Report No. 53704-UG Uganda Agriculture Public Expenditure Review February 28, 2010 Agriculture Rural Development Unit (AFTAR) Sustainable Development Department Country Department 1: Tanzania, Uganda and Burundi Africa Region Document of the World Bank Acronyms and Abbreviations AAMP Area Agricultural Modernization Program ADB African Development Bank AgPER Agriculture Public Expenditure Review ARD Agriculture and Rural Development A­SWG Agriculture Sector Working Group BCC Budget Call Circular BFP Budget Framework Paper CAADP Comprehensive Africa Agriculture Development Programme COFOG UN Classification of Functions of Government COMESA Common Market for Eastern and Southern Africa DDA Dairy Development Authority DfID Department for International Development (UK) DRA Direct Rate of Assistance DSIP Development Strategy and Investment Plan ECSSD Environmental and Socially Sustainable Development EPRC Economic Policy Research Centre (Kampala) FAO Food and Agriculture Organization of the United Nations FITCA Farming in Tsetse Areas of East Africa Project GDP Gross domestic product GoU Government of Uganda IFPRI International Food Policy Research Institute ISFG Integrated Support to Farmer Groups JICA Japan International Cooperation Agency LDCP Livestock Disease Control Project LGDP Local Government Development Program M&E Monitoring and evaluation MAAIF Ministry of Agriculture, Animal Industry, and Fishery MoFPED Ministry of Finance, Planning, and Economic Development MTEF Medium-Term Expenditure Framework MWLE Ministry of Water, Lands, and Environment NAADS National Agricultural Advisory Service NARO National Agricultural Research Organization NBFP National Budget Framework Paper NEPAD New Partnership for Africa's Development NGO Non-Governmental Organization NFA National Forest Authority NLPIP National Livestock Productivity Improvement Project NSCG Non-Sectoral Conditional Grant NWSADP North-West Smallholder Agricultural Development Project OPM Oxford Policy Management PAF Poverty Action Fund PEAP Poverty Eradication Action Plan PER Public Expenditure Review PMA Plan for Modernization of Agriculture ReSAKSS Regional Strategic Analysis and Knowledge Support System PETS Public Expenditure Tracking Survey PMU Project Management Unit SACCO Savings and Credit Cooperative SBFP Sectoral Budget Framework Paper SFDP Support to Fisheries Development Project SIDA Swedish International Development Agency TRA Total Rate of Assistance TTL Task Team Leader UCDA Uganda Coffee Development Authority UCDO Uganda Cotton Development Organization UK United Kingdom USAID United States Agency for International Development UShs Ugandan shillings VAT Value-added tax VODP Vegetable Oil Development Project iv Contents Executive Summary ............................................................................................ x 1. Introduction .................................................................................................. 1 2. Level and Composition of Agricultural Sector Expenditure in Uganda .. 4 2.1. Trends in agricultural sector expenditure ............................................................. 4 2.2. Agricultural policy environment ........................................................................ 12 2.3. Economic composition of the sector budget ...................................................... 14 3. Budget Process and Performance ........................................................... 29 3.1. Budget preparation ............................................................................................. 29 3.2. Budget execution ................................................................................................ 35 3.3. Budget performance monitoring ........................................................................ 40 4. Efficiency and Effectiveness of Agricultural Sector Expenditures ....... 42 4.1. Methodology and analytical approach ............................................................... 43 4.2. Sample selection and procedures for field work ................................................ 45 4.3. Implementation arrangements of MAAIF development projects ...................... 47 4.4. Analysis of technical efficiency ......................................................................... 48 4.5. The impact of NAADS ....................................................................................... 57 4.6. Incidence of agricultural public expenditure ...................................................... 60 5. Conclusions and Policy Recommendations ........................................... 67 5.1. Agricultural sector budget: recent developments ............................................... 67 5.2. Agricultural price distortions ............................................................................. 68 5.3. Allocative efficiency of public expenditure ....................................................... 68 5.4. Technical efficiency of public expenditure ........................................................ 