Poverty Reduction 43849 288 March 2008 Findings reports on ongoing operational, economic, and sector work carried out by the World Bank and its member governments in the Africa Region. It is published periodically by the Operations Results and Learning Unit on behalf of the Region. The views expressed in Findings are those of the author/s and should not be attributed to the World Bank Group. Uganda's Virtual Poverty Fund Pro-Poor Spending Reform C an governments and Introduction donors ensure that re- sources are effectively tar- The provision of debt relief to geted to poverty reducing HIPCs commencing in the programs? late 1990s, and the growing interestamong donors in The question raises difficult providing direct budget sup- issues both regarding the port, increased donor focus anticipated effect of a spe- on national budget systems. cific spending policy on pov- Given that debt relief and aid erty and the quality of budget resources are fungible, do- management. The Heavily nors were concerned that Indebted Poor Countries such debt relief be verifiably (HIPCs) program attempted to used to benefit the poor in earmark resources released the recipient country. In by debt relief to programs effect, the World Bank and that would benefit the poor. the International Monetary Various developing countries Fund (IMF), acting on behalf with weak public expenditure of donors, asked that HIPC management systems--such governments put in place as Tanzania, Ghana, Chad, systems to track the use of Honduras, and Zambia--have resources freed up by debt established special poverty relief and show that these funds, along the lines or were in fact used to finance similar to Uganda's Poverty pro-poor programs. This Action Fund. required governments to There are, however, key have the capacity to identify lessons from Uganda's expe- policies and programs that rience, both regarding the would benefit the poor and to design of Virtual Poverty effectively channel and track Findings Funds (VPFs) and the defini- resources to such programs. tion of pro-poor programs, Budget systems are notori- that deserve careful consid- ously weak in most low in- eration in similar initiatives come countries. Just how elsewhere. The issue as- weak was made evident by sumes particular significance the 2002 HIPC assessment of in light of the recent decision budget systems conducted of the G-8 to expand the scope jointly by the IMF and the of debt relief significantly. World Bank, which concluded that 15 out of 24 HIPC coun- dedicated poverty fund with · Protecting budget disburse- tries needed substantial special implementation and ments to these programs upgrading of public expendi- reporting arrangements. ture management (PEM) · Monitoring of performance systems. Typically such coun- But such a short-term of these expenditures. tries demonstrate: remedy would create a paral- lel financial arrangement · An inability to identify that would divert scarce This note considers the Poverty ReductionStrategy capacity and undermine the Uganda VPF to understand Paper (PRSP) priorities integrity of the overall budget how well it served to allocate within the existing budget management system. The resources to pro-poor pro- classification system alternative was to create a grams and what weaknesses · Budget allocations and out- VPF as a bridging mechanism were observed that may need turns that do not reflect for tracking pro-poor expendi- to be corrected as other PRSP priority programs tures in the budget, whilst countries employ mecha- budgetwide mechanisms nisms similar to the VPF. · Unpredictability in budget were being established and allocation/ implementation strengthened. A well de- The Poverty Action Fund-- processes and the inability signed VPF would, in prin- Uganda's VPF to track expenditures dur- ciple, allow for: Uganda was the first country ing budget implementation. · Maintaining the integrity of to benefit from debt relief There was, therefore, an budget management and under the original HIPC and inherent conflict between the systemic reforms enhanced HIPC initiatives. In capabilities of recipient bud- 1997, prior to receiving HIPC, get systems and the expecta- · Adapting the existing budget the Government of Uganda tion of HIPC donors. Given classification system to "tag" (GOU) developed its own such problems, the tempta- pro-poor programs (hence comprehensive strategy to tion to address the immedi- "virtual" poverty fund) tackle poverty, the Poverty ate pressure from donors to · Linking specific (e.g., HIPC) Eradication Action Plan account for the use of HIPC resources to these budget (PEAP). Subsequently, the resources was to introduce a allocations government introduced the Poverty Action Fund (PAF) in 1998 to reorient government Box 1: Tracking Pro-Poor Spending expenditures towards imple- menting its PEAP as well as The performance of a country's public financial management (PFM) to account for HIPC resource system in terms of its ability to allocate and execute budgets and to track use (see box 2). and report on poverty reducing spending is monitored annually as part of oversight of the HIPC Initiative. One of the indicators relates to the exist- Successes of the PAF ence of effective pro-poor tracking mechanisms for the HIPC funds chan- Over time the scope of the neled through the budget. The definition of what qualifies as a poverty reducing expenditure is determined by country authorities and varies from PAF budget increased. Ex- country to country, but generally includes social sector spending, some- plicit