spending, such as the building of additional classrooms to accommodate an expanding school age population or higher enrolments, or construction of new rural water supplies, or clinics, simply become the capital components of an overall program, such as the provision of secondary schooling, rural water supplies, or health services, in which the focus is on the total funding of the service, leaving the responsible departments to propose the optimum mix of recurrent, transfer and capital funds (and possibly cost recovery, and contracting out) to achieve efficient and effective service delivery. 47. Adopting a MTEF based budget means discarding the mechanism of a centrally determined Public Investment Program. There would be no point in a MTEF based budget system of maintaining (as is currently done with the NDP system) a separate government-wide PIP. PIPs are systems for central control of the composition of capital spending, and, as such, are part of the older development planning paradigm of public resource allocation. However, this would not preclude departments with a heavy capital component in their spending, such as roads, from developing a sector specific PIP. However, this would be to better plan and manage their departmental spending within overall MTEF limits, optimizing the balance between maintaining existing roads better and upgrading or building new ones, either directly, or through performance based PPP arrangements for maintenance and operation of parts of the core road network. MFDP would still need to concern itself with proposed large capital projects (such as the hubs) to ensure that the sponsoring departments were properly estimating costs, both capital and recurrent, accurately projecting their development benefits and realistically assessing their risks. 48. Large projects will continue to feature in the government’s spending plans, but need to be evaluated critically against alternative uses of funds. Very large investment projects (such as infrastructure spending for a “hub”) would have to be evaluated on their merits at the time of (or before) annual budget preparation, and decisions made by Cabinet, with the advice of MFDP, as to whether the implementing department’s total budget ceiling should be adjusted upwards to accommodate large project spending, and what spending plans elsewhere in the budget should be postponed or foregone, within the overall MTFF. This would make clear that large project spending has a cost in terms of resources forgone for other purposes, and thus must demonstrate clearly superior economic rates of return. 49. There remains a need to improve the quality of departmental project preparation and scrutiny. Earlier this chapter commented on the decline in the quality of development project preparation and scrutiny, reflecting, in part, the loss of MFDP dominance and in part the expansion in the number of projects in the budget and the delegation of project oversight to departments and agencies. As part of the budget reform process, there is a need to bring up to date project planning manuals, restore standards in project preparation, and train staff in line ministries. Since this would be taking place within the context of a wider move towards program budgeting, manuals would need to be updated to ensure that projects were now being carried out within a program context. Likewise, there would have to be extensive training of staff on what the shift to a program structure for the budget means. 50. The change to a program structure implies a complete re-ordering of budget incentives and a change in the roles of MFDP and line ministries. This last point is a vital one, since it goes to the heart of how government is managed. Recasting the budget in program format is not just presentational, to make budgets more transparent and enable legislators relate the resources they are approving to the policies they have endorsed in the NDP. There is a more fundamental goal to change the way departmental heads manage their budgets. Here, the objective is to terminate the incrementalism that has characterized annual budgeting in Botswana for so long, in which line ministries “gang up” against MFDP to obtain more resources, and instead create an entirely new set of budget incentives which recognize the authority of MFDP to set the overall fiscal strategy and to advise on development strategy, while at the same time acknowledge the political nature of allocation decisions, and the responsibilities of departments to manage budget resources, both staff and money, against policy, program and project objectives. Since the present planning and budgeting system is so firmly embedded (both the formal system and the informal practices that have grown up), an explicit change management strategy will need to be developed. 51. Roles and responsibilities, together with methodologies for monitoring will need to be defined afresh. Monitoring requires strengthening both at the centre and also in line ministries, and its scope broadened. The latter is necessary because the traditional monitoring system is project based, and needs to be expanded to encompass programs. There is no point in focusing only on the timely and cost effective execution of a hospital building program, if the policy goal is the delivery of health care services and better health outcomes. There needs to be both a strategic monitoring and reporting capability for programs at the centre (run by either Government Implementation Coordination Office or MFDP), and departmental monitoring, to provide feedback to managers on how well programs are being implemented, and whether additional management interventions are needed. 52. Budget reform needs to be developed in tandem with public service reform. There is little point in moving to performance management on the budget side if civil service rules remain hierarchical and bureaucratic, and vice versa. A program structure in the budget, if it is to achieve the results intended, has major implications for how the public service is run. Up to now, efforts by DPSM to make the public service more performance oriented have had mixed results. Permanent secretaries have resisted being placed on performance contracts, and managers are quick to shift the blame for shortcomings to others (lack of funds, procurement delays, staff vacancies etc). The new system almost certainly requires both additional delegations and a greater holding of managers to account for how effectively they have used the financial and staff resources at their disposal. Accompanying the development of a program structure for the budget in the pilot ministries would have to be changes in personnel rules and procedures. This in turn implies a more harmonized approach to running government by the Office of the President, DPSM and MFDP.84 53. Ultimately, what matters most for both budget reform and public service reform is how from the top down, the leadership of the executive chooses to run government. Even when working together, there are limits on what Human Resources Management (HRM) and PFM systems can do to improve government performance. Ultimately, success depends on how senior policy-makers at the head of the executive choose to run government. Just as with the “old” civil service (but now in a more explicit way) the “new” civil service should be merit based, intolerant of poor performance. Much depends on how senior public service managers are chosen, what performance is expected from them, and what happens if they fail to deliver. 54. Reform is likely to require changes to the legal framework for PFM. Introduction of a new planning and budget system is likely to require changes to the legal framework for government, both the long standing Finance and Audit Act and its successor legislation, and Financial Regulations, and their equivalent on the public service side. For a start, the distinction between the Consolidated Fund and the Development Fund is no longer relevant. Unified budgeting requires a single common fund. Just as the existing law forbids Development Fund spending on projects that are not part of an “approved plan”, so the making of annual budgets within a MTEF might be made mandatory, together with additional transparency measures requiring the government to publish its medium term fiscal strategy, the economic assumptions on which it rests, and the long run values for critical variables such as total spending to GDP, deficit share of GDP and debt to GDP. The requirement to save surplus revenues above, say, 30% of GDP, could be statutorily reflected. Changes to the legislation could either precede the introduction of new processes, or follow the latter. 55. Changes to the government accounting system are required to support budget reform, but full accrual accounting needs to be approached cautiously. The Business and Economic Advisory Council (BEAC) report recommended a move to accrual accounting by the government, but did not specify how this should be done. Full accrual accounting has been adopted by only a small fraction of OECD countries, and implies changes to budget systems that go far beyond what is outlined in this chapter. Almost certainly, full accrual accounting would be a “bridge too far”. 85 Progressively moving from cash to modified cash (which effectively Botswana long ago did) and then to modified accrual, once the business case is clear, makes more sense. However, introducing a program structure for the budget means modifying the Chart of Accounts, and adjusting the expenditure classification code string to capture spending by program. This is technically straightforward. 56. Once the budget system is changed, a fresh look can be taken at the role, content and format of future NDPs, or whether they need to be continued at all. Finally, once the new budget system has been tested in the pilot ministries, rolled out to all line ministries and institutionally embedded, the opportunity can be taken to revisit the future of national development plans. There is likely to be continuing value in a document that lays out the development challenge, and reviews existing policies and programs and their affordability and effectiveness within a consistent macro-fiscal framework. This could be a shorter document than 84 A start has been made by PSRU, but so far this has been unconnected with budgeting or the NDP. It draws from the Vision 2016 document and not the NDP, and appears more “needs” based than “availabilities” based. There is a need to assess and harmonize both HRM and FM reforms. 85 As introduced, for example, in New Zealand or the UK. the current NDP, shorn of the PIP and individual projects, but still a guide to making policy and program decisions in a transparent and inclusive manner in the annual MTEF based budget. 57. Since changing the planning and budget system is being done not for its own sake but to change the way the Government performs, the key policy-makers will need to understand both the rationale for change, the essence of that change and the steps to implement it. Ministers and senior civil servants, and, in due course departmental managers in the sector ministries, need to understand what is entailed by a change in planning and budgeting models. Most particularly, economists and budget staff in MFDP will need to thoroughly grasp the new approach, since it will fall on them to guide colleagues in sector ministries through the process of MTEF based budgeting and the development of a program structure. 58. A strategy for managing the changes needs to be developed, understood and owned by the political and bureaucratic leadership. Although the need to change the way government is managed is made more pressing by both the immediate and longer term challenges to revenues, a sense of urgency alone is insufficient. 59. The following are suggested as next steps: ƒ MFDP officials to review afresh the May 2005 IMF FAD report on “Strengthening the Budget System”, which has a detailed section on implementing a MTEF, and any subsequent advisory work by the IMF’s Fiscal Affairs Department. ƒ MFDP and DPSM to discuss and agree on a joint program of public sector modernization in Botswana based on integrating PFM and HRM reforms. ƒ Establishment of MFDP led technical groups to develop a sequenced plan of implementation for introducing the required changes to the public financial management system, such as macro-economic forecasting, modification of fiscal rules, accounting changes, budget preparation calendar,publication of a Budget StrategyPaper, changes to financial regulations, and the Finance and Audit Act. ƒ Identification of pilot ministries to begin the process of changing from a dual line item budget to an integrated program structure. ƒ Updating of policy making practices, to ensure that new policies or changes to existing policies and programs are discussed by cabinet against the background of the MTEF and the aggregate fiscal framework. This includes decisions on “hubs”. ƒ Development of a capacity building program for economic, budget, HR and finance staff at MFDP and line ministries, updating of existing and development of new procedural manuals, and a change management strategy. Chapter 4: Public Investment I. Introduction 1. Public investment of mineral revenues has transformed Botswana and played a central role in making the country the development success story that it is. Since 1985, 31 percent of the budget has gone to development spending. This has been done in the context of Botswana’s NDP system, with a government-wide Public Investment Program (PIP), and a dual annual budget comprising a recurrent budget and a development budget. This chapter reviews the performance and evolution of this planning and budgeting system as it relates to public investments in particular. The following section gives a brief overview of public investment and access to some associated services in Botswana; section 3 describes the process of public investment planning in the country and discusses some of the problems that have emerged in the current system; section 4 concludes and suggest some options for policy-makers’ consideration. II. Public Investment in Physical and Social Infrastructure 2. Botswana has re-invested an enormous amount of its mineral revenues since independence, in social and physical infrastructure and services, largely through government budget. Table 1 shows central government allocation of development spending by functional category during the most recent NDP 9 (2003-2008). 3. A great deal has already been achieved as a result, and as noted in this report, but significant gaps remain. Botswana lags comparator upper middle income countries in terms of youth and adult literacy rates despite high levels of budget outlays. Infrastructure capacity and quality have not kept pace with growth, and the electricity, transport and water sectors are now presenting bottlenecks to sustained growth. Capital expenditures are therefore likely to continue to constitute a significant portion of public spending for some time to come, as noted below. 4. Roads. When Botswana achieved independence in 1966 it had 12 km of paved road. Today it has 6,000 km of paved road and 12,000 km of unpaved network. The length of the network grew rapidly in the 1970s - 90s, but slowed down significantly over the last five years, as the competing demands for limited public financial resources have increased. The current transport networks are basically developed to service the relatively small economy. However, all these land and air transport facilities are aging and require serious rehabilitation. The government budget is facing problems with funding the increasing needs of the transport sector. This pattern started to appear with an accumulation of periodic maintenance backlogs. Taking road sub-sector for example, the ministry's total expenditures on roads in 2008 was US$100 million (about 2.0 percent of total budget expenditures), of which about US$80 million was for development and US$20 million for maintenance. This is far from sufficient for maintaining the system of assets worth about US$2 billion. Emerging regional integration and diversification needs means that more resources are likely to be needed in the roads sector (World Bank, 2009). Table 1: Development Budget Allocations in NDP 9 (2003/04 – 2008/09) – (%) 2003/04 2004/05 2005/06 2006/07 2007/08 2008/09 Revised General Public Services 7.0 7.8 7.0 9.9 8.1 10.7 General Administration 3.5 4.3 4.5 6.9 5.4 6.1 Public Order and Safety 3.5 3.4 2.4 3.0 2.7 4.6 Defence 5.9 6.0 3.7 4.7 5.4 3.1 Education 9.3 7.8 7.0 7.5 12.3 8.8 Health 13.1 17.6 19.3 15.2 12.9 8.8 Food and Social Welfare Pgms. 2.7 1.0 2.4 0.8 2.8 1.0 Housing, Urban & Reg. Dev. 7.0 2.8 5.5 6.9 9.5 10.9 Other Community & Social Svc. 1.0 1.5 1.0 0.7 1.1 3.1 Economic Services 23.1 20.5 20.8 21.1 22.6 34.9 General Administration 0.0 0.1 0.2 0.6 0.5 0.5 Agriculture, Forestry & Fishing 2.3 0.8 4.8 1.7 1.6 4.0 Mining 0.5 0.7 0.6 0.0 0.3 0.5 Electricity and Water Supply 9.2 11.4 11.2 8.8 6.0 16.7 Roads 5.5 5.7 2.4 7.4 6.3 5.6 Air Transport 0.0 0.1 0.2 1.3 4.5 5.7 Rail Transport 2.0 0.0 0.0 0.0 0.0 0.0 Post and Telecommunication 1.7 0.5 0.6 0.5 2.6 0.9 Other Transport 0.7 0.1 0.2 0.3 0.2 0.2 Promotion of Commerce & Industry 1.2 1.1 0.7 0.5 0.7 0.8 Unallocated Expenditure 30.8 34.9 33.3 33.1 25.2 18.7 100.0 100.0 100.0 100.0 100.0 100.0 5. Water and Sanitation. Botswana has one of the highest access rates to improved water sources with 95 percent of the total population having access, slightly higher than the 93 percent access of upper MICs, and far ahead of the rest of SSA. Central government spending on water infrastructure is high relative to other countries because Botswana is a large, sparsely settled area with sometimes vast spaces between settlements, so economies of scale are sometimes hard to realize. Also due to high evaporation rates and low topography, ground water capture in daps is relatively inefficient, so more collection per unit supplied is required. Third, most of the rural supplies come from boreholes which are often quite deep (hundreds of meters sometimes as the Kalahari layer can be quite thick) Exploration and operation of such deep wells is difficult and expensive. The recent Africa Infrastructure Country Diagnostic report (AICD, see Box 1) notes that Botswana has one of the lowest reservoir capacities in Southern Africa. 6. In terms of improved sanitation, only 41 percent of the population has access, less than half the access level for upper MICs (at 86 percent) and similar to access in low-income countries. Rural access to improved sanitation is just 25 percent, far behind 76 percent for upper MICs. Urban access stands at 57 percent of the population. 