70 5.5. Budget processes ................................................................................................ 73 Annex 1: Recommendations to Improve the Guidelines of the Sector Budget Framework Paper ................................................................................ 75 Annex 2: Distribution of Goats for Restocking as at 30th September 2007 in NLPIP ................................................................................................................. 78 Tables Table 1: Shares of the agricultural sector budget in the national budget and GDP, 2001/02­ 2008/09 (percent) ......................................................................................................... 7 Table 2: Medium-term expenditure framework for the agricultural sector budget, 2008/09­ 2012/13 (UShs billions) ............................................................................................... 7 Table 3: Expenditures on agriculture based on the COFOG classification, 2001/02­2005/06 (UShs billions) ............................................................................................................. 9 Table 4: Off-budget disbursements to agriculture (based on the COFOG classification) by two donors, 2000/01­2006/07 (UShs billions) ................................................................. 10 Table 5: International comparison of budgets for agriculture, average for 2002­04............... 12 Table 6: Distortion indicators for Ugandan agriculture, five-year averages............................ 13 Table 7: MAAIF recurrent budget allocation, 2001/02­2008/09 (UShs billions) ................... 16 Table 8: Allocation to grants in the MAAIF approved recurrent budget, 2001/02­2008/09 (UShs billions) ........................................................................................................... 16 Table 9: Percentage distribution of MAAIF's recurrent budget by department (excluding grants and domestic arrears), 2005/06­2008/09 ........................................................ 17 Table 10: Main parts of MAAIF's recurrent budget (excluding grants and domestic arrears), 2005/06­2008/09 (UShs billions) .............................................................................. 18 Table 11: Sources of funding for the development budget, 2005/06­2008/09 (UShs billions) ........................................................................................................... 19 Table 12: The economic composition of MAAIF's development budget, 2005/06­2008/09 (UShs billions) ........................................................................................................... 19 Table 13: The detailed economic composition of MAAIF's development budget, 2005/06­ 2008/09 (UShs billions) ............................................................................................. 20 Table 14: Economic composition of MAAIF's budget (excluding grants and domestic arrears), 2005/06­2008/09 (UShs billions) .............................................................................. 22 Table 15: Proportion of MAAIF budget allocated to DSIP priority areas, compared with DSIP projections, 2005/06 to 2007/08 (percent) ................................................................. 27 Table 16: Evolution of 2007/08 budget ceiling projections for agriculture, March 2005 to March 2007 (UShs billions) ...................................................................................... 30 Table 17: Evolution of the 2006/07 budget ceiling projections for agriculture by Vote (UShs billions) ........................................................................................................... 31 Table 18: Approved, revised, and released recurrent expenditures for MAAIF, 2000/01­ 2005/06 (UShs billions) ............................................................................................. 36 Table 19: Released recurrent expenditures as a percentage of approved recurrent budget allocations for MAAIF programs, 2000/01­2005/06 ................................................ 36 Table 20: Released recurrent expenditure as a percentage of revised recurrent budget allocations for MAAIF programs, 2000/01­2005/06 ................................................ 37 vi Table 21: Disbursement/release of MAAIF of donor- and government-financed development funds as a percentage deviation from development budget allocations, 2001/02­ 2005/06 ...................................................................................................................... 37 Table 22: Approved and released development budgets in the selected MAAIF development projects, 2005/06­2007/08 ........................................................................................ 38 Table 23: Shares of selected projects in MAAIF's development budget, 2005/06­ 2007/08 ...................................................................................................................... 