criteria for programs to times water, and in a few cases other infrastructure needs. qualify for inclusion were Of the 27 HIPC countries that were monitored in 2004, only 14 developed. The focus of at- (Cameroon, Democratic Republic of Congo, Ethiopia, The Gambia, Ghana, tention moved to the actual Guyana, Honduras, Malawi, Mali, Niger, Rwanda, Sierra Leone, Tanzania, performance of PAF pro- and Uganda) had satisfactory tracking mechanisms for pro-poor spend- grams. ing. Of these, Honduras, Sierra Leone, and Uganda have established a virtual fund and protected them against budget cuts. Rwanda and Tanzania Although the predictability have resorted to mechanisms similar to a virtual fund. The remaining of disbursements facilitated countries used existing budget classifications to identify pro-poor expendi- better performance, the tures, but have no mechanisms to protect these expenditures from budget guarantee was qualified, that cuts. is, only those programs ac- Table 1: Reorienting national and Improved budget predict- sector allocations towards the PEAP ability, transparency, and 1997/98 1988/89 2002/03 accountability: The govern- (pre-PAF) ment guaranteed that all PAF programs as % of national budget (excluding interest) budgeted resources would be Social services1 17 21 27 made available in full for Productive sectors2 1 2 4 disbursement to PAF pro- Others 1 2 5 grams, regardless of re- Total 19 25 36 PAF expenditures 163 267 692 source shortfalls. The PAF (Uganda shillings 002/03 provided a platform for estab- prices lishment of an open and 1Primary education, primary healthcare and water. transparent process of bud- 2Agriculture, rural roads, and strategic exports. get reporting and review, and improved the focus on the counting for funds and per- with a level of comfort in results of government's forming satisfactorily were terms of allocation, imple- programs. A system of activ- guaranteed funding. Unless mentation, and transparency, ity-based budget reporting the performance of programs the PAF enabled the donors' was introduced in local was put under scrutiny, it shift from project to budget government for PAF condi- was felt that there would be and sectorwide approach tional grants. Five percent of little incentive to perform. (SWAP) support. all PAF resources were set Reorienting budget alloca- More recently as governance side for oversight institutions tions towards pro-poor ser- concerns have emerged in and local government to vice delivery and demon- Uganda, the PAF has proved improve monitoring and strating the additionality of important in justifying contin- accountability. debt relief: The PAF ensured ued provision of budget support While a number of the that additional HIPC debt by demonstrating orientation achievements can be directly relief and donor direct budget of the budget to pro-poor ex- attributed to the PAF, it support were channeled into penditures. must be emphasized that specific PEAP priority pro- grams, helped reorient allo- cations within sectors to- Figure 1: Earmarking HIPC and budget support to the PAF wards pro-poor expenditures, may actually distort budget allocations and increased the funds channeled to local govern- ments. Expenditures on PAF programs grew from 19 per- cent to 36 percent of a rap- idly expanding government budget between 1997/8 and 2002/3 (see table 1). Mobilizing donor resources and harmonizing conditions: Over time, the PAF has also contributed to the mobiliza- tion of sector-specific donor resources through budget support, which increased from $20 million in 1998/9 to more than $130 million in 2001/2. By providing donors ity, and agriculture that Box 2: The Key Elements of the Uganda Poverty Action Fund may have had a larger impact on income pov- The PAF identifies and gives special treatment to specific pro-poor sectors/ erty. Conceptually, a subsectors/programs in the budget. budget allocation such · PAF criteria--PAF programs are defined as only those that are in the as B rather than C may PEAP or PRSP, are directly poverty reducing, delivering a service to the have had a greater poor, and have a well developed plan. The five major areas are primary positive impact on the education, primary healthcare, water and sanitation, rural roads, and agriculture extension. poor. This has been · Matching resources to expenditures--A PAF table matches specific compounded by the resources from HIPC, donors, and the government to the budget difficulty government allocations for PAF programs. encountered in develop- ·Additionality of resources--PAF resources were shown as additional to the ing appropriate public government's own budget allocations to PAF programs in the 1997/8 sector programs that budget. Since 2000, the GOU has made a commitment that PAF will promote the private consistently grow as a proportion of the overall budget. sector, exports, and · Protection of disbursements--Government guarantees that PAF programs more generally economic are protected from budget cuts during implementation, provided that growth. What is required performance is on track. · Reporting and transparency--There are specific requirements for local is a more balanced governments (LGs) and other government departments to report approach so as to provide disbursements on PAF programs, and progress in implementation. greater flexibility within Reports are made public and discussed in open quarterly meetings, where the budget to allocate civil society, the press, and donors are present. resources to pro-poor · Monitoring--Five percent of