7. Power supply only reaches 56 % of the population; access in a typical upper middle income country stands at 87 percent. The national target is to reach 80 percent access by 2016 through large investments in the sector. Under-investment in power supply has been identified as the most important infrastructure –related constraint on continued growth of the economy (AICD, 2009). Box 1: Ten Main Messages of the AICD Report for Botswana AICD is a project designed to expand knowledge of physical infrastructure in Africa. It provides an overview of in-country the status of public expenditure, investment needs, and sector performance in each of the main infrastructure sectors, including energy, information and communication technologies, irrigation, transport, and water and sanitation. For Botswana, the AICD report can be summarized in the following ten main messages, although data limitations make some of these statements provisional: 1. Infrastructure’s contribution to growth has been high, with the power sector holding the economy back. (NB. This is calculated as contribution to per capita growth, with infrastructure defined as Telecoms, Power and Roads, and excludes the water sector for which no data was available.) 8. In health, Botswana has emphasized the widespread provision of primary health care rather than the narrow provision of hospitals and doctors to the few. Virtually all urban residents are within fifteen kilometers of a primary health care facility as are over 80 percent of rural residents. This approach to health care resulted in a dramatic fall in infant mortality, from 100 per 1000 live births in 1971 to 45 in 1991, although with the HIV/AIDS epidemic it had risen to 85 per 1000 live births by 2005. Before the impact of AIDS, widespread public health services also resulted in an increase in life expectancy at birth from less than fifty years at Independence to the upper sixties in the 1990s. Botswana includes the bulk of its HIV/AIDS related spending in the development budget. Looking ahead, the costs of the national HIV/AIDS program are projected to remain high, slightly increasing from an estimated 3.2 of GDP to 3.4 percent of GDP in 2025 (Haacker, 2009). Relative to government revenues, the costs of the national HIV/AIDS program are therefore projected to increase from the current 7 percent of revenues (excluding grants) to 10.3 percent by 2025. 9. In education, primary school enrolment has gone from 66,100 in 1966 to 327,600 in 2000. Secondary school and university enrolment have both grown at double-digit rates. The proportion of the population with completed primary education or better has moved from 1.5% in the 1964 census, to 43 percent in the most recent, 2001 census (Leith, 2005). Looking forward, development expenditures on education are expected to remain high, with NDP10 continuing to focus in particular on secondary and tertiary education to build the skills judged to be necessary for diversifying Botswana’s economy through increased regional and international competitiveness and to support private sector growth. Initiatives include the construction of more secondary schools, the Botswana International University for Science and Technology (BIUST), expansion of the University of Botswana campus to Maun; and the establishment of a teaching hospital. 10. Botswana has also initiated the development of six “Hubs” (in Diamonds, Health, Education, Agriculture, Innovation, and Transport) to promote diversification and growth. These are being undertaken outside the NDP framework on the whole; only expenditures associated with the Innovation Hub are clearly identified in the NDP 10 PIP, although it is almost certain that much of the required capital investment will need to come from the budget, with the hope of spurring increased private investment further down the line. III. Challenges Facing the Current System 11. Botswana has been seen as a best practice leader in terms of its programming of public investment and (during the 1970s and 1980s) the management of its external assistance. Botswana uses a classic NDP/PIP/DB system, with a 6-year National Development Plan that incorporates a PIP and which is translated into annual, dual, budgets. At first glance, Botswana’s system incorporates most of the “must have” features of an efficient public investment system.86 1) The NDP provides broad strategic guidance for public investment, anchoring government decision and guiding sector-level decision-makers. 2) Projects are screened, first by the responsible ministry during the preparation of the TNS and later by the MFDP at the Project Memorandum stage prior to release of funds; larger projects undergo more complete economic feasibility or cost-benefit analysis. 3) There is a formal project appraisal process, which is reviewed at the central level by MFPD. 4) The process establishes envelopes for public investment: the PIP includes the Total Economic Cost of each project, a 5-year projection of annual costs, along with the balance of funds required to complete them; TECs are revised annually during budget preparation to ensure that they adapt to a changing environment. 5) There is an effective process to control the gates to the PIP; the vast majority of projects are included in the NDP, and adding a project once the NDP has begun (“climbing the fence”) is difficult, although it can be done under exceptional circumstances. 6) Project evaluation is undertaken in the context of the Plan’s Mid-Term Review and during the preparation of the subsequent Plan. Box 2 outlines the main features of the public investment process in Botswana. Box 2: How the Development Planning Process Works The size of the development budget is largely determined during the preparation of the six-year National Development Plan, as are the component projects. Project Identification: Projects, in theory, can be identified almost at any level. But the formal demarcation of a project is made by the implementing agency (IA), usually a government department or ministerial planning unit (MPU). Consultations with stakeholders are undertaken at this stage through a National Stakeholders Conference during the preparation of the NDP. At the grassroots level, communities are able to access the NDP consultation process through their constituency Members of Parliament, the Kgotlas and through district and local authorities. Project Screening and Selection: Initial screening is conducted by the IA and the MPU. A Thumbnail Sketch (TNS) is prepared containing the background of the project, purpose of the project and ties to a particular policy. The TNS should include a general technical description, an outline of the project’s benefits, both quantitative (as applicable) and qualitative, the project’s Total Economic Cost (TEC) and projections of the capital and recurrent cost of the project for the implementation period in current and constant prices. TNSs should assist IAs in screening out lower priority projects. A second screening is conducted by MFDP at meetings of Project Review Committee (PRC). The Project Review meeting is where each ministry presents its NDP (and annual) bid for each project to the PRC. During the project review exercise, sources of finance for the development budget are identified (largely domestic). A project is likely to be approved as long as it fits within the ministerial ceiling and is broadly in line with national objectives. These approved projects are then incorporated in the NDP upon the approval of cabinet and parliament. No economic appraisal techniques are applied at the screening and selection stage. Adding a project during the course of the Plan period is difficult, though possible. Approval. The NDP, with the PIP (the compilation of TNSs) is approved by Parliament. While approving the 86 R NDP, Parliament also approves project TECs. Therefore revisions to the TECs must be approved by Parliament through the annual budget or supplementary estimates. In urgent cases or emergencies, TEC revisions can also be sought through Cabinet memoranda, and submitted to Parliament for ratification at its next sitting. Appraisal. Thus, before projects become part of NDP, no economic or financial appraisal techniques are applied beyond the TNS. Once the NDP is approved, a more in-depth analysis is required for funds to be withdrawn to implement the project. This is done in the form of a Project Memorandum (PM) submitted by the IA and relevant MPU to MFDP. The PM provides a detailed plan for project implementation, a basis for final performance. Quarterly Box 3: Concerns About PIPs sector progress reports are Over the years, PIPs have raised concerns. One frequent concern is that submitted to GICO, projects included in the PIP cannot be ranked by economic rate of return. analyzed and then When a project idea is first considered, there is insufficient dat. But by the time the latter has been assembled, it is usually too late in the project cycle submitted as part of a to deny the project a place in the PIP. A momentum has already built up. Cabinet Information Note Even if rate of return ranking were possible, Internal Rates of Return (IRR) on progress, also on a are not comparable across sectors, and for many projects they cannot be quarterly basis. More calculated. Where they can be calculated and data are available, few comprehensive policy or development countries could deploy enough economists to cover all project proposals. (Chile’s internationally recognized system of project preparation program evaluations are and evaluation is based on 3,000 professionals.) And if the gatepost for the envisaged every 3-5 years. PIP is a threshold IRR, line ministries will cook the figures. Furthermore, 12. Despite these PIP planners never start with a clean sheet of paper. Up to 90 percent of the features, Botswana today available space in the PIP may be preempted by ongoing projects. ______________________ has an over-expanded Adapted from World Bank, 1998, Public Expenditure Management government, as noted in Handbook, Box 3.12 previous chapters, and a weakening public investment process. And despite good intentions, Botswana’s reliance on dual budgeting and a PIP has encouraged a capital-budget led system with a focus on projects, and with policy and program considerations coming as an afterthought. In investment led growth models, political pressure to commit government to the highest possible growth path can quickly lead to too many projects being initiated, causing either slow implementation or recourse to inflationary spending. Competition between ministries for budgets ends in escalation of costs, expansionary spending and a growing government, and a gradual loss of the discipline behind project screening and appraisal. In the case of Botswana this has been made worse by a soft budget constraint, where the cost of approving poor projects with low rates of return was less apparent and did not lead to any serious trade-offs. “In theory, the PIP/DB mechanism was not supposed to affect fiscal discipline. In practice, the preference for projects undermines fiscal discipline in most developing countries.” 87 (World Bank, 1998) 13. Although the NDP sets out government plans and priorities, these are not clearly linked to the expenditure allocations. The focus of expenditure planning is on individual projects, rather than on a wider program of activities and/or projects aimed at achieving government objectives. TNSs begin to look more like capital budget shopping lists than well rounded initial project/program summaries, despite the fact that the macro-fiscal chapter presents a picture of constrained resources. The proper function of the NDP sector chapters is to lay out policies, programs and projects, confront trade-offs and make priorities consistent with available resources, informed by the medium and long term adjustments the country must make. This has not happened in NDP 10. Generally there is an absence of a discussion regarding the resources required to achieve Plan goals, other than in the macro-fiscal aggregate chapter. 87 Discipline appear s to have been lost gradually over time. In terms of being a control document, for instance, NDP 10 is less effective than previous plans. Macro fiscal projections are presented in the aggregate only. There is not apparent conversion of the aggregate 5-year revenue and expenditure tables into a spending frame for each ministry. The sector chapters include goals and performance targets, but little on resource requirements or staffing needed to achieve these targets. As there are no cost figures and no apparent ceilings, no tradeoffs seem to have been made between competing programs. 14. Discipline appears to have been lost gradually over time. In terms of being a control document, for instance, NDP 10 is less effective than previous plans. Macro fiscal projections are presented in the aggregate only. There is no apparent conversion of the aggregate 5-year revenue and expenditure tables into a spending frame for each ministry, and so no simple MTEF at the heart of the Plan as there used to be previous Plans. There is also no summary table – i.e. a PIP – with the total cost of all the projects contained in the Plan, both recurrent and capital. This suggests that the total capital spending approved in the Plan was not totaled. Summing all the capital expenditures included in the TNSs reveals that these exceed the total development spending included in the macro-framework of the Plan by about 15 percent. The link between the development budget and fiscal reality is weak. 15. The capacity to shift resources across priorities is limited by the expansive nature of the development budget and by project “overhang” . The NDP 10 fiscal framework sets out strategy for broadly maintaining government spending levels during the Plan period (provided that mineral revenues recover quickly), but shifting the focus of government spending towards the needs of the private sector, for example, by spending on education, training and maintenance of existing infrastructure. Yet the capacity to shift resources is limited. There is little fiscal for new projects under the Plan’s base case scenario. Projected resource needs for development projects envisaged during NDP10 are about 15 percent higher than the development budget ceiling in the fiscal framework. Of the projects included in the Plan’s TNSs, 77 percent represent on-going projects, and 23 percent are new investments. Assuming that given currently constrained resources, fiscal space for new projects will only become available once on-going projects have been substantially completed, less than 10 percent of NDP10 investment is likely to be on new projects. A discussion of the transition from ongoing to new projects is absent from the Plan. The constrained resource envelope projected in the fiscal framework for the development budget is likely to be pre-empted almost entirely by incomplete NDP 9 projects. 16. The linkages between development spending and the recurrent implications of the development budget are weak. While much effort and thought has in the past gone to ensuring that development projects include a fair assessment of their associated recurrent projected recurrent expenditures in the TNSs, a general characteristic of the dual budget system is that it fails to capture the total recurrent costs that arise from development expenditure. This is reflected in project delays and escalating costs. During the period covered by the MTR of NDP8, for instance, development spending was characterized by “front-end loading”, or the bunching of development spending at the beginning of the Plan; something similar is taking place in NDP10. At the time, this was an attempt to reduce the backlog of delayed projects, but resulted in the associated recurrent expenditures not being appropriately budgeted for. Considerable strain was placed on limited resources, especially skilled manpower available to the government and private sector contractors. As a result, problems of cost escalation and poor workmanship emerged. Costs increased because the initial estimates of projects’ outlays and scope were often inadequate. Similarly, the upsurge in development spending seen at the end of NDP 9 and planned for the initial years of NDP 10 will require increased recurrent spending to maintain the additional projects, although it is not at all clear that the required linkage between project costs outlined in TNSs and the overall macro framework has been made. In the reverse case, when project implementation is delayed – as has more usually been the case in Botswana – recurrent spending continues to grow. (MTR NDP 8) 17. Poor planning, including poor financial management and procurement planning, is evidenced by constant delays in project implementation. Close to 50% of all projects suffer implementation delays on one form or another.88 Project Memoranda are not submitted on time or it takes a long time for them to be approved. Poor planning is also evidenced in delays in the tendering process and frequent cost overruns, a lack of skilled staff to implement project investments, the breaching of MDA manpower ceilings, and ultimately an inability to staff and maintain completed projects. All these issues are supposed to be identified at the screening and appraisal stages, but are not. Deterioration in project performance has ensued. With poor ex ante scrutiny of economic benefits, ex post returns from public investment have fallen, even if this has not been accurately measured.89 18. The emphasis on new projects in a capital-budget led system leads to the postponement of operation and maintenance spending, and a cycle of neglected maintenance, collapse and rehabilitation. A substantial decline on the returns to investment is the result. As a case in point, the development program for NDP 10 contains a number of large projects which are the backlog of maintenance or refurbishment that should have been undertaken under the recurrent budget. This cycle of neglect has been a pattern in Botswana, particularly in recent years. 