46 Table 24: Contract performance within NWSADP ................................................................... 51 Table 25: Percentage of capital expenditures (approved budgets) of NWSADP, NLPIP, and SFDP in the budget of these projects, MAAIF's capital budget, and MAAIF's development budget, 2005/06­2008/09..................................................................... 52 Table 26: Financial resources transferred by the NWSADP Project Management Unit and received by districts (UShs millions) ......................................................................... 54 Table 27: Financial resources transferred by NLPIP Project Management Unit and received by districts (UShs millions) ............................................................................................ 54 Table 28: Physical resources reportedly transferred by the LDCP Project Management Unit and received by districts (000 doses of vaccine) .............................................................. 55 Table 29: Physical resources reportedly transferred by NLPIP Project Management Unit and received by districts (numbers of animals) ................................................................ 55 Table 30: Flow of financial resources from the NAADS Secretariat to districts, 2005/06­ 2006/07 (UShs millions) ............................................................................................ 58 Table 31: Incidence of public and private services across wealth (per capita) quintiles, per capita in 2007............................................................................................................. 66 vii Figures Figure 1: Trends in nominal and real approved budgets for agriculture, 2001/02­2008/09 (UShs billions) ......................................................................................................... 4 Figure 2: Released versus approved budgets for the agricultural sector, 2001/02­2005/06 (UShs billions) ......................................................................................................... 5 Figure 3: Structure of the approved budget for agriculture, 2001/02­2008/09 (UShs billions) ......................................................................................................... 6 Figure 4: Agriculture's share of the national budget in relation to other sectors, 2007/08­ 2012/13 (UShs bilions) ............................................................................................ 8 Figure 5: Agricultural sector Total Rate of Assistance, Uganda, 1961­2004....................... 14 Figure 6: MAAIF recurrent and development budgets, 2001/02­2008/09 ........................... 15 Figure 7: Economic composition of MAAIF's budget (excluding grants and domestic arrears), 2005/06­2008/09 (UShs billions) ........................................................... 22 Figure 8: Economic composition of the agricultural sector budget, 2005/06­2008/09 (UShs billions) ....................................................................................................... 23 Figure 9: Functional composition of the agricultural sector budget, 2006/07 ...................... 25 Figure 10: NAADS (Districts) budgeted and actual cash flows, 2005/06 (UShs billions) ....................................................................................................... 39 Figure 11: Relationship between inputs, outputs, and outcomes in agriculture ...................... 43 Figure 12: Implementation arrangement of most MAAIF agricultural development projects .................................................................................................................. 48 Figure 13: Incidence of the public agricultural expenditures by household wealth quintile... 62 Figure 14: Incidence of total support by wealth quintile ........................................................ 62 Figure 15: Proportion of households receiving support by wealth quintile ............................ 62 Figure 16: Incidence of support for NAADS and non-NAADS participants.......................... 63 Figure 17: Amount of support received by NAADS and non-NAADS participants by wealth quintile ................................................................................................................... 63 Figure 18: Incidence of credits and public expenditure by wealth quintile ............................ 63 Figure 19: Incidence of Agricultural Advisory Services by wealth quintile ........................... 64 Figure 20: Incidence of Advisory Services by NAADS membership and wealth quintile ..... 64 Figure 21: Access to information on crop production methods and practices by wealth quintile ................................................................................................................... 