PAF funds are earmarked for enhanced issues. monitoring and accountability. The Tanzanian model too far towards the direct has some useful fea- broader reform initiatives provision of basic social tures that provide such as the medium term services to the poor, illus- greater budget flexibility expenditure framework trated by the movement from in future, (see box 3), (MTEF), SWAPs, and the point A to point C in figure 1. although early ap- PRSP combined with a sup- This reflects the difficulty of proaches to poverty portive political, institutional, predicting the ex-ante impact reducing spending had and policy environment have of a chosen budget allocation. a similar social sector played a major part. But Government's commitment to bias (see figure 2). equally, by maintaining the the size of the PAF and donor Biased budget imple- integrity of the budget whilst preference for the social mentation: The bias in channeling HIPC resources, sectors has limited the abil- budget allocation has the VPF contributed to the ity to reallocate away from been magnified by a success of the PRSP and established PAF sectors. similar bias in budget budget reforms. Thus more than 70 percent of execution. With the Negative aspects of the PAF additional PAF resources growth of the PAF, other were spent on basic social sectors have borne the Some aspects of the PAF are services between 1998/9 and brunt of budget adjust- problematic and could poten- 2002/3, despite the budget ments to resource short- tially be undermining the already being oriented to- falls. The protection of achievement of Uganda's wards those services. disbursements under poverty reduction goals: It can be argued, retrospec- PAF is required only Unbalanced budget alloca- tively, that there has been because the major tions: The PAF may have underinvestment in areas causes of skewed budget allocations such as roads, rural electric- underdisbursement, such as the serial over- Figure 2: Social service bias in Uganda and tanzania priority poverty reducing expenditures (total public expenditure excluding interest) spending of some government A VPF should be introduced institutions, have not been temporary mechanism with a in a way that supports rather addressed to date. sunset clause at the outset. than replaces the implemen- tation of such comprehensive Partial monitoring and Key lessons improvements in budget evaluation: The PAF added a Given weak initial PEM capa- preparation and implementa- layer of monitoring and bilities, VPFs allow priority tion. Donor dialogue and evaluation (M&E) and exter- poverty reduction programs to conditions should be based nal verification processes be implemented without on achievement of such that has diverted attention undermining reform of public improvements, and not away from the overall budget expenditure management solely the meeting of VPF and led to unbalanced scru- systems (PEMs). To be effec- commitments. tiny of government expendi- tive, a VPF should be simple tures. Although initially an and limited to the identifica- Protection of VPF budgets improvement, currently tion of PRSP priority expendi- may be necessary, but it these systems have not been tures in the budget classifi- would be important to control mainstreamed into govern- cation system. overspending in other parts ment systems, and the cover- age of the budget M&E im- provements remains partial. Box 3: Tanzania--different approaches to pro-poor spending No exit strategy: While the Tanzania originally adopted a more flexible approach for allocating Ministry of Finance has expenditures to priority PRSP areas by including all spending on expressed a desire to phase the seven broad priority sectors. A system of reporting on pro-poor priority expenditures was also introduced. The broader definition out PAF, government agen- allowed greater flexibility in budget formulation and execution, cies within priority sectors whilst neither creating an artificial enclave nor undermining the and donors supporting those flexibility necessary to complement priority expenditures with other sectors want the preferential activities. However, the disadvantage was that it treated all lines of PAF treatment continued, expenditures within a priority sector as equally important. Since the making it politically difficult 2004/05 budget, the priority sectors have been replaced by an to remove PAF protection, allocation system that aims to make a more comprehensive link especially given that it was between the second PRSP and budget allocations by requiring all not clearly designed as a spending agencies to identify how their expenditures are contributing to PRSP objectives. of the budget to limit the This represents the greater shocks to unprotected sec- challenge in the context of tors. countries with weak policy, A VPF does not bypass the planning, and budgeting This Findings was originally need to have a PRSP and an processes. The danger of a published as World Bank effective budget process that VPF is that it can create PREM Note Number 108, identify priority pro-poor incentives for development March 2006, and was written expenditures to be included partners to predominantly by Sudharshan Canagarajah in the VPF as part of a fund social sectors that can (scanagarajah@worldbank.org) broader policy framework for distort the implementation of of the World Bank and Tim growth and poverty reduc- more balanced and appropri- Williamson tion. ate strategies for improving (t.williamson@odi.org.uk) of the poverty outcomes. Overseas Development Institute, London.