19. A recent tendency has been to approve large projects of “national importance” outside the scope of NDP. Investments in the Hubs described above appear to be taking place outside of the normal process of NDP project screening. The only expenditure clearly identified in NDP 10 is that for the Innovation Hub. This undermines the competitive project selection process even further, and reduces transparency and accountability for the investments that undertaken in this manner. Neither actual nor projected expenditures – be it for capital or recurrent expenditures – on the Hubs are reported anywhere in the NDP. While international experience suggests that there are valid reasons at times for unscreened projects entering the development budget, an effective gate-keeping process would limit this to a very small number of emergency projects. Long-term projects of national importance such as the Hubs, geared towards supporting economic diversification in Botswana, do not fit this description, and to ensure that the NDP remains authoritative and comprehensive, need to be incorporated into the Plan. 20. The tangible outcome of these challenges is that during implementation, Government normally falls short of its promised development program. This has been a chronic issue in Botswana. Table 2 shows the growth of government spending during NDP 9, and the significant under- implementation of the development budget. As can also be noted, this was rather spectacularly reversed during the last two years of the NDP 9, with over 50 percent growth in actual development spending during each of those years. There was also acceleration in recurrent expenditure, but on a much smaller 88 BIDPA, 2009. 89 Some evidence of this is in the change in the marginal productivity of capital in the national accounts, as the return on the re-invested capital from the exploitation of the diamond wealth declined. Botswana moved from a low incremental capital-output ratio (ICOR) of 2-3 units of capital per unit of additional output in the 1980s to an ICOR of close to 5 in the early 2000s. This was reflected in the Vision 2016 strategy, where in order to achieve the 8% p.a. real growth required to triple per capita incomes by 2016 (compared to 1996), Botswana would need to have an investment rate of 40%. (Exchange with J. Salkin). scale. Table 3 gives more information B : Investment Inefficiencies The World Bank has categorized the types of inefficiency in on the credibility of the budget, investment management that are observed in many countries under showing budgeted allocation of the following six dimensions: development spending against actual • Poor project selection, including wasteful “white elephant” expenditures, by government agency. projects As noted, 2007/08 and 2008/09 are • Delays in design and completion of projects exceptional in terms of the level of • Corrupt procurement practices development spending executed, and • Cost overruns 2006/07 is more typical, with the • Incomplete projects capital budget under-executed by • Failure to operate and maintain assets effectively so that the benefits are less than they should be. nearly 30 percent. In 2008/09 the Evidence of these failures can weaken the case for expanding revised budget estimates were 54 investment spending, particularly in a crisis when there is more of percent higher than the initial a need for speed and quality. estimates (Government of Botswana, __________ Source: World Bank, 2009. A Diagnostic Framework for Assessing Public NDP 10). Investment Management. 21. How did this rapid increase in development spending come about? Various reasons have been given for the under spending of development budgets. These include a lack of implementation capacity in some government ministries, slow land allocation, and difficulty with filling project posts because of unattractive packages and the insecurity associated with project positions. Land availability, the need to perform environmental impact assessments and design issues were noted as constraints on implementation of some major projects in the capital budget in 2007. The requirement for Environmental Impact Assessments was waived on a case by case basis to accelerate implementation. In addition, ministries were allowed to proceed with pre-project planning for projects included in the budget, but before the budget is approved by Parliament. This was allowed for the first time in the period between January and March 2007 (just before the beginning of the 2007/08 budget year). Ministries have also been given the ability to approve, through their Accounting Officers, their Project Memoranda up to a threshold of BWP 10 million, rather than having to seek approval from MFDP. Ministries are also now permitted to directly hire consultants without tendering for projects costing under BWP25 million. (Budget speech, 2007) Table 2: Annual growth of government spending (constant prices) - % 2003/04 2004/05 2005/06 2006/07 2007/08 2008/09 Recurrent 4.1 -205 -4.6 2.0 8.9 16.5 Development -5.6 -14.1 -12.0 -3.0 50.2 53.4 Total -3.4 -1.7 -6.3 1.3 17.5 23.8 Source: NDP10, Table 4.6. 22. It appears that these changes were effective in accelerating development spending, although the reforms do not address the main underlying problems of poor planning and diminished capacity and incentives to carry out proper project screening and appraisal. The reaction to the under-implementation of the development budget has been to address the superficial symptoms but not the underlying problems. The 2010 Budget Speech notes that “there was a general problem in budgeting and financial management during the 2009/10 financial year, across most Ministries. This is evidenced by factors ranging from cost-overruns caused by inadequate budgeting, award of tenders without enough funds, change of scope of projects without authority for additional funds and transferring of funds between expenditure heads without re-prioritizing.” As a result, the dispensation to grant authority to Accounting Officers to approve funding for projects which cost P10 million or less has been amended to exclude the approval of Project Memoranda on consultancies and computerization, which will now revert back to MFDP. While streamlining processes and easing controls might expedite the expenditure, there can be a trade-off in terms of the quality of investment, particularly if oversight and accountability are not strengthened. 23. Finally, the development budget falls short on the monitoring and evaluation of outcomes. It does not provide incentive for government to effectively evaluate its earlier measures, rethink its key policy objectives and prioritize expenditures in the best possible way. A major effort, spearheaded by GICO, is underway to remedy the weakness of monitoring and evaluation. Care should be taken, however, not to put too much emphasis on the ability to implement a project as compared to ensuring that the project offers good value for money. This is one of the criticisms of the Malaysian system where there is a fear of undertaking a that is difficult to implement, and where Departments do not want to get caught by the Implementation Coordination Unit. IV. Conclusions 24. The combination of PIPs and dual budgeting has been associated in practice with many dysfunctional budgeting, resource allocation and financial management practices. The spending dynamics and incentives of systems based on annual budgets and government-wide investment plans tend to encourage investment-led budgeting, an overexpansion of government, and a neglect of policies and programs and ultimately, service delivery. While many of the important elements needed to implement a good public investment management system are in place, but the incentives and spending dynamics inherent in a dual budget/PIP system have meant that the positive elements have weakened over time. 25. In Botswana today there is an ongoing discussion regarding the merits of the current planning and budgeting the possible gains from reforming it. In the context of the on-going debate, this chapter recommends a shift away from dual budgets and PIPs towards a more fully integrated approach to programs that integrate recurrent and capital spending in a medium term expenditure framework in order to address the structural failings of the current system. The PIP in Botswana, as in many other countries, emerged as a reflection of external financing of the government budget in the decades immediately following independence. The PIP as a tool for better managing donor aid has, however, served its purpose, and Botswana no longer relies on external financing to any great extent. 26. Even if the shift towards MTEF were to be undertaken, there would still be a need to strengthen the role of formal project screening and appraisal process both at MFDP and in line ministries even as the need for a government-wide PIP disappears. While capacity to screen projects has declined in line ministries and MFDP, this can be rebuilt, and the long standing but now neglected Planning Officers’ Manual can be resuscitated. There is a need for systematic, high quality project assessment, including a better assessment of options, with formal and well publicized guidance on the technical aspects of project appraisal. These also need to be appropriate to the technical capacity of ministries. All this would require an explicit strategy to strengthen the capacity and skills of the civil service to manage public investment in its programs. Specialized training covering project appraisal, procurement, project management and policy analysis is needed. A stronger monitoring and evaluation system to ensure that there is learning and feedback from successes and failures is another requirement. Proposals for public investment usually exceed resources available, and choices have to be made between different projects being proposed. Thus a necessary, though not sufficient, condition would be to restore the central role of MFDP in the screening and appraisal of project, debilitated by years of plentiful resources and pressures to continually increase spending. 27. As discussed in chapter 3, Botswana is technically well placed to take on this reform, provided the requisite political support and ownership is present. An adequate and reliable resource envelope is still available to allow genuine prioritization, and prudent overall macroeconomic and fiscal policy-making, a well-functioning financial management system, and minimal incidence of corrupt practices are key advantages. Retaining and strengthening the existing system would be a risky undertaking, given the apparent shift in the balance of power in government away from technocrats towards the politicians who are less interested in details such as cost-benefit analysis and rates of return in the face of short-term interests. Table 3: Budget Credibility – Budgeted and Actual Development Expenditure Credibility of the Development Budget Thousand Pula 2006/07 2007/08 2008/09 1 * Organization Budget Actual Actual/Budget Budget Actual Actual/Budget Budget Revised Revised/Budget % % % Parliament 2,715 2,393 88.1% 8,950 705 7.9% 14,015 13,271 94.7% State President 1,480,203 679,623 45.9% 1,930,407 1,656,007 85.8% 1,918,042 1,882,228 98.1% Ministry of Finance and Development Planning 65,807 28,658 43.5% 134,667 127,907 95.0% 191,316 191,316 100.0% Ministry of Labour and Home Affaris 105,831 62,227 58.8% 274,330 118,247 43.1% 174,255 220,827 126.7% Ministry of Agriculture 113,952 101,433 89.0% 162,047 150,596 92.9% 208,875 729,760 349.4% Ministry of Education 485,620 233,257 48.0% 584,479 594,284 101.7% 1,037,597 796,384 76.8% Ministry of Trade and Industry 71,534 65,181 91.1% 93,317 65,214 69.9% 96,705 130,000 134.4% Ministry of Local Government 973,604 907,876 93.2% 989,926 1,421,553 143.6% 1,326,180 2,231,680 168.3% Ministry of Works and Transport 728,018 584,172 80.2% 785,539 688,491 87.6% 988,980 1,982,868 200.5% Ministry of Minerals, Energy and Water Affairs 471,519 396,235 84.0% 637,062 466,645 73.2% 653,935 2,512,072 384.1% Ministry of Health 508,535 539,366 106.1% 635,579 366,854 57.7% 575,540 575,540 100.0% Administration of Justice 31,057 26,005 83.7% 66,131 29,777 45.0% 64,085 134,000 209.1% Ministry of Foreign Affairs and Int’l Cooperation 14,900 13,248 88.9% 51,600 22,378 43.4% 64,085 92,817 144.8% Independent Electoral Commission 612 268 43.8% 3,800 2,101 55.3% 6,000 3,410 56.8% Ministry of Lands and Housing 352,765 256,047 72.6% 363,021 356,795 98.3% 280,115 280,115 100.0% Ministry of Communications, Science and Technology 110,648 105,287 95.2% 346,983 290,833 83.8% 315,735 315,735 100.0% Ministry of Environment, Wildlife and Tourism 85,527 53,654 62.7% 188,712 156,908 83.1% 283,115 530,798 187.5% Total 5,602,847 4,054,929 72.4% 7,256,550 6,515,296 89.8% 8,198,575 12,622,821 154.0% 1 Excludes Attorney General’s Chamber, Auditor General, Office of the Ombudsman and Ministry of Youth, sports and Culture 2 Actual figures not yet available. These reflect revised estimates Source: Government of Botswana Annual Statements of Accounts, various editions; Estimates of Expenditure From the Consolidated and Development Funds, various edition Chapter 5: The Social Safety Net in Botswana: Spending, Coverage, Challenges I. Introduction 1. While Botswana, as has been noted in the previous chapters, has a high poverty rate, with most recent estimates putting it at 30 percent of households living below the national poverty datum line, a compensating factor has been the relatively high level of social provision of services such as education, health care, clean water, etc. available to the entire population either free or at heavily subsidized rates. In 2002/03, government social spending averaged just over BWP 17,000 per household (about US $3400), a very high level relative to the income of the poor90. Nevertheless, it is fair to conclude that Botswana’s development strategy has been more effective in improving services and general conditions than in alleviating poverty. Disaggregated data on social indicators show that certain groups and geographical areas continue to lag in terms of improvements in social conditions; these include women, female-headed households, remote area dwellers, the illiterate and the disabled. 2. Two key elements of Botswana’s strategy to spread the benefits of mineral-led growth have been employment generation and social provision, but as growth slows, this strategy of spreading growth in income through growth in employment may be reaching its limits. Heavily subsidized social expenditures including in health and education are also likely to face the discipline of budget constraints and a greater need for cost recovery as economic conditions change. This is happening at a time when a large portion of the population still lives below the poverty line. A rethinking of government’s approach to ensuring social justice and to sharing the benefits of economic growth may thus become necessary, including a refocusing of social safety net expenditures. 3. Social Safety Nets (SSN) are one component of poverty reduction strategies. Public policy statements, including Botswana’s National Strategy of Poverty Reduction (NSPR) and the NDP 10 highlight the importance of SSN in providing support to those who are unable to take advantage of income earning opportunities, “diminishing their exposure to risks, enhancing their capacity to protect themselves against psycho-social and economic hazards and the loss of income.”91 4. Government provides a broad spectrum of SSN programs, designed to fulfill special needs rather than generalized income support. Existing SSNs address the risks associated with malnutrition, HIV/AIDS, unemployment, disability and old age, and appear to be comprehensive over a life-cycle. A 2007 study found, however, that SSNs overall cover only a small proportion of the poor (19 percent), although coverage varies widely by program, and that many beneficiaries are non-poor households (57 percent).92 5. This chapter assesses Botswana’s SSN system, highlights some of the challenges that policy makers are likely to face as demand for broader coverage increases with changing economic 90 The latest Household Income and Expenditure Survey was conducted in 2002/03. Government social expenditures include spending on: education; health; food and social welfare programs; housing, urban and regional development; and other community and social services. 91 NDP 10 Chapter 12: A Compassionate, Just and Caring Nation. 92 BID PA circumstances, summarizes relevant international experience, and provides some suggested avenues for policy looking forward. The next section describes existing SSN programs in Botswana; problems of estimating expenditures on SSN in Botswana are discussed in section 3; section 4 outlines some of the key challenges to building a well-functioning SSN system in Botswana and relevant lessons from international experience; section 5 concludes with some policy options. Box 1. What Are Social Safety Nets?