65 viii Acknowledgments This report synthesizes findings from a joint effort to comprehensively review the agricultural public expenditures in Uganda by the Ministry of Agriculture, Animal Industries and Fisheries (MAAIF), the World Bank, and the UK Department for International Development (DfID). The review was conducted under the auspices of the Agriculture Sector Working Group (ASWG) and with the full support of the Ministry of Finance, Planning, and Economic Development (MoFPED) and the other Agriculture Development Partners of Uganda. The team gratefully acknowledges the financial support received from the World Bank­DfID program on Agriculture Expenditure Reviews (AgPER), DfID Uganda, the National Agricultural Advisory Services (NAADS) Secretariat, and the Regional Strategic Analysis and Knowledge Support System (ReSAKSS) node for Eastern and Central Africa to undertake the background studies for this review. The team particularly appreciated the strong support and encouragement of Limin Wang (DfID), Christopher Delgado (ARD), and the World Bank­DfID AgPER program. The core study team included Keizire Boaz Blackie (formerly CAADP and AgPER focal point in MAAIF, now with the African Union Commission); Fred Mayanja (MAAIF); Madhur Gautam, Sergiy Zorya, Wilson Onyang Odwongo, Varun Kshirsagar (World Bank); Alan Tollervey (formerly with DfID Uganda, now DfID London); Peter Oumo (DfID Uganda); and Godfrey Bahiigwa (formerly with RESAKSS and now PMA, MAAIF). This synthesis report was prepared by Madhur Gautam and Sergiy Zorya (Task Team Leader and co-TTL for the World Bank). It is based on background studies conducted in three phases by teams from the World Bank, Oxford Policy Management (OPM), London; Economic Policy Research Centre (EPRC), Kampala; and the International Food Policy Research Institute (IFPRI). The Phase I study on levels and patterns of allocation of agricultural expenditures was prepared by the OPM team, led by Lawrence Smith and including Martin Fowler, Frederick Mugerwa, and Milton Ogeda, with support from Stephen Akroyd. The Phase II study on budget process and performance was conducted by the same OPM team. The Phase III analysis of the efficiency and effectiveness of agricultural public expenditures consisted of three studies: (i) a public expenditure tracking study undertaken by the EPRC team, led by Nyende Magidu and supported by Mugisha Fredrick and Omiat Omongin, with technical support from Sergiy Zorya; (ii) an impact evaluation of NAADS, conducted by an IFPRI team led by Sam Benin and Ephraim Nkonya and including Geresom Okecho, Josée Rdriamamonjy, Edward Kato, Geofrey Lubade, Miriam Kyotalimye, and Francis Byekwaso; and (iii) an incidence analysis of agricultural expenditures undertaken by Varun Kshirsagar and Madhur Gautam. Dina Umali-Deininger (Sector Manager, ECSSD), Mona Sur (Senior Economist, ARD) and Christopher Delgado (Advisor, ARD) served as Peer Reviewers for the study. Limin Wang, Christopher Delgado, Wilson Onyang Odwongo, and Stephen Mink provided advice and comments throughout the preparation of the report. Kelly Cassaday edited this report. ix Executive Summary This Agriculture Public Expenditure Review (AgPER) analyzes the efficiency and effectiveness of agricultural sector expenditure in Uganda. Demand for this AgPER comes from several quarters. The Ministry of Agriculture, Animal Industry, and Fisheries (MAAIF) has a strong interest in using the results of a comprehensive AgPER to better align future expenditure to its priorities, improve the quality of public service delivery, and make the case for an appropriate level of funding for the sector. This Review is also a part of a larger annual Public Expenditure Review, a core diagnostic undertaken collaboratively by the Government of Uganda (GoU) and World Bank. An additional justification for this AgPER has its origins in the commitment by member states of the New Partnership for Africa's Development (NEPAD) to allocate 10 percent of their annual budgets to agriculture. Under NEPAD's Comprehensive Africa Agriculture Development Programme (CAADP), one of the first tasks for individual countries is to undertake a public expenditure review to document the level, composition, and quality of agricultural sector expenditure. Uganda is one of the first five pilot countries to carry out the CAADP and has committed itself to conducting an AgPER. The UK Department for International Development (DfID) and the World Bank have established a global program to support PERs for the agricultural sector. Uganda was selected as one of the case study countries for the program in response to the demand expressed by the government--the Ministry of Finance, Planning, and Economic Development (MoFPED) and MAAIF. The demand for this analysis also comes from the local Agriculture Development Partners (in particular the DFID and World Bank country offices), and the Regional Strategic Analysis and Knowledge Support System (ReSAKSS) node for Eastern and Central Africa, which is mandated by the Common Market for Eastern and Southern Africa (COMESA) to monitor agricultural sector spending by governments in the region. Against this background, the objectives of this AgPER are to help the Government of Uganda better align future expenditures to its priorities, assess the robustness of the agricultural public expenditure management system to maximize the effectiveness of scarce resources, and make the case for the need and absorptive capacity for additional expenditure targeted at high-priority areas. The agricultural sector budget in this AgPER includes the following agencies: (i) MAAIF, (ii) National Agricultural Research Organization (NARO), (iii) National Agricultural Advisory Services (NAADS) at the central and district levels, (iv) Uganda Cotton Development Organization (UCDO), (v) Uganda Coffee Development Authority (UCDA), (vi) district agricultural extension staff, and (vii) Non-Sectoral Conditional Grants. x Trends in agricultural sector expenditure in Uganda Between 2001/02 and 2008/09, agricultural sector expenditure can be divided into two distinct phases. During the first phase, from 2001/02 to 2003/04, the budget for agriculture fell sharply in nominal and real terms (Figure E1). It began to recover in 2004/05, and by 2008/09, cumulative growth in the sector budget was 46 percent above 2001/02 level in nominal terms. In real terms, although the budget was 38 percent higher in 2008/09 than in 2004/05, it was at about the same level as in 2001/02. As a share of the national budget, the budget for agriculture had fallen to 3.8 percent in 2008/09 from 5.7 percent in 2001/02. Finally, in terms of gross domestic product (GDP), the sector budget remained stable, albeit low, at 1.6 percent. It is important to emphasize that these are the trends in the approved budget. The released budget was on average 10 percent lower, reducing the share of sector expenditure in GDP to a very low 1.2 percent. The agricultural sector budget is not much larger when measured according to the United Nations classification of functions of government (COFOG). This classification, recommended by CAADP/NEPAD for comparing agricultural sector expenditures across countries in Africa, increases the sector budget by about 1 percent of the national budget (to 5.4 percent for the released budget in 2005/06, for example).1 There is substantial off-budget spending by some donors, however, which is estimated to be equivalent to 10­20 percent of the current sector budget. Information about off-budget spending remains fragmented and difficult to obtain. Because it lies largely outside the purview of planners and policy makers, off-budget spending makes coordination, prioritization, and sector planning difficult. Even when off-budget funds are taken into account, Uganda's agricultural sector budget is still about two times lower than it must be to meet Uganda's Maputo Declaration pledge to allocate 10 percent of the national budget to agriculture under CAADP. In addition to these public expenditures, it is important to note the significant volume of expenditures by nongovernmental organizations (NGOs). Estimated at about 10 percent of the present budget, these expenditures pose additional coordination and harmonization challenges. 1 The COFOG classification increases the budget for agriculture with the addition of the land and forestry sectors under the Ministry for Water, Lands, and Environment (MWLE) and National Forest Authority (NFA). xi Is Uganda's agricultural sector expenditure expected to grow in the near future? Rather than growing, public expenditures on agriculture are projected to keep declining, according to the Medium-Term Expenditure Framework. In 2012/13, the agricultural sector expenditure is expected to be 3.2 percent of the national budget, compared with 3.8 percent in 2008/09 and 4.6 percent in 2001/02 (Figure E2). Budget resources will be shifted to other areas, notably infrastructure and social development, which would provide for indirect positive benefits to agriculture. Within the agricultural sector budget, resources will continue to shift from the national to local level, mainly to NAADS in the districts. Given this outlook, it is more critical than ever to ensure that scarce budgetary resources are used as effectively and efficiently as possible at all levels in the agricultural sector. It is also vital for MAAIF to improve collaboration with local governments, which increasingly will be responsible for implementing agricultural programs. How does Uganda's spending on agriculture compare to spending by other countries? An international comparison shows that Uganda spends relatively less than other countries (Table E1), when spending is measured as the share of agricultural budget in GDP (adjusted by the size of the sector). Uganda spends about as much as other Sub- Saharan African countries but less than middle- and high-income countries. This finding is consistent with Uganda's narrower fiscal capacities and its greater need to support competing public investments in infrastructure and social sectors compared to higher- income countries. International experience has shown, however, that lower public spending does not necessarily mean lower agricultural competitiveness. It is well known that in the United States and the European Union, for