* Safety nets are defined by the World Bank as noncontributory transfer programs, targeted in some manner to the poor and those vulnerable to poverty and shocks, and the term is analogous to the U.S. term “welfare” and the European term “social assistance”. Safety nets are themselves part of a broader poverty reduction strategy—interacting with and working alongside social insurance, health, education, and financial services, the provision of utilities and roads, and other policies aimed at reducing poverty and managing risk. Safety net programs can play four roles in development policy: • They redistribute income to the poorest and most vulnerable, with an immediate impact on poverty and inequality. Most societies hold strong convictions that adequate provision for the poor is required, though they may differ in how this should be achieved. • Safety nets can enable households to make better investments in their future. In this role, safety nets basically act to remedy credit market failures, allowing households to take up investment opportunities that they would otherwise miss—both in the human capital of their children and in the livelihoods of the earners. • Safety nets help households manage risk. At minimum, safety net programs help households facing hard times avoid irreversible losses, allowing them to maintain the household and business assets on which their livelihoods are based, and to adequately nourish and school their children. At best, they can provide an insurance element that lets households make choices about livelihoods that yield higher earnings. Safety nets thus both protect households and promote their independence. • Safety nets allow governments to make choices that support efficiency and growth. An adequate permanent social assistance system can fulfill whatever redistributive goals the society has, freeing other sectors from the role and letting them concentrate on efficient provision of services. Thus, for example, energy sectors can price for efficiency, and trade policy can focus on growth rather than job protection. Short-term safety net programs can compensate those negatively affected by needed reforms or who may oppose and stall these reforms. The kinds of programs typically included as common elements in a safety net are: cash transfers or food stamps, whether means tested or categorical as such as child allowances or social pensions; in-kind transfers, with food via school feeding programs or mother/child supplement programs being the most common, but also of take- home food rations, school supplies and uniforms, and so on; price subsidies meant to benefit households, often for food or energy; jobs on labor-intensive public works schemes, sometimes called workfare ; in-cash or in-kind transfers to poor households, subject to compliance to specific conditionalities on education or health; and fee waivers for essential services, health care, schooling, utilities, or transport.” Botswana’s SSNs cover most of these areas, and are described in the next section. _____________ * This Box is based on Grosh et al., 2008 II. The Social Safety Net in Botswana A. Program Descriptions 6. Botswana has nine major SSN programs. These are described below, and more succinctly, in Table 1. In rough order of coverage, these include (1) primary school feeding program (2) vulnerable group feeding program (3) old age pension scheme (4) the destitute persons program (5) orphan care program (6) community home based care program (7) labor based public works and drought relief programs (8) remote area development program and (9) world war II veterans program. In addition to the formal government-provided nine main SSN programs, Botswana outsources significant social care services to the Non-Governmental Sector. Botswana also has a traditional, informal safety net that has eroded as the country’s population has become more urbanized, concentrating in the Gaborone area and growing suburbs. Primary School Feeding Program 7. In terms of size and coverage, the largest programs are the feeding programs (Primary School and Vulnerable Group Feeding Programs, covering a total of about 580,000 people in 2006/07, about a third of Botswana’s population. In 2009/10, according to the Ministry of Local Government’s Food Relief Services Division, the primary school feeding program provided 331,000 school children with both a snack and a full, cooked midday meal, whose caloric value is one third of the child’s daily need. The primary objective of the PSFP is to minimize child malnutrition and stunting and to enhance learning. This program and the Vulnerable Group Feeding Program have been key to improving nutrition in children and have supported the success of HIV/AIDS treatment programs. All children attending public primary schools are eligible, and there is no means testing. Feeding is also provided to secondary students and students who board in special educational institutions. For the school year starting in 2009, the Ministry of Education and Skills Development reported that approximately 165,000 secondary school students were fed. Budgeting for these programs is split among several different agencies. MLG FRSD is the budget holder for the primary school feeding programs and the vulnerable groups feeding programs, while Ministry of Education’s budget covers secondary school students and students in specialized boarding institutions. Vulnerable Group Feeding Program 8. This is one of the oldest social safety nets for children and vulnerable groups. It was established at independence in 1966 and aims at distributing meals and nutritional supplements to people who are vulnerable to malnutrition and women of child bearing age from poor or low income households. The Program is implemented by the Ministry of Health and local authorities. Beneficiaries are pregnant and lactating mothers, nutritionally at risk under-fives, and tuberculosis patients. During drought years supplementary feeding is provided to all under-fives as well as food rations for lactating mothers. During non-drought years, supplementary feeding is based selectively on the weight progression of the child. In July 2005, there were 268,000 beneficiaries registered under this scheme. The vulnerable groups comprised approximately 131,000 children aged 6-36 months, and 122,000 children aged 3-5, with the remaining beneficiaries being lactating mothers, pregnant women, and TB outpatients. Table 1. Botswana Social Safety Nets93 SSN Program Eligibility Criteria Packages Beneficiaries # Primary School • All children attending public primary schools are • The food basket caters for a third of the daily caloric 331,000 Feeding Program eligible. (Universal benefit) requirements for children. primary school • Two meals are provided, a mid-morning snack and students lunch. (2009/10) Vulnerable Group • During non-drought years, the program covers • Tsabana (a fortified soghum and soya product) for 268,000 people Feeding Program medically selected under-five children and pregnant children between 6 and 36 months. and lactating women (including those who are • Dry skimmed milk for children aged 37-60 months anemic, not gaining enough weight teenagers and medically selected pregnant and lactating between 13 to 18 years, and TB and leprosy women. patients); blanket coverage is implemented during • Sunflower oil for children ages 6-60 moths. drought years. (Selection by age and medical • Fortified precooked maize for children aged condition) between 37-60 months and TB out-patients. Old Age Pension • All citizens aged 65 years and over are entitled for • Pensioners received P220.00 per month (2009). 95 percent of Scheme benefits. (Universal benefit) those eligible, 90,639 in October 2009 The Destitute • Individuals unable to engage in sustainable • According to the policy, permanent rural destitute 42,381 in Persons Program economic activities, due to disability or chronic persons receive food packages amounting to October 2009 health problems. P211.90 per month. Their urban counterparts • Individuals with insufficient assets or income receive food packages worth P211.40 per month. sources, should have less than 4 livestock units, and However, the figures may go up, depending on the should earn income less than P120/month without cost of the approved food basket. dependents or P150/month with dependents. • According to the policy, temporary destitute persons • Individuals who due to physical or mental disability receive monthly food rations valued at P181.90 in are incapable of engaging in sustained economic rural areas, and P181.40 in urban centers. However, activity, as determined by a health practitioner. the figures may go up, depending on the cost of the • Individuals who due to emotional or psychological approved ration. disability are incapable of engaging in sustained • Both permanent and temporary destitute persons economic activity, as determined by a social worker. receive an additional P81.00/month (in cash) for • Children under 18 living under difficult personal (non-food) items. circumstances. • Provisions are made for shelter (if needed), medical • Individuals who are terminally ill. care, occasional fare, funeral expenses, and exemption from service levels, taxes, water charges, (Means-tested) street licenses and school fees, and tools required for rehabilitation. 93 This table is based on BIDPA 2007, with updated figures based on information from the Ministry of Local Government and NDP 10. for rehabilitation. Orphan Care • Children under 18 years of age who have lost one • According to the policy, the food baskets 46,833 children Program (single) or two (married) parents (biological or (amounting to P216.60) are provided through local October 2009 adoptive). (Registration) retailers; each basket is based on nutritional requirements by the age of the child. However, in practice, the cost of the food basket depends on prevailing local prices. • Clothing, toiletry, transport fees, school fees, etc. Community Home • Terminally ill patients. While it was established in • Provides optimal medical care for terminally ill 3557 people Based Care Program response to the HIV/AIDS epidemic, the Program patients in their local environment. October 2009 covers those patients with other conditions as well. • Provides food baskets based on recommendations • Food basket is provided to needy patients only. by a doctor or dietician; thus no price is attached to • Assessment guidelines for the destitute program are the food basket applied. • In practice, the cost of food baskets has ranged from (Selection based on medical condition) P200 to P1,500 per patients per month. Labor Intensive • Self selection; poor adults • Provides income support to the poor (LIPW) and Number of Public Works temporary support during periods of drought; beneficiaries (LIPW) and workers are engaged in labor intensive maintenance varies per year Drought Relief and construction activities. Laborers receive P15.00 Program while supervisors receive P20.00 per six-hour day. World War II • All WW II veterans are eligible for benefits under • The beneficiary receives P359.00 per month. 3,005 (October Veterans Grants this program 2009) • When the veteran dies, his widow receives payments. • If both the veteran and his spouse are deceased, their children under the age of 21 receive payment. Remote Area • Targets 64 identified marginalized communities in • Provides basic facilities to communities; facilities Development the remote areas of Botswana. include education, health, drinking water, and Program (Geographical selection) vulnerable group feeding schemes. • Promotes access to land and water through water rights. • Promotes income generating opportunities for remote areas dwellers. • Promotes self reliance, social integration, etc. Old-Age Pensions. 9. The Old Age Pension (OAP) scheme was introduced in Botswana in 1996 and is an entitlement scheme, thus not means tested. It is administered by the Commissioner for Social benefits in the Ministry of Local Government. The objective is to provide financial security to the elderly citizens who may otherwise be without means of support. The 2002/03 HIES indicates that over 95% of the elderly are now registered for this program. Eligibility is for all Batswana citizens 65 years and older.94 Botswana citizens receiving other pensions are not excluded. Identification to confirm eligibility is through the national Registration Card. Pensioners under the scheme who get paid by check or by proxy are required to sign a “Life Certificate” every three months as proof that they are still alive. Beneficiaries in 2007 received a cash component of P166 per month, which is adjusted for inflation each financial year. Some issues are that. Beneficiaries receive their allowances from post offices. 10. In 2008/2009, there were 97,114 recipients of the old-age pension according to MLG making the old-age pension program the second largest official safety net program in Botswana after the feeding programs. There is some lack of clarity regarding the actual number of beneficiaries, however, and other MLG information estimated 98,208 beneficiaries. 95 The discrepancies among the MLG data point to systemic issues about public expenditure reporting in Botswana that are discussed later in this chapter. Destitution program. 11. The destitution program was adopted in 1980, and substantially revised in 2002, as a response to the erosion of the traditional safety net in pre-Independence Botswana.96 The objectives of the scheme are “to ensure that government provides minimum assistance to the genuine destitute persons to ensure their good health and welfare.” The program thus targets the few who have absolutely no other sources of support, and there is an extensive procedure and form for means-testing eligibility. Eligibility for destitute benefits is therefore targeted and conditional. The Ministry of Local Government staff work closely with traditional authorities in the rural areas, and this appears to work as an additional means-test. Destitute persons are classified into Permanent or Temporary categories. Permanent destitute are those whose age, physical or mental conditions render them completely dependent and they are thus eligible for benefits for life, with no conditions apart from an annual assessment by social workers. Temporary destitute are those temporarily incapacitated by ill health or natural disasters until they can support themselves. 12. The number of destitute has been growing since the program was implemented and has been paid out to approximately 38,000 permanent beneficiaries during the past four fiscal years, as well as 25,000 needy students and children. The destitution program consists of food packages and a small cash allowance for non-food needs, access to social services as well as various subsidies (in education, health, transportation and housing). In 2007, a permanent destitute person was entitled to P256.90 worth of food items per month in rural areas, plus P61 cash for personal items. This amount can rise to P400 depending on the rising costs of commodities. Adjustments for inflation are made on a yearly basis. Children in the destitute program, in addition to receiving food rations, receive school 94 The age limit of 65 Is not harmonized with the retirement age of 60 for public servants 95 From IBM-style printouts obtained from MLG DSS HQ. A handwritten note claims 99,224 OAP beneficiaries in 2009/10. 96 Revised National Policy on Destitute Persons, July 2002. uniforms, transport, tuition etc. All categories of destitute persons are exempted from payment of publicly provided services such as medical fees, school fees, water charges, service levy and electricity charges. Orphan Care Program 13. The Orphan Care Program was launched in 1999 and responds to the immediate needs of orphans, including food, clothing, shelter, education, protection and care. An orphan is defined as a child below 18 years who has lost one or two biological or adoptive parents. With the advent of social change coupled with the escalating rates of HIV/AIDS, the number of orphans has risen. Unlike the destitute program, the orphan care program is a social allowance and is not means-tested. In 2002 there were 39,571 registered orphans; today there are close to 50,000. This figure is thought to be a serious underestimation as some relatives are reported to refuse to register orphans because of the stigma associated with the HIV/AIDS epidemic. It is estimated that only about 60 percent of the eligible orphans are registered. In 2007, 92 percent of registered orphans received assistance, with only 8% being entirely supported by relatives. Orphans receive a food basked of P216 per month, which is regularly adjusted for inflation at the beginning of each financial year. In terms of public expenditures, the Orphan Care Program is one of the largest SSN programs in Botswana today. Community Home-Based Care 14. This program was established in 1995 in response to increased illnesses due to HIV/AIDS. The aim of Community Home-Based Care (CHBC) is to provide quality care from at the home setting. The program is implemented through a partnership between the Ministry of Health and the Ministry of Local Government. Patients enrolled in the program benefit from clinical medical assistance as well as a food basket that is aimed at meeting the nutritional needs of the patients. The pula value of the food based depends on the nutritional needs of patients. Eligibility for assistance under this program is not means tested. There has been a significant decrease in the number of supported CHBC patients over the past five years, with numbers dropping from about 9,500 to 3,400 beneficiaries between 2004 and 2008.97 This may be attributable to factors such as the availability of Anti-Retroviral drugs. Labor Intensive Public Works and Drought Relief Programs 15. The Labor Based Drought Relief scheme was started in the 1960s as an emergency response to alleviate the effects of drought. In the 1960s and 1970s payment for participating in drought programs was in the form of “food for work”. The objectives have now shifted to focus on the provision of temporary supplement to the rural poor through wage employment for those most affected by drought. There are also Labor Intensive Public Works (LIPWs) programs that are not dependent on declaration of drought with the same overall goal of supplementing incomes and creating socially useful or productive. The programs are coordinated by the MLG and implemented by the local authorities. Beneficiaries are all poor able-bodied individuals in the rural areas. No means testing is conducted, but wages are set at a low level to encourage self-selection. Under the 2003/04 drought, the cumulative number of people employed was 121,599 workers, against a total of 1362 projects. To address poverty and unemployment problems brought about by the global crisis, the Ipelegeng (self-reliance) LIPWP was introduced in July 2008, and 97 NDP 10, Table 10.1. by October 2009 the cumulative number of beneficiaries under the program was 327,494 people.98 The beneficiaries of these programs tend to be mostly women (on the order of 80% or so). World War II Veterans Allowance 16. The WWII Veterans Allowance is another universal entitlement program which is not means tested. An allowance is paid to each WWII veteran or his surviving spouse or children less that 21 years of age, in recognition of the services they rendered In 2006, there were 4033 beneficiaries. The program is administered bythe Commissioner for Social Benefits and implemented under the office of the District Commissioner/Officer in various districts. Beneficiaries receive their allowances from post offices. Remote Area Development Program 17. The Remote Area Development Program (RADP), has after a number of reviews, evolved to focus on ensuring that beneficiaries achieve sustainable social and economic development and tthat they benefit from the rapid economic development of the country. The origins of the Remote Area Development Program date back to the 1970s when it was called the Bushmen Training and Settlement Project. Beneficiaries are marginalized communities that mainly live in remote and arid parts of western Botswana where there is very little economic activity. These communities have traditionally been nomadic hunter-gatherers who did not engage in arable agriculture. Many are Basarwa, but there are other groups as well. Over time there has been encroachment into the traditional hunting and gathering areas of these groups and many are today characterized by sever poverty, loc of income, low literacy levels and lack of employment opportunities. The Program is implemented by the MLG in over 60 designated settlements through the Department of Social Services. Beneficiaries receive the destitute rations and allowance. Other SSN provided include the old age pension scheme, assistance for orphans and vulnerable children and services for people who are on home based care. In addition to SSN benefits, an Economic Promotion Fund has also been established to create employment opportunities for remote area dweller, and provides funds for productive and business oriented activities including game ranching, harvesting and utilization of veldt products and arable agriculture. 