example, expensive public support in the form of huge farm subsidies has failed to keep farmers competitive in world markets, whereas lower spending in Brazil, Australia, and New Zealand has not prevented their farmers from being highly competitive. xii Three conditions must be met to maximize the impact of public spending on agriculture in Uganda. These conditions will be even more critical if spending begins to rise again. First, public expenditures on agriculture must be supported through an enabling environment in which agricultural prices are subject to few distortions. It is counterproductive to raise the public expenditure on agriculture when farm-gate prices are depressed. Second, the mix of spending should be efficient (allocative efficiency). Spending that does not contribute (or does not contribute as much) to growth and poverty reduction, relative to alternative goals, is allocatively inefficient or unproductive. Third, technical efficiency should be high. Technical efficiency in the public sector involves making the best use of inputs to provide outputs in the form of public services. (Put simply, technical efficiency is doing things well, and allocative efficiency is doing the right things.) When those three conditions are met, even small budgets will be well- positioned to generate economic growth and encourage more private investment. It is also very likely that when these conditions are met, the scaling up of agricultural sector expenditure in Uganda will bring the highest rates of return. Agricultural price distortions Uganda has successfully addressed one requisite for the efficiency of public expenditure: Agricultural price distortions have been largely eliminated, and most farm-gate prices are at the level of reference border prices adjusted for marketing costs. In 2001­04, the rate of assistance to agriculture was estimated at 1 percent, with no taxation to exportables and about 13 percent support to importables such as rice. The nonagricultural rate of assistance has also decreased notably, reducing the prices of farm inputs and stimulating resource flows to agriculture. The policy environment in Uganda can be described as quite conducive for public expenditure to have a lasting impact. Allocative efficiency of public expenditure The allocative efficiency of public expenditure can and must be improved. The first drawback is the small share of capital expenditure in the sector budget. An economic decomposition of Figure E3: Shares of development budget and capital outlays in the public spending illustrates approved MAAIF budget, 2005/06-2008/09 that in Uganda "development 100% expenditure" is not 90% synonymous with "capital 80% Share of expenditure," as usually 70% development expenditure assumed. Development 60% expenditure makes up about 50% 80 percent of the agriculture 40% budget but is heavily 30% Share of capital expenditure oriented to nonwage 20% recurrent expenditures rather 10% than to capital expenditures. 0% 2005/06 2006/07 2007/08 2008/09 The share of capital outlays in the 2008/09 approved budget is 6 percent, falling from an already low level of 11 percent in 2005/06. The share of capital expenditures is higher in MAAIF's budget than in the total sector budget; it xiii was 22 percent in 2008/09 (Figure E3). This budget share is far less than that of the development budget and may convey the erroneous impression that capital expenditures dominate public spending. In reality, too little is invested in rural community roads, bridges, and wholesale and livestock markets. Only a minor share of the budget is allocated to irrigation and to veterinary, sanitary, and phytosanitary laboratories and equipment, despite their considerable importance for raising agricultural productivity. Furthermore, the high levels of technical inefficiency in capital expenditures reduce the impact of even that small capital outlay to a very low level. Uganda's low level of capital expenditure is bad news, because the condition of rural infrastructure strongly affects agricultural trade, farm diversification, and ultimately poverty. A study in Uganda by Balat et al. (2008) demonstrates that greater trade in cash crops (coffee, tea, cotton, and fruits) will have a strong effect on poverty reduction. High marketing costs caused by poor rural infrastructure prevent many Ugandan farmers from participating in markets, leaving them little choice but to grow staple food crops in quantities sufficient for domestic consumption.2 The second drawback of agricultural sector expenditures is that they are increasingly dominated by nonwage recurrent expenditures, mainly subsidized inputs and other goods. Between 2005/06 and 2008/09, nonwage recurrent spending, mainly for farm inputs, comprised 64.6 percent of MAAIF's development budget on average. The share of nonwage Figure E4: Economic structure of the sector budget recurrent spending in MAAIF's total budget grew 250 from 49 percent in 2005/06 to 80 percent in 2008/09. Inputs 200 Capital are distributed by many outlays development projects and UShs billion 150 increasingly by NAADS. In Non-wage 2008/09, NAADS received 100 recurrent about UShs 37 billion "for inputs to the small-scale 50 Wage bill farmers who cannot afford to purchase the necessary inputs" - 2005/06 2006/07 2007/08 2008/09 (Figure E4). Making inputs accessible is integral to raising agricultural productivity, but the current approach of distributing inputs is unlikely to achieve this objective, because it does not strengthen private input suppliers, is not well targeted, or fails to invest in infrastructure to ensure that farmers can obtain inputs easily. Most input distribution and promotion appears to be supply driven and to favor wealthier farmers. It does not meet the requirements of market-supporting smart subsidies. As allocations for subsidized inputs grow, the fiscal burden and economic distortions (waste and leakages, 2 Balat, J., I. Brambilla, and G. Porto (2008): Realizing the Gains from Trade: Export Crops, Marketing Costs, and Poverty. World Bank Policy Research Working Paper 4488, Washington, DC. xiv as well as the displacement of input sales by the private sector) will also grow, if the mechanism for delivering inputs remains unchanged. The third problem with the allocative efficiency of the agricultural sector budget is the share of wage and operational expenditures going to MAAIF Headquarters versus other MAAIF departments. The recurrent budget of MAAIF Headquarters is 35 percent of MAAIF's recurrent budget. This large share is partially explained by higher wage allocations for senior staff, high transport costs to Kampala from Entebbe, and other recurrent expenses paid from Headquarters for services used by all MAAIF departments. The current allocation of operating costs between Headquarters and the technical/operational departments needs to be seriously reviewed. The current ratios of wage to nonwage recurrent costs are very high for the technical departments (with the exception of the Fisheries Department for the past two years), including such core functions as pest and disease control, while the ratio for Headquarters is in a quite comfortable range. The result is that operating funds for the technical departments are severely constrained, rendering them ineffective. Another way of assessing allocative efficiency is to analyze the functional composition of public expenditures. At first glance, the functional composition is quite efficient, given that the largest share of the budget is allocated to areas with the highest potential for enhancing pro-poor growth, namely advisory services and agricultural research. These categories absorbed 57 percent of the sector budget between 2005/06 and 2007/08 (Figure E5). Even so, many other core public goods remain underfinanced, undermining the potential impact of core functions like research and advisory services. Among these critical Figure E5: Functional composition of the sector budget areas are water capacity Physical infrastructure Promotion building, rural Processing & 15% 2% marketing Research infrastructure, pest and 3% Seed capacity 19% disease control (for plants development 5% and livestock), regulatory Water capacity services, and institutional building 4% development. These are Institutional the core public goods and development 2% essential investments Planning & policy 2% required at Uganda's Livestock & fish Advisory services current stage of regulatory services Plant pest and 38% 2% diseases Livestock diseases 7% development to catalyze 1% private investment in agricultural production and agribusiness. Investments in these public goods are also critical for ensuring that the geographically focused interventions of development projects can be sustained, in the sense that farmers retain access to markets and services even after development projects have ended and that technologies promoted by NAADS are available at the lowest possible cost. Given these drawbacks, how can the allocative efficiency of public expenditure be improved? More specifically, how can Uganda increase its expenditure on critical public goods without undermining fiscal sustainability? In the short run, it might not be xv possible to shift substantial resources to critical functions within the existing budget, because these resources are committed to ongoing development projects. Resources for underfinanced critical public goods would need to come from additional sources, including donor funds. In the medium to longer run, however, development resources must be shifted strategically from private to public goods and from nonwage recurrent to capital expenditures. Furthermore, the technical efficiency of agricultural spending requires improvement, as indicated in the next section, and those improvements will increase the impact of better-allocated expenditure on agricultural growth. Technical efficiency of public expenditure The analysis of technical efficiency indicates that much work remains to be done to make the case for substantially scaling up funding