18. In addition to the programs mentioned above, there are a variety of other schemes that would certainly classify as SSN programs, but information on them is fragmentary and incomplete. These include such programs as Housing for the Poor, child day care centers, exemption from secondary school fees for certain population categories, additional programs for vulnerable children, community homes, and others. 19. In particular, a significant proportion of spending in the agriculture sector could also be seen as directed towards income support and risk management for the rural poor. This includes such programs as Field Crop Production Improvement for Communal Farmers, Small Stock Program, Integrated Support for Arable Agricultural Development (ISPAAD) and its predecessors the Arable Lands Development Project (ALDEP) and the Accelerated Rainfed Arable Program (ARAP). Many of the agricultural subsidy programs have been in existence for decades under different guises, but with an unaltered objective of supporting smallholder arable agriculture. They have almost invariably enjoyed little success and reviews have generally noted that the large sums of money spent have had no discernible impact on agricultural production. Many intended beneficiaries are today quite diversified 98 State of the Nation Address, 13 November 2009. and are no longer as dependent on subsistent rain-fed agriculture, suggesting that a rethinking of these subsidy schemes is in order. B. Does SSN spending in Botswana reach its targeted beneficiaries? 20. The evidence is mixed as to whether the various SSN programs reach their intended beneficiaries. The large SSN programs in Botswana – the Old Age Pension, and Primary School Feeding - are universal and not means-tested, and the question of targeted beneficiaries is somewhat irrelevant. Others are based on narrow categories of the population (Orphans and Vulnerable Children), where targeting is simplified, but where special eligibility criteria, beyond the simple fact of being poor, reduce coverage. 21. A satisfactory answer to the question is made more difficult by the paucity of data. The available Household Income and Expenditure Survey (HIES) data are severely out of date. Botswana carries out a HIES only every decade, while best international practice is every 5 years at a minimum, and ideally, annually. A detailed incidence analysis was conducted by BIDPA on HIES 2002/03, and some of the BIDPA tables and findings are replicated in Appendix 3. The main findings are: • Government programs reduce poverty by 7 percentage points • Targeting efficiency ratios vary by program type • Coverage rates are quite low 22. In some respects, SSN impacts have been quite good. SSNs focus largely on children and the elderly, and these are the population categories characterized by the highest poverty headcounts. As the BIDPA study found, if government transfers were to be removed, household level poverty would rise by 7 percentage points, and it would increase by 10 percentage points at the individual level, and the impact is largest in rural areas where poverty is highest (Annex Table A.7) .99 Another study finds that the orphan allowance has had a significant poverty-mitigating impact. Modeling done by Econsult (2006) indicates that the orphan allowance reduced poverty nationwide by at least 1 percentage point. The poverty- reducing impact of this program is greatest in the Rural North-East (where poverty was reduced by 3 percentage points) and Rural South-West regions (2 percentage points). Supplementary feeding programs, as noted above, reach about a third of the population and have been central to improving nutrition in children. 23. Botswana’s SSNs cover all the age cohorts and address some of the risks that each cohort faces. The orphan food basket and the VGFP address the risk of malnutrition in under-fives; school feeding, orphan care and the needy student allowance (under the Destitute Program) cover some of the risks faced by older children; working-age adults of 21-64 years are assisted largely by LIPWs and a few can find assistance in CHBC; and the elderly have the OAP and the Destitute program. 24. Regarding effectiveness of SSNs, there are concerns, however, with the adequacy of the programs in terms of their coverage, targeting and their ability to address risks and vulnerabilities. Many needy Batswana are not provided with a SSN program if they fall outside the narrow definitions of 99 This can be seen as an upper bound of the impact of SSNs on poverty as it assumes that households will not explore alternative livelihood in the absence of government transfers. vulnerable groups. Universal programs like school feeding and the old-age pension do cover a large part of the population, but they do so only incompletely.100 Most beneficiaries of SSN programs are not poor. The poor, those falling in Quintile 1, account for just one third of beneficiaries (Annex Table A.2). Government transfers constitute a significant portion of expenditures for households in quintiles 2 and 3, and there is no discernible relationship between the size of mean government transfers per household and per capita expenditure quintile (Annex Table A.3). Many non-poor are be benefiting from SSNs. 25. All the SSNs combined are associated with very low coverage, reaching only an estimated 19 percent of poor households. With the exception of the school feeding program, all programs had coverage ratios of less than 10 percent of poor households. In terms of targeting efficiency (the inverse of the leakage ratio), it is estimated that less than half – or 43 percent – of households covered by SSNs are poor, although this varies by program, with the RADP and labor based drought relief programs performing very well, followed by the destitute package and the clinic ration.101 The OAP, and school meals perform less well as they are universal and not means tested, as does the CHBC program (Annex Table A.5). III. Overall SSN Expenditure Level 26. Estimating SSN spending in Botswana is difficult. As was noted in Chapter 3, expenditures overall are not are not classified on a programmatic basis and SSN and other poverty-reducing expenditures are neither clearly defined nor tracked in any manner. Information is aggregated in an often confusing way. For instance, the functional classification of expenditures includes a “Food and Social Welfare Programs” category, but without further disaggregation by program or program type.102 Upon closer inspection of this category it becomes clear that expenditures which would not normally be considered part of the SSN umbrella are included, such as “Trade Fairs, Agricultural Shows and Exhibitions”, “Construction of Border Fences”, “Food Research and Testing Centre” etc.. There is no one place where SSN expenditures are consolidated, and it is not easy to reconstruct the amount spent on SSN from the budget books as it is difficult to ascertaining where certain program expenditure occurs. Programs sometimes shift from one ministry to another, are allocated to recurrent or to development budget in no easily understandable manner, etc. Information obtained from MLG departments does not always correspond to figures found in the budget estimates. 27. Nevertheless, an attempt has been made at a best estimate of SSN spending by the public sector, covering the nine major programs outlined above. The information is based on the annual Estimates of Expenditure from the Consolidated and Development Funds and the Annual Statement of Accounts. Table 2 and Figures 1and 2 summarize this information. 28. Contrary to some previous assessments, SSN spending in Botswana is quite high, and has been increasing as a proportion on total government spending and GDP in recent years. The 100 In particular, there is no mention of any feeding programs for children outside of primary school for the weekend or during school breaks. 101 The clinic ration is part of the VGFP which is supposed to cover all under-fives during drought years, and medically targeted under-fives during normal years. It appears that in reality all under-fives are covered in all years. 102 The Central Statistical Office 2008 yearbook, Tables 2.7.4A and 2.7.5A. largest program in terms of expenditures in recent years has been the Orphan Care program. It is followed by the Old Age Pensions and the two Supplementary Feeding Programs. These three programs constituted fully two thirds of total SSN spending in 2007/08. Depending on economic and weather- related circumstances, the Labor Intensive Public Works program has also represented a significant portion of SSN spending in some years. With the onset of the global crisis, resources have been diverted to the Ipelegeng LIPW, which may today be the largest SSN program in Botswana. Table 2: Social Safety Net Expenditures, million pula P O A P S F P O C C H B C R A D P LIPW D R WWII V D T SSN S I P S V G F P Figure 1: Social Safety Net Program Composition Figure 2: SSN Expenditures, % of Total Spending & GDP RADP WWII D CHBC F OAP LIPW O GDP Source: Bank Staff calculations based on Annual Statements of Accounts and CSO 29. It is important to note that the numbers for many of these programs include only the benefits received, and do not include the associated administrative costs of delivering the benefits. On the other hand, the expenditures on LIPWs and the RADP do include the cost of materials and supplies used to construct or maintain the works in question, and so exaggerate the benefit received. In the absence of better information, however, these are our best estimates and provide a reasonable idea of the public expenditure effort in the SSN area as well as spending trends over recent years. 30. How do SSN spending levels compare internationally? Again, comparisons are difficult due to the challenges surrounding adequate information, outlined above. Cross-country comparisons are problematic as SSN programs will not be fully comparable. But at 3.2 percent of GDP, Botswana spends significantly more on its safety nets than the country average of 1-2 percentage points of GDP as suggested in Grosh et al. 2008 (Figure 3). Box 2 summarizes some of the characteristics of SSNs found in MICs. Figure 3: SSN Expenditures as % of GDP in Selected Countries, 2000s C C M M R B A C R V T E B S A U M E Source: Tesliuc, 2008, compiled from World Bank Public Expenditure Reviews 31. Resources allocated to SSNs do not appear to be an issue. Public spending on SSNs in Botswana is significant. A rough calculation, consisting of dividing estimated SSN spending (Table 2) by the number of poor Batswana (under the assuming that it is still at 30.6 percent of the population as in 2002-2003) indicates that per capita SSN expenditure stand at BWP 2000 annually, not counting the myriad smaller programs aimed at supporting the incomes of the poor. Per month, this amounts to about BWP 167, very close to the Poverty Datum Line for a household with a single individual (which is on the order of BWP 200). IV. Major Challenges Facing the SSN in Botswana and Some Options to Tackle Them 32. There are a number of important areas that require the attention of policy makers if Botswana’s SSN is to adequately fulfill the needs of the poor. These are outlined below. A. Better information is needed to identify the poor and to assess the impact of SSN programs 33. Identifying the poor and their characteristics is a first required step to the design of appropriate SSNs. But information on this is sorely lacking in Botswana. As already noted, there is no up-to-date poverty assessment and the latest HIES was conducted ten years ago. A HIES is being conducted now by the CSO, but ten-year intervals are simply too long. The Safety Net has not been studies systematically (except for the BIDPA study cited here) and there are no impact evaluations for any SSN program. Access to high quality, current information as a basis for policy-making needs to receive higher priority. Without adequate information it is today not possible to assess whether universal, categorical, or means-tested benefits would be the most appropriate. An impact evaluation of SSN is needed in conjunction with the HIES. 34. A recommendation is that at a minimum, a HIES with an impact evaluation module should be carried out every five years, and ideally this would be done annually (such as in Turkey after it reformed its SSN to provide conditional cash transfers targeted to the poorest). Further, a beneficiary assessment is needed to provide information on whether or not SSN spending is helping people to move out of poverty.103 B. Improved targeting of beneficiaries would increase the efficiency and effectiveness of SSNs 35. If Botswana wishes to improve the impacts of the SSN and the use of public monies, more thought needs to be given to targeting beneficiaries. While Botswana is spending a significant share of GDP, well in the medium-range for middle-income countries, poverty and exclusion and leakage rates are quite high. Poverty affects nearly 31 percent of the population, and at first glance, the fact that 26 percent of households report some receipt of a transfer, it could appear that Botswana’s safety net is adequate. However, those 26 percent in receipt are not the poorest. Only one-third of the lowest 20 percent by income received a benefit, while 6 percent of the richest quintile report receipt. The exclusion problem (two-thirds of the poorest 20 percent of the population get nothing) is worse than the leakage problem, but any leakage to households above the poverty line reduces the amounts available for targeting to the poorest. 36. Botswana’s safety net would be improved if it could cover more of the poorest. Few benefits target the poor as a broad group. Some programs are universal and not means tested in any way, others benefit only a very specific sub-group of the poor. The Orphan Allowance program does not attach any conditions on school attendance or health clinic visits. While an income test is used for the destitution benefit that relies on extensive reporting and involvement of social workers. The destitution allowance may also benefit from an expansion and a refocusing to allow for coverage of poor children who have not lost parents, but whose parents do not earn enough income to cover the needs of the family. 37. There are three large universal/ categorical programs which be definition benefit far more non-poor than poor people. These are the old age pension, the orphan benefit and school feeding. All of these programs would benefit from enhanced targeting, although it is unclear whether Botswana’s social compact would allow a change to the eligibility rules for these programs. Nevertheless, if the SSN budget is to remain constant the only way to ensure that SSNs reach the poorest is to institute targeting to untargeted programs. 38. One recommendation would to introduce targeting in at least school feeding programs, such as has been done successfully in other middle-income countries such as Jamaica and Chile. “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. If Botswana were 103 The World Bank is automating tools to be used with STATA for poverty analysis, including targeting and measuring errors of inclusion and exclusion (ADePT and SP Adept). These new tools paired up with more frequent HIES data, could guide Botswana in a reform of its SSN. The ADePT tools are described at: http://go.worldbank.org/CXMO0VQ9D0 able to generate significant cost savings from targeting school feeding (currently 75 percent of expenditure on school feeding goes to non-poor households), after administration costs, there might well be enough savings to warrant a moderate expansion in the best-performing programs, armed with impact evaluations and after such programs had been reformed with better targeting (Bundy et al, 2009). 39. To improve targeting many middle income countries have experimented with an alternative approach – the proxy means test – developed in Latin America but now in use in MICs across the world. The advantage of proxy means testing is that it requires less information than true means testing, and yet is objective. In proxy means tests, rather than trying to measure total income perfectly, information is collected on items which are much easier to measure and verify, such as the number of children in the family, the location and quality of its dwelling, its ownership of durable goods, demographic structure of the household, and the education and, possibly, the occupations of adult members. These variables should be ones which are known to correlate with poverty in the country, and ideally, which are easy to measure and thus require little administrative cost to verify. Because it does not measure income itself, proxy means testing may discourage work effort less than a means test would. 