for agriculture in Uganda. Project implementation suffers from many serious problems, although not all of the blame rests with MAAIF. Many factors that are not specific to agriculture and often beyond MAAIF's immediate control significantly affect project implementation. They include Parliament's long delays of one to one-and-a-half years in ratifying loans, untimely and insufficient release of counterpart funds by MoFPED (with the outturn from 10 percent to 50 percent, depending on the development project), public officials' weak procurement and fiduciary capacity, and insecurity in northern Uganda. These constraints seriously affect the implementation of projects and increase their transaction and implementation costs. MAAIF needs to be much more proactive in bringing these issues to national attention and looking for concerted remedies at the national and local level. The remaining constraints on technical efficiency are significant and need to be addressed by MAAIF and local governments. Most MAAIF development projects with infrastructure components are characterized by long delays (three to five years) in building and rehabilitating rural infrastructure. Aside from the issues just mentioned, important problem areas include improper appraisal and feasibility work, poor coordination of preparation and implementation between MAAIF and local governments, inadequate operating budgets for local technical staff, ineffective project procurement, and problems related to land tenure.3 These result in high cost overruns, low-quality work, and other kinds of wastage. Given that capital expenditures account for such a large share of the development projects reviewed (81 percent of MAAIF's capital budget), MAAIF needs to take immediate action to devise remedies to deal with these problems. These lessons from experience should be taken into account when new investment projects are prepared. Other important lessons on technical inefficiencies emerge from studies of how goods and services are delivered to farmers and frontline service providers. First, the unit costs of goods procured at the central level is usually 20­50 percent higher than comparable market prices or unit costs at the local level. Second, the delivery of centrally procured goods is prone to wastage. The mortality rate of livestock distributed through 3 Land tenure problems are often related to improper appraisal and weak coordination of activities with local governments. xvi several projects is unacceptably high (ranging from 7 to 38 percent across interventions). In several instances, a significant number of cattle and goats died in the first months after delivery. The distribution of vaccines entailed low value for money and substantial wastage, because vaccines were overpriced and storage facilities inadequate. Records of financial and physical resource transfers from the central management units to the districts often do not match, suggesting wastage, diversion, or leakages, although the source of the problem remains unclear (possibilities include poor record keeping and thus low accountability, diversion from intended beneficiaries, corruption, resource misallocation, improper accounting, or simply insufficient information from project implementation units). Leakages were more pronounced for in-kind transfers than financial transfers, suggesting that inputs/goods are more easily directed away from their intended beneficiaries or uses. A key lesson from these experiences is to decentralize procurement to reduce costs and to place resources directly at the command of the beneficiaries to improve accountability. For example, such costs are lower for NAADS, because most procurement is decentralized and involves the local communities and beneficiaries. The impact of individual projects is considerably circumscribed by the lack of a strategic approach for using public expenditures to support agriculture. One example is the current strategy for improving farmers' access to inputs through subsidized input distribution by development projects, complemented by farmer training. This strategy fails to resolve the real causes of farmers' poor access to inputs, including high transaction costs (caused by poor rural infrastructure), slow progress in microfinance development, and weak technical and business capacity of private agro-dealers. Once the projects are phased out, inputs are still not accessible. Project expenditures are likely to have been wasted, given that they could have been used for infrastructure or other public investments with more sustainable outcomes. Another example is the promotion of animal vaccines in the absence of complementary disease control measures. Only 7 percent of Uganda's agriculture budget is allocated to controlling livestock diseases, and even that small sum is irregular and unpredictable. MAAIF can detect and control only one-sixth of the dangerous diseases listed by the World Health Organization and only those caused by ticks and tsetse. Diseases caused by viruses cannot be