104,105 Box 2 - Safety Nets in Middle Income Countries “Most middle-income countries have partial, but significant, safety net systems, although a great deal of diversity is apparent. In countries in the Middle East and North Africa, expenditures can be significant, and the main instruments are general food subsidies and, often, fuel subsidies. Many Latin American countries have a long tradition of truncated welfare states, with social protection systems built on contributory social insurance based on attachment to the formal labor market but large informal sectors. Because as many as half of all workers, and a much higher proportion of the poor, work in the informal sector and are not covered by social insurance, they derive little benefit from such schemes. An increasing number of these countries have begun to develop the social assistance side of the welfare state by creating conditional cash transfer (CCT) programs, several have adopted noncontributory pensions, and a few have extended health insurance or fee waivers to the poor. Countries of the former Soviet Union have moved away from the Soviet legacy of categorically based in-kind “privileges” to targeted cash transfers and child allowances. East Asian countries’ publicly supported safety nets are still quite small, and the countries tend to rely on high growth and family support to prevent poverty. Complete safety nets cover the chronic poor, those hit by shocks, and those with special vulnerabilities. Few middle-income countries have achieved full coverage in all three areas; in general, the programs for the chronic poor are the best developed. These countries are relatively unconstrained in their choices of 104 how to construct their safety net systems. Their selection will depend on the starting point. Many countries The first large-scale use of proxy means-testing occurred in Chile in the late 1970s and 1980s, in a program are the called introducing Ficha CAS targeted cash (card for transfer social systemsSince assistance). to meet theCosta 1994, needsRica of the chronically and poor,adopted but for proxy Colombia have different means- tests for some reasons: in of their social Europe Asiaprograms, assistance and Central to replaceas have Mexico, a system Honduras, of privileges Argentina, unrelated Venezuela to economic need;andinNicaragua. Latin Outside Latin America America, and Turkey, the Caribbean towith World address theBank support, truncation of uses a proxystate; the welfare means test and informula for a large a few countries national elsewhere social assistance/conditional cash program. The Caucasus countries (Armenia, to replace substantial, untargeted commodity subsidies. Programs aimed at vulnerable groups Georgia, Azerbaijan) are also using a are Proxy Means Test 105 increasingly a focus of attention, in Europe and Central Asia because the inherited system of The indicators used in calculating this score and their weights are derived from statistical analysis (usually institutionalized regression analysis or social services yielded care components) principal poor of data fromcare and was detailed unsustainably household surveys of a sort tooand expensive, in to other costly be carried for out to regions all applicants address to large needs that programs. The had previously information met. provided by the applicant is usually partially not been Cash by verified will play the either collecting transfers information a larger role inon a visit middle-income stable, to the home by a program than official countries in or by having other the settings, applicant in part bring written verification of part of the information to the program office. Eligibility is determined because means testing and proxy means testing are feasible and will permit the good targeting required for by comparing the household’s score against a predetermined cutoff. From the World Bank website significant cash transfers. The programs may use one or more features to keep labor disincentives low. For http://go.worldbank.org/55Q72CB9Q0 instance, they might configure benefits to be higher for households with higher dependency ratios; they might add a requirement for public service, work, or enrollment in some sort of training or job search; or they might be complemented with an earned income tax credit. In countries with extensive health and education networks but inadequate use of these, benefits might be conditional on households obtaining adequate preventive care for themselves, undertaking health education, and/or enrolling their children in 40. Botswana is a good candidate for introduction of proxy means testing. Proxy means tests are most appropriately used in stable, middle income countries, where there is reasonably high administrative capacity, for programs meant to address chronic poverty in stable situations, and where they are used to target a single program with large benefits or to target several programs so as to maximize return for fixed overhead. The latter description would fit Botswana well. There are some drawbacks. Administering proxy means testing requires a body of literate and computer-trained staff, and moderate-to-high levels of information and technology. There is also an inherent inaccuracy at household level, since the formula is only a prediction, though good results on average have been observed. The formulae used usually rely on indicators that are fairly stable and may distinguish well chronic poverty, but can be insensitive to quick changes in household welfare or disposable income, which may be frequent and large when an economy is suffering a large downturn. C. Reaching the poor at reasonable cost 35. Until recently Botswana relied heavily on a cumbersome system of distributing benefits in kind as a major component of its SSN. While this may be appropriate for a few areas where e.g. food is not locally available, it is problematic in that in-kind benefits are often difficult to move around especially to remote areas and has high overhead costs. In 2009, a coupon system was introduced for destitution, orphan, and vulnerable groups food beneficiaries. Prior to this recipients had to go to designated stores to pick up their food rations, and there were anecdotal reports of senior citizens being robbed of food on the way back to their homes, while disabled elderly lacked the strength to pick up their food packages at a single time, necessitating many trips. Meanwhile, the old-age pension was distributed in cash, raising security issues. The new “coupon” system is actually not a paper coupon, but a smart card that can be used at any retail outlet that has the reader. However, the term “coupon system” is the common and accepted name of the program in Botswana hence its usage herein. 36. Expanded use of smart cards may also help reduce the capacity constraint that has been noted as an important issue in administering the SSN. There is an acute shortage of professional social workers to manage the SSN case loads. Social workers, are, for instance, required to conduct thorough assessments and registration of destitute persons. On average, one social worker covers at least five villages. In the case of the Orphan Care program, there were only 82 posts to manage a case load of 51,600 registered orphans, working out to a case load of 630 orphans for each post (2007). Yet these scarce social workers spend much of their time on clerical duties such as preparing tenders for food rations and clothing, ensuring provision of supplies, etc. 37. A general recommendation, given the success of the coupon system to date, is to expand the use of smart cards to all benefits. This will greatly facilitate access for beneficiaries and will cut down markedly on administrative costs. As of June 2009, approximately 45,000 beneficiaries out of the total for 96,000 in the three programs where the smart card was introduced had been registered in the new coupon system and 21,000 had already drawn benefits. It is expected that the new coupon system will greatly reduce costs to recipients and enable them to withdraw their benefits gradually over the month rather than picking food up in bulk once or twice a month. It is also expected that if the pilot continues to go well, that other benefits including the two benefits in cash (the old-age pension and the cash allowance for destitute persons) could be switched over to the coupon system.106 Smart cards could also unify the two halves of the destitution benefit (cash and food). D. Planning, Budgeting and Monitoring for Service Delivery 38. Given the uneven progress in poverty reduction in the midst of decades of sustained economic growth, income inequality and the negative impact of HIV/AIDS on poverty incidence, an important question is how has public spending helped to alleviate the condition of the poor? Unfortunately, this question cannot be answered empirically in Botswana. Given poor data, analysis and tracking of expenditures, it is not possible to know if public spending levels are adequate, if they are having the intended impact in an efficient and effective way, if targeting has been effective in ensuring that the poor benefit, or if exit mechanisms have been effective and if SSN programs are affordable in the long run. 39. At the most fundamental level, there is no clear idea of what is being spent today on the SSN. Botswana, as has been noted above, does not have automated systems to track and monitor SSN spending by program. The Budget Estimates themselves do not indicate expenditure on SSN programs, with the treatment of food expenditures being particularly opaque. As of this writing and after much dialogue, food spending on primary and secondary schools has been manually identified, but not on the food portion of the destitution benefit. No operating costs associated with SSN programs are reported in the budget and so it is impossible to establish how much it costs the government to transfer the program’s multiple benefits. Government spending figures are reported by individual program benefits, but often are either incomplete or inconsistent. The emphasis is on closely tracking inputs and output, but less on tracking progress towards attaining intended intermediate and final outcomes – outcomes which in any case are not often clearly articulated. 40. There is also no information on who the beneficiaries of SSN programs are. There has been no analysis on the incidence of public spending, and none of the programs are tracking beneficiary-level data. As a result, there is little knowledge of the characteristics of the beneficiaries or the benefit levels accruing to them. No impact evaluation planning exists (although this may be now changing with renewed emphasis on M&E, spearheaded by the Government Implementation Coordination Office, GICO). 41. The introduction of a public expenditure tracking system (PETS) would help fill the information gap regarding the effectiveness of SSNs. PETS is a technique for tracking the effect of public expenditure on growth and/or on social outcomes. It explores the ways in which public expenditures become public goods. A PETS enables a country to gauge the level of central budget allocations that are actually spent on the final service objectives. For Botswana, this would consist of mapping expenditure flows from the Estimates to specific SSN allocations and then tracking that money down to the local level. PETS can help identify bottlenecks to reform and suggest alternative channels for funds. By following the flow of funds within and between organizations, PETS yield useful data for institutional analysis. 106 For the Old Age Pension benefit and the cash portion of the Destitute benefit, the distribution of cash in cash is risky and administratively costly. Botswana lacks a broad banking network outside of Gaborone but cash could be sent through new innovative methods, such as the case in neighboring countries which have successfully used cell phones to transfer money. Mobile telephone coverage in Botswana exceeds 100 percent. 42. As diamond revenues decline, responding to emerging resource constraints in a rational manner is complicated by the way planning, budgeting and monitoring is carried out for SSN programs and advocates for a shift to budgeting for programs and for a transition to a MTEF as argued in chapter 3. With no information on how much is being spent overall on SSNs, who the beneficiaries are, whether or not spending under the various programs has been effective, and in the absence of clear targets or objectives for expenditures, properly informed planning and budgeting simply cannot take place. A case in point has to do with HIV/AIDS. HIV/AIDS is a major SSN challenge, but budgeting for future expenditures cannot rely on the current practice of examining what was spent the previous year and then adding on a small percentage. A long-range forecast for the number of OVCs to be cared for under the SSN program is urgently required, for instance. Current budgeting practices make this difficult. Moreover, the fragmentation of the SSN into diverse programs means that it is not possible to monitor and evaluate whether overall spending is effective in helping the poor. V. Conclusions 43. This chapter has suggested that the SSN in Botswana is fragmented and hidden in various indirect social expenditures by the State; it does not adequately cover the poor. Though a substantial portion of the poor in Botswana are probably provided with some basic essential services to sustain their livelihoods, these schemes are partial and are not intended to provide skills to enable the poor to get out of the poverty trap. Broader programs covering a larger proportion of the poor would probably be a more equitable and more efficient way for government to target social spending towards the worse-off, along with requiring the better-off to meet an increasing proportion of the cost of the social services they consume. For instance, structural poverty in rural areas increases vulnerability to drought and there is a case of reorienting or augmenting the existing drought-relief machinery to provide some form of poverty relief or social security in both drought and non-drought years and to support a considerably expanded program of longer term interventions with the aim of alleviating poverty. Botswana also still needs to define minimum standards of social security that depend on what is efficient, equitable and affordable. 44. Given the high levels of poverty the need and demand for SSNs is unlikely to subside, yet there is real likelihood of constrained budgets in the years to come, as discussed in chapter 2 of the report. In order to achieve a higher impact with the same level of resources, which in any case appear to be adequate based on international comparisons, a number of options have been suggested (Box 3 notes some of the pitfalls of reforming SSNs that should be kept in mind). The main recommendation is to improve targeting of SSNs. A suggested avenue is to begin targeting the school feeding program, and to move on from there to other well-performing universal programs. The option of using proxy means testing as opposed to full means testing could be considered given conditions in Botswana. A critical requirement for a successful refocusing of SSNs, however, is the availability of much better information on who the poor are and on the impacts of SSNs. More frequent HIES, with impact evaluation modules are needed; a PETS would also be helpful in gauging the effectiveness of public spending. 45. A transition to a planning and budgeting system that enables the organization of expenditures by programs and encourages forward-looking budgeting would help policy makers identify some of the inefficiencies and challenges outlined in this chapter. It would also facilitate the tracking, monitoring and evaluation of expenditures, and ultimately, the improvement of service delivery. An effective welfare scheme needs to, at a minimum, answer two key questions: Who are the target beneficiaries? And How can resources best be delivered to them? The current collection of disparate programs neither asks nor answers these questions in a satisfactory manner. Botswana already allocates a significant amount of budgetary resources to SSNs as defined by the nine programs outlined in this chapter. But the results in terms of poverty reduction have not been commensurate with the volume of resources. If additional resources such as agricultural subsidies are added to “core” SSN spending, it becomes fairly clear that Botswana could alleviate more poverty for less money than is happening today. Box 3: Common Pitfalls in Reforming Safety Net Systems When developing reform plans, avoiding common pitfalls such as the following is important. • Having unrealistic expectations. Safety net programs can never fully compensate for macroeconomic instability or eliminate the causes of poverty, although they can be helpful when used in conjunction with policies that address the root problems of these. • Avoiding conflicts between policies and programs. Safety net programs cannot be expected to, for example, solve an unemployment problem caused by skills shortages, rural poverty caused by distortions in agriculture markets, or excessively restrictive labor market regulation. • Avoiding having too many programs. International experience is rife with countries that have too many programs, each with low coverage, low benefits, inadequate administrative systems, and high overheads. Having fewer, larger programs would allow them to achieve economies of scale. In countries with too many programs, they often overlap and are not sufficiently coordinated to achieve the best possible synergies. • Preventing an imbalance in target groups. Programs may be excessively based in the formal sector or favor “virtuous groups” such as children or the elderly while failing to cover other groups, such as minorities or those with disabilities. • Rationing entry into a program by budget rather than by eligibility threshold. When the funding for a program is insufficient to allow all those who meet the defined eligibility criteria to be included, horizontal inequity is created, transparency declines, and opportunities for rent seeking are created whereby eligibility intake officers may demand bribes or give favorable treatment to those with whom they share an interest or affiliation. • Having insufficient administrative effort, monitoring, and evaluation. Programs are often set up in a hurry with only rudimentary systems. They may die altogether or fade away after a few years, especially if a change in government occurs. Developing and fine-tuning the most effective safety net systems takes time. • Creating a new intervention. Introducing a new safety net intervention is often tempting, especially when a major risk or cause of poverty is largely unaddressed. Despite a valid justification for the program in such cases, the value of the new program with regard to other uses of funds must be assessed. Establishing a new program that addresses an issue that should have been addressed by other, poorly performing programs is tempting, and particular care must be taken in such situations. Sometimes starting a new program is appropriate, as when small local programs cannot be scaled up without losing their effectiveness, but in many cases, starting a new program rather than resolving an old one’s flaws can prove costly in the long run. The forces that led to the need to reform the old program or that made it difficult to reform may, over time, affect the new program, leaving the country with two poorly performing programs. Moreover, neither program will have as much opportunity to achieve full economies of scale. ___________________ From Grosh et al., 2008, pp. 398-399 Appendix 1 : Summary of Baseline Assumptions A1.1 Macroeconomic Assumptions Real mining GDP growth rate Calculated based on the real mining GDP in NDP 10, Table 6.1. Diamond price and production volume are adjusted to be consistent with real growth rates of mining GDP. Real non-mining GDP growth rate Derived from the real growth rates of non-mining private sector and government provided in revised NDP 10 Table 6.1. Real GDP growth rate Derived from real mining and non-mining GDP growth rates. Nominal GDP Derived from real growth rates and inflation. See the table below Inflation Projections taken from IMF WEO (October 2009) in CY converted to FY. Exchange rate Projections taken from IMF WEO (October 2009) in CY converted to FY. Nominal US interest rate Projections taken from IMF WEO (October 2009) in CY converted to FY. Nominal interest rate on new Taken from GDF (CY converted to FY). See the table below. external borrowing Nominal interest rate on domestic Assume a spread of 2 percentage points over the nominal domestic interest government debt rate. See the table below. 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 Nominal GDP (millions of Pula) 84,184 92,330 108,085 127,776 153,977 172,968 178,627 Mining GDP (millions of Pula) 21,617 24,066 33,362 45,505 62,787 71,063 64,659 Non-mining GDP (millions of 62,567 68,264 74,723 82,271 91,190 101,905 113,968 Pula) Diamond production (millions of 22.1 23.4 29.5 33.5 37.0 38.1 33.0 carats) Diamond exports (millions of US 1,765 1,888 2,442 3,176 4,175 4,489 4,085 dollar) Nominal US interest rate (percent 1.2 1.4 4.8 4.8 4.8 4.8 4.8 p.a.) Nominal interest rate on 9.86 8.28 7.81 7.35 7.22 8.00 8.00 government debt (percent p.a.) Nominal interest on new external 0.68 1.40 2.51 4.09 4.18 4.50 4.50 debt (percent p.a.) A1.2 Fiscal Policy Assumptions Revenue Revenue policy No new revenue policy: no new revenue instruments; no increase in tax rates; no increase in fees and charges in real terms. Expenditure Wage bill Employment to increase by 1,610 in 2009/10 based on the 2009/10 Budget. From 2010/11, there will be no net increase in established posts (in line with NDP10 para. 6.51). From 2016/17, the establishment is assumed to grow, in net terms, by 1.5 percent each year. No cost of living adjustment in 2009/10 in line with the 2009/10 Budget. Thereafter, wages will be adjusted annually at the inflation rate. Other recurrent spending Assumed to grow at the rate calculated from Table 6.6. The 2009/10 figure is taken from the Budget (Table V). Development expenditure For 2009/10, the number is taken from the Budget table. From 2010/11, assumed to grow at the rate calculated from Table 6.7. On-lending No new on-lending. Financing External financing New external borrowing in 2009/10 as advised by the MFDP (August 2009). The government is assumed to comply with the legal external debt limit of 20 percent of GDP. As external PPG exceeds the limit by end-2009/10, no new borrowing is assumed to take place until subsequent GDP growth provides room for additional external borrowing. Once some room becomes available, it is assumed that the government prefers external borrowing to domestic borrowing due to lower costs. Domestic financing Borrowing plan for 2009/10 provided by the MFDP (P 1.5 billion) as T bills and government bonds. The government complies with the legal domestic debt limit of 20 percent of GDP. Domestic borrowing serves as a main borrowing instrument while external debt constraint is binding. A1.3 Fiscal Projections Revenue Mining royalties, dividends, and Projected by assuming 60% of diamond exports as government’s share of mineral income tax Debswana’s profit (based on historical pattern). Non-mining income tax Projected by applying the effective tax rate of 7% (historical average) on non-mining GDP. Export duties Assumed to grow at inflation rate. SACU revenue Projected based on the medium-term projection of SACU revenue pool (available from South Africa’s budget, Feb 2009) and historical average of Botswana’s share in the revenue pool (20%). After 2012/13, SACU revenue is assumed to grow at the rate of inflation. VAT Projected by applying effective tax rate of 3.5% (historical average) on GDP. Other taxes Assumed to grow at inflation rate. Interest receipts Projected by applying the effective interest rate of 7% (historical average) on outstanding balance of PSDF/RSF/DF at end of previous FY. Balance of PSDF/RSF/DF for 2008/09 taken from BoB FS. BoB revenue Projected by applying the projected US nominal interest rate on the balance of GIA outstanding at end of previous FY. Fees and charges Assumed to grow at inflation rate. Grants US$300 million assumed every year Expenditure Wages and salaries Projected based on actual employment and average salary per person per year. Actual employment in 2008/09 derived by assuming that employment in 2007/08 (provide in NDP10, Table 6.10) grew by 4.4% (average employment growth rate during NDP9). 2009/10 employment is derived by adding newly established posts (1,600). The 3,926 vacant positions (outstanding at end- March 2008) are assumed to be filled evenly during 20010/11-2014/15. External interest payment Derived by constructing payment schedule for each of Botswana’s 101 external loans, including new borrowing expected within 2009/10. Payment schedules are calculated using original currency and converted to US$ using projected exchange rates (Oct. 2009 WEO) up to 2014/15, and thereafter assume the same rates as 2014/15 throughout. Interest payments on new external borrowing after 2009/10 are projected by assuming nominal interest rate on new external borrowing and balance of new borrowing at end of preceding fiscal year. Domestic interest payment Projected by applying projected domestic interest rate on outstanding balance of domestic debt at end of preceding fiscal year. Repayment of PSDF/RSF Assume outstanding balance at end 2008/09 will be repaid by 2013/14. Financing Amortization of external debt Derived from the payment schedules constructed for each of 101 loans, based on the information provided by the Common Wealth Secretariat. For new borrowing after 2010/11, assume 5% of outstanding balance of preceding FY is amortized. Amortization of domestic debt Assume all government securities have maturity of 2 years. Changes in cash balance Assume this reflects only capital paid into (negative) and withdrawal from (positive) from the Pula Fund. Appendix 2 Public Expenditure and Financial Accountability Assessment Summary of Findings Botswana undertook a Public Expenditure and Financial Accountability (PEFA) PFM Performance Assessment during 2009. The assessment approach was based upon a careful consideration of demonstrated observable public financial management (PFM) systems, procedures and practices in Botswana, across the standard 31 high-level indicators. It was evidenced through interviews with Government officials, reviews of official documents and reports and corroborating evidence from a variety of independent sources where possible. The purpose of the PFM Performance Report is to assess the current status of the public financial management system of the central government, in order to support fiscal discipline, strategic allocation of resources, effective and efficient service delivery and accountability. In so doing the Government is assisted in defining its PFM reform priorities, subsequent reform activities and sequencing of implementation. Further, it should serve as a useful tool to the dialogue between Government and cooperating partners about the provision of harmonized support to the Government’s PFM reform efforts. The performance of PFM systems, procedures and practices as measured through the PEFA PFM Performance Assessment, are described in terms of six critical dimensions of PFM : 1) credibility of the budget; 2) comprehensiveness and transparency; 3) policy based budgeting; 4) predictability and control in budget execution; 5) accounting, recording and reporting; and 6) external scrutiny and audit. Donor practices were also assessed and considered by way of their impact on each of the six critical dimensions of PFM. The overall assessment results of all 31 indicators are presented in Table 1 below. Table 1 – Overall summary of PFM indicator scores Scoring Dimension Ratings Overall PFM Performance Indicator Method i. ii. iii. iv. Rating A. PFM-OUT-TURNS: Credibility of the budget PI-1 Aggregate expenditure out-turn compared to original approved budget M1 C C PI-2 Composition of expenditure out-turn compared to original approved budget M1 C C PI-3 Aggregate revenue out-turn compared to original approved budget M1 A A PI-4 Stock and monitoring of expenditure payment arrears M1 NS D B. KEY CROSS-CUTTING ISSUES: Comprehensiveness and Transparency PI-5 Classification of the budget M1 C C PI-6 Comprehensiveness of information included in budget documentation M1 B B PI-7 Extent of unreported government operations M1 A A A PI-8 Transparency of inter-governmental fiscal relations M2 D B C C PI-9 Oversight of aggregate fiscal risk from other public sector entities M1 C D D+ PI-10 Public access to key fiscal information M1 B B C. BUDGET CYCLE C(i) Policy-Based Budgeting PI-11 Orderliness and participation in the annual budget process M2 B C A B PI-12 Multi-year perspective in fiscal planning, expenditure policy and budgeting M2 D C A C C+ C(ii) Predictability and Control in Budget Execution PI-13 Transparency of taxpayer obligations and liabilities M2 B C B B PI-14 Effectiveness of measures for taxpayer registration and tax assessment M2 A A D B PI-15 Effectiveness in collection of tax payments M1 B C D D+ PI-16 Predictability in the availability of funds for commitment of expenditures M1 D A C D+ PI-17 Recording and management of cash balances, debt and guarantees M2 D C C D+ PI-18 Effectiveness of payroll controls M1 A B A B B+ PI-19 Competition, value for money and controls in procurement M2 D C C D+ PI-20 Effectiveness of internal controls for non-salary expenditure M1 B A C C+ PI-21 Effectiveness of internal audit M1 B C B C+ C(iii) Accounting, Recording and Reporting PI-22 Timeliness and regularity of accounts reconciliation M2 B B B PI-23 Availability of information on resources received by service delivery units M1 A A PI-24 Quality and timeliness of in-year budget reports M1 C A A C+ PI-25 Quality and timeliness of annual financial statements M1 B B C C+ C(iv) External Scrutiny and Audit PI-26 Scope, nature and follow-up of external audit M1 B D B D+ PI-27 Legislative scrutiny of the annual budget law M1 B A B A B+ PI-28 Legislative scrutiny of external audit reports M1 C A A C+ D. DONOR PRACTICES D-1 Predictability of Direct Budget Support M1 A D D+ D-2 Financial info provided by donors for budgeting/reporting on proj./program aid M1 C D D+ D-3 Proportion of aid that is managed by use of national procedures M1 D D Summary findings 1) Credibility of the budget Botswana has a dual (separate capital and recurrent) annual budget system which is based on a medium term national development plan. The durations of the plans are normally 6 years with a review half way. The development / capital expenditures are financed through the Development Fund (DF), mainly by development aid, loans and transfers from the Consolidated Fund. Recurrent expenditures of ministerial votes are funded through revenue deposited into the Consolidated Fund. Botswana bases its budget ceilings on an annual fiscal forecast and resource envelopes of are pre-announced, with strong bottom-up inputs from the budget entities. Separate procedures and timetables are applied to prepare the two budgets, through comprehensive budget circulars. Budgets are normally approved four to six weeks before commencement of the financial year. Revenue Robust revenues from mining royalties and taxes ensures revenue adequacy, but the revenue estimate system lacks careful macro-economic considerations in the development of a multi-year revenue forecast frame and favours conservative projections. Although on the one hand this appears like a sound starting point for a credible budget system, it may limit opportunities for achieving full effectiveness with respect to stated national strategic objectives. Coupled with a successful debt management strategy, evidence suggests that the excess revenue over budget has been channelled principally to investment funds and debt reduction, leading to a dramatic decline in debt from 13.5% (2004/2005) of GDP to about 10.2% (2006/2007). This has provided space to develop a meaningful fiscal framework to serve as an effective and realistic top-down budgetary discipline tool, which in principle should allow Botswana to meet service delivery demands. Expenditure At the aggregate primary expenditure level Botswana presents a mixed picture with respect to the credibility of the budget. It is clear from Table 2 below that aggregate expenditure out-turns has not matched the budget estimates very closely, bringing into question budget credibility and the capacity to implement the budget as approved. Three issues appear to underpin this result – (i) weak procurement planning; (ii) weak procurement management; and (ii) weak capacity to implement development projects. Table 2: Budget outturn Revenue 2005/06 2006/07 2007/08 Revenue budget 24,907 29,629 34,585 Revenue Actual 26,098 31,386 37,170 Deviations, Pula million 1,191 1,757 2,585 Deviation % 5% 6% 7% Expenditure Budget 19,285 22,191 26,342 Expenditure actual 17,620 19,823 22,444 Deviations, Pula million -1,665 -2,368 -3,899 Deviation % -9% -11% -15% In summary, much remains to be done to increase the perceived credibility of the budget as a tool for the implementation of national strategic objectives and policies. The budget chapter provides more detailed analysis and recommendations in this regard. 2) Comprehensiveness and transparency There is a fairly active culture of transparency with regards to fiscal information. Budget documentation is complete, comprehensible and comprehensive including the macroeconomic assumptions, the fiscal balance, the external debt profile and status, the financial assets, the historical budget outturns and some explanation on the impacts of new major revenue and expenditure policy initiatives. Budgetary, tax revenue, procurement and audit information is made available in a timely fashion on the Internet, through the government bookshop and through public and academic libraries. There is also information available on the amounts of resources received by the front line facilities such as primary schools and primary health care facilities. The Government has adopted GFS / COFOG classification standards for the budget formulation and execution, reflecting both economic and administrative classifications in budget and accounting information. There are adequate controls to ensure that revenue and expenditure are comprehensively controlled - all revenues generated directly by line ministries are transferred to the Consolidated Fund and all expenditures are made through a centralized payments system, using a single treasury account concept. The budget process and documentation provide transparent and reliable information on the allocations to be made to Sub-national Government, well in time before the start of their detailed budgeting processes. Subsequent budget releases are also timely and fully predictable. While fiscal reporting of provincial government is consolidated into the national financial reports, financial reporting of the Local Authorities is delayed and is still to be consolidated into annual reports consistent with central government fiscal reporting107. There are no unreported Public-Private Partnerships, nor unreported unconventional financing instruments for addressing losses of Public Enterprises (including any foreign exchange losses of the Bank of Botswana). All security agency funds are reported on in aggregate, even if details of expenditure remain undisclosed. The financial reports consolidation process includes a reconciliation process between sources of funds and applications which reasonably assures that there are no significant extra-budgetary funds outside cooperating partner funds. There is some financial monitoring of Public Enterprise operations but not yet a fairly careful risk assessment using a comprehensive risk analysis framework. Not all Public Enterprises are submitting timely audited financial reports. 3) Policy-based budgeting The national vision, mission and development objectives have been articulated within the Government’s National Development Plan. The strategic plan is updated every three years through a mid-term review. It represents a consolidation of fully costed sector strategies, which are developed within a fiscal frame that is consistent with macroeconomic forecasts. There is some consultation with Parliament on key issues during the budget preparation process. The annual budget process however does not rely on policy input by the cabinet at the beginning of the budget process. Line ministries select projects based upon program priorities that are 107 H determined by the Sector Strategies. There is approval by the cabinet at the end of the budget cycle and timely approval of the budget by Parliament. However, it is unclear whether the recurrent cost implications considered as part of the prepared strategies and the development budget are consistently included in the recurrent budget. There is also an apparent disconnect between budgets and the capacity to execute the approved budgets. 4) Predictability and control in budget execution Revenue and cash management Predictability in budget execution is premised upon cash adequacy, in itself critically dependent on sound revenue management. The newly formed Botswana Unified Revenue Service (BURS) is still evolving and developing its management systems. However, many elements of revenue management has been found to work reasonably well – including a complete database of taxpayers (linked to other relevant systems such as vehicle registration), clarity of taxpayer obligations and liabilities, constraints on officer discretion in the application of penalty waivers and rates, sustaining of tax awareness and educational programs, and the implementation of tax audits. Areas of concern include – (i) timely reconciliation of collections and transfers to the Consolidated Fund; and (ii) all categories of tax are not yet timely transferred to the Consolidated Fund. Cash and debt management are well managed to facilitate highly predictable budget releases. A centralised payment system for the Consolidated Fund is in place, providing for daily bank balance consolidation (except for the treasury accounts in some districts). Debt is monitored using the CS-DRMS system and regularly reconciled and reported on with respect to stock as well as debt service. Expenditure management The central government deploys a centralized off-the-shelf computerized financial management system (“GABS”) for budget and expenditure management, accounting and financial reporting by all votes. Budget release discipline is strong, supported by commitment control functionality in GABS. There are well controlled and documented virement procedures, supporting the transparent use of supplementary budgets with Parliamentary oversight. However, some expenditure is still managed manually with commitment control outside of GABS, leading to concerns that commitment control for development expenditure is not fully effective. Adding to that are concerns about – (i) procurement planning capacity; (ii) reporting on commitments; and (iii) a rush to spend unutilised funds at the close of the year, which contributes to working around procurement procedures and controls. Expenditure arrears accumulation is not monitored. This can be justified because of the cash accounting basis adopted by Botswana, but such an argument would not benefit improved financial management in the government. GABS does facilitate the recording and monitoring of accounts payable, which could be used to provide insight into the position on outstanding commitments and expenditure arrears. In addition, when one considers the variance of actual expenditure to budgets there is evidence that control over budget execution is not yet optimal. It points to a weak coupling between the budget formulation and implementation, leading to frequent in-year budget adjustments (see PI- 16). Payroll management is facilitated using a computerised payroll system, INFINIUM. This system directly links three databases: the establishment of posts, the personnel database which serve as control files, and the payroll database. Changes to these databases leave an audit trail and permit only selected access dependent upon function. The databases are encrypted. All civil servants are registered through INFINIUM that include appropriate fields to protect against duplication. There are effective controls with respect to the creation of new posts, that include budgetary controls, the hiring of new employees (controlled by the posts database), and the assignment of promotions, transfers, allowances and terminations. Further, through the use of verification procedures, exception reports and regular physical payroll audits, there is fair level of assurance of the integrity of the payroll management system. The Government is careful to address extra-budgetary risks by controlling line ministry revenues, unbudgeted bail outs, and unreported unconventional debt instruments to address losses in Public Enterprises (see PI-7(i)). However, factors that can undermine both the credibility of the budget and expenditure management include the poor monitoring of fiscal risk, debt and contingent liabilities that might emanate out of unforeseen expenditure burdens of sub national government or public enterprises (see PI-9). Controls and internal audit While effective controls exist for each of the main steps of the expenditure cycle, procurement controls remain less than fully effective primarily due to capacity constraints. Internal audit is fully operational across most line ministries, with audit work plans suggesting that approximately 50% of staff time is applied to systems audit and regular reporting on audit activities. Internal Audit adopts audit standards specified in internal audit manuals. Quality assurance exercises, to ensure compliance with the standards, are carried out on a periodic basis. Internal audits are carried out at least annually for each Ministry, following a risk-based audit plan and focusing on specific issues revealed by the previous audit. Responsive action by management of most line ministries is fairly good across central government entities and generally there were prompt responses to internal audit findings, in compliance with the regulatory period of 4 weeks. 5) Accounting, recording and reporting The Ministry of Finance operates a conceptual Treasury Single Account system in the Bank of Botswana for the Consolidated Fund. It also operates a number of Treasury Accounts to facilitate the collection of MDA revenues as well as facilitate payments in the district offices of the MDAs. This arrangement facilitates a reconciliation of the cash book on a monthly basis, which takes place within 4 weeks of the close of each month. The GABS system provides an automated basis for assisting in the completion of the reconciliation process. The Accountant General also prepares on a monthly basis a cash flow statement that aggregates the revenues, MDA requisitions and net borrowings, reconciled against changes in bank balances within 4 weeks of the end of the month. Typically reconciliation and clearance of suspense accounts and advances are done at the close of the year, within two months after the close of the fiscal year. The Accountant General publishes monthly on the government intranet, within 21 days of the close of the month, the individual and consolidated Monthly Expenditure Returns presented in a format consistent with budget documentation and segregated by vote and classified by current expenditure, transfers and subsidies, and capital expenditure. This format permits the direct comparison of revenue and expenditure to the original budget allocations which are included in the tables. The format of the Monthly Expenditure Returns reflects expenditure only at the time of payment and does not reflect commitments. The Local Authorities are responsible for primary schools and primary health care clinics respectively. The Ministry of Local Government publishes a Recurrent Budget report that includes both estimates and actual outturns for the Education Department and Health Clinics. There is a similar report published for capital expenditure in each district segregated in the same way. While there are no summary extracts pertaining to expenditure by primary health care clinics or primary schools, such summaries can be readily extracted from the budget documentation consolidated by the Ministry of Local Government. A consolidated central government financial statement is prepared annually (8 months after close of the financial year) that includes full information on revenues, expenditure and financial assets including revenue arrears. These annual statements however do not provide a full reporting on liabilities - they do not provide any information on expenditure arrears or accounts payable. 6) External scrutiny and audit Botswana is characterised by a democratic system underpinned by Parliamentary oversight over the government’s budget formulation, its actual expenditure achievements and the quality of expenditure management. The Parliamentary debates cover fiscal policies and the details of revenue and expenditure estimates, seeing the annual budget approved in law before the start of each fiscal year. Supplementary Budget Estimates are presented by the Minister of Finance several times a year, but Parliament has indicated that it will not review supplementary budgets until the last quarter or so, as a way to limit the number of supplementary budget proposals submitted. Clear rules exist with respect to in-year budget amendments by the executive with respect to item, programme and vote amendments. A review of audited financial statements, reporting expenditure anomalies, is done through a Public Accounts Committee (PAC). There have been systematic, though not always timely, follow up on PAC recommendations across all line ministries. The office of the Auditor-General (AG) generally meets the standards of independence set by INTOSAI for supreme audit institutions. These include the legal requirements with respect to the appointment and termination of the Auditor-General, the financing of the budget, the hiring of staff, the auditor’s jurisdiction and the timing and extent of dissemination of audit reports. In practice not all central government entities are audited every year. It is estimated that 80% central government expenditure is audited annually. The audit covers revenue, expenditure as well as assets and liabilities. A variety of audits are performed, including systems audits, financial and compliance, procurement, some performance audits, payroll and recently also Information Technology audits. Public Enterprises are audited by private audit firms. The audited financial statements are submitted to Parliament within 4 months of receipt from the Ministry of Finance. The actual audit reports are submitted within 13 months after the end of the fiscal year. There is room for systemic improvement in the system to respond and follow-up on audit findings. Assessment of the impact of PFM weaknesses When viewed from the perspective of key objectives of a sound PFM system, namely aggregate fiscal discipline, strategic allocation of resources, efficient delivery of services and effective accountability, Botswana scores well in many respects. Concerns do however exist in the ability of the PFM system to allocate resources strategically and to support an acceptable level of service delivery and accountability. 1) Aggregate Fiscal Discipline There is a fair measure of fiscal discipline that is achieved through effective top-down discipline to the budgetary process coupled with effective cash and debt management, effective exclusion of unreported extra-budgetary transactions; credible if excessively conservative fiscal forecasts, fair budget release predictability, expenditure control and sound salary management (see PI-1, PI-2, PI-3; PI-7; PI-16; PI-18; and PI-20). The story is however more mixed with respect to the effectiveness of the procurement system. 2) Strategic Allocation of Resources Botswana has in place several important steps towards achieving a budgetary process that is capable of the strategic allocation of resources (PI-11 and PI-12). However, there are still a number of important steps that remain to be fully implemented, including the adoption a functional and programme based budget classification (PI-5); the stronger policy inputs at the start of the budget preparation cycle (PI-11); and the improvement of the development of fully costed inputs along with recurrent cost implications considered in the development of the budget. Although the development objectives do not rely heavily upon development partner inputs, there are missed strategic opportunities that arise to the lack of a close alignment of cooperating partner grants with the budget process and an almost total absence of timely reporting on project and programme achievements consolidated into the national consolidated financial reporting framework. The effectiveness of the central government’s success in allocating resources strategically followed by disciplined budget releases in accordance with such strategic considerations benefits from the tracking resources received by front line service delivery units such as primary schools and primary health care facilities (see PI-23). 3) Efficient Service Delivery Efficient revenue collection and cash management contribute to efficient service delivery. There are also areas that Botswana has had much success in contributing to efficient service delivery, including sound payroll management and timely financial reporting. One area that appears to adversely affect the efficiency of service delivery is procurement management (see PI-19). 4) Accountability There is scope for improvement in the timeliness of submission to and examination by Parliament of audit reports. While the audited financial statements are submitted to Parliament within 5 months of receipt from the Ministry of Finance, the audit reports are not submitted to parliament till 13 months after the close of the fiscal year (see PI-26 and PI-28). Prospects for PFM reform planning and implementation The commitment to continuing improvements in PFM in Botswana has political championship at the very highest levels through the Minister of Finance. Implementation oversight and monitoring is the responsibility of the respective departmental heads and coordination of the reform efforts is the responsibility of the Corporate Services Department within the MFDP. Botswana has neither adopted a formal PFM reform strategy, nor has a formal public sector reform strategy been issued or approved by cabinet. However, there is a Public Service Reform Coordinating Unit in the Office of the President. There have been important reforms undertaken recently including the implementation of the government accounting and budgeting system (GABS), improvements in the revenue administration, strengthening of external audit and the implementation of the new public procurement legal and regulatory framework. Against the backdrop of the 2009 PEFA PFM Performance Assessment and the findings of this PER, opportunity exist for implementation of important reform initiatives that could have positive fiscal management implications for the Government. Key focus areas for a program of PFM reform over the medium term could include – • The system to develop a reliable multi-year perspective in fiscal planning, expenditure policy and budgeting, including the development of an integrated budget and further improvements to the budget classification system • Improved monitoring of the fiscal risk emanating from sub-national government and other public sector entities • Improved capacity for project and procurement planning and execution • Presentation of financial statements in accordance with International Public Sector Accounting Standards. In this regard the Government has already requested the Bank to assist with the determination of policy issues that need resolution should accrual accounting be implemented. Several other areas of concern will also be addressed through the implementation of accrual accounting, including – o Management of all assets and liabilities (including commitments and expenditure arrears) o Reconciliations between tax assessments, collections, arrears records and receipts by the Treasury • Submission of audit reports together with audited financial statements to the Parliament, which will assist to facilitate an earlier review by the PAC • Agreeing with donors a system whereby donor funded projects will be executed through the PFM system Appendix 3 SSN Analysis based on the 2002-2003 Household Income and Expenditure Survey Table A.1: Household Benefitting from Social Safety Nets Benefit Status Sample Households Population Households Number Percent Number Percent Did not benefit from SSN(s) 4,510 74.5 267,864 68 Benefitted from SSN(s) 1,543 25.5 126,408 32 Total 6,053 100.0 394,272 100 Source in original BIDPA report (2007) was given as "Author computed from 2002-2003 HIES data." Table A.2: Household Receiving Government Transfers by Expenditure Quintile Sample Quintile Benefit Status Number Percent Q1 Y < 465.22 509 33.0 Q2 465.22< Y 821.73 422 27.3 Q3 831.73< Y 1410.34 324 21.0 Q4 1410.34< Y 2803.40 194 12.8 Q5 Y > 2803.40 94 6.1 Total 1,543 100.0 Source in original BIDPA report (2007) was given as "Author computed from 2002-2003 HIES data." Table A.3: Share of Beneficiary Expenditure from Government Sources Quintile Main monthly total household Mean monthly household expenditure Share of government expenditure from all sources from government transfers transfers Q1 613.53 104.02 18.9 Q2 1,129.90 166.88 16.4 Q3 1,758.51 146.60 10.3 Q4 3,200.22 100.14 4.7 Q5 11,197.68 196.26 2.8 Total 2,841.07 208.78 17.6 Source in original BIDPA report (2007) was given as "Author computed from 2002-2003 HIES data." Table A.4: Poverty Headcount Index by Age Group Number of People Age group Non-poor Poor Total Headcount index 0-5 139,970 95,277 235,247 40.50 6-11 170,354 92,947 263,301 35.30 12-20 238,813 104,407 343,220 30.42 21-39 347,215 104,305 451,520 23.10 40-64 180,728 707,66 251,494 28.14 65+ 55,710 32,432 88,142 36.80 Total Population 1,132,790 500,134 1,632,924 30.63 Source in original BIDPA report (2007) was given as "Author computed from 2002-2003 HIES data." Table A.5: Targeting and Coverage Effectiveness Ratios Effectiveness ratio SSN Program Implementation Targeting Leakage ratio Coverage ratio ratio (IR) expenditure (LR) (CR) ratio (TER) All government transfers 0.4300 0.43 0.57 0.1900 War veteran pension 0.0084 0.25 0.75 0.0021 Old age pension 0.2700 0.48 0.52 0.0129 Assistance from drought relief 0.0043 1.00 0.00 0.0430 Destitute package 0.0170 0.64 0.36 0.0110 Needy student package 0.0029 0.50 0.50 0.0014 Orphan ration 0.0024 0.43 0.57 0.0010 RADS package 0.0018 1.00 0.00 0.0018 CHBC allowance 0.0012 0.00 1.00 0.0000 Clinic ration 0.0027 0.79 0.21 0.0021 School meal 2.7700 0.25 0.75 0.6400 Source in original BIDPA report (2007) was given as "Author computed from 2002-2003 HIES data." Table A.6: Program Targeting (PDL Analysis) SSN Program Population Households TER Non-poor Poor Total % poor (percent) War veteran pension 1,396 584 1,980 29.49 25 Old age pension 36,168 21,494 63,662 43.19 48 Assistance from drought relief 3,761 6,029 9,790 61.58 100 Destitute package 6,438 12,090 18,528 65.25 64 Needy student package 4,132 5,688 9,820 57.92 50 Orphan ration 9,887 6,372 16,259 39.19 43 RADS package 1,218 1,795 2,943 60.99 100 CHBC allowance 1,236 804 2,040 39.41 0 Clinic ration 17,820 14,736 32,556 45.26 79 School meals 144,722 51,993 196,715 26.43 25 Total 82,056 75,592 157,578 47.9 43 Source in original BIDPA report (2007) was given as "Author computed from 2002-2003 HIES data." Table A.7: Poverty Rates with and Without SSNs Region Household Level Individual Level With GT* Without Change With GT Without Change GT GT Gaborone 6.3 8.2 1.9 6.5 9.4 2.9 Francistown 11.5 15.9 4.4 14.1 20.1 6.0 Other Cities and Towns 11.3 14.2 2.9 14.3 19.0 4.7 Rural South-East 21.7 30.6 8.9 29.4 40.8 11.4 Rural North-East 29.0 38.0 9.0 38.5 49.8 11.3 Rural North-West 28.4 36.6 8.2 42.3 55.2 12.9 Rural South-West 35.0 43.8 8.8 49.8 61.6 11.8 Grand Total 21.7 28.8 7.1 30.6 40.5 9.9 Source in original BIDPA report (2007) was given as "Author computed from 2002-2003 HIES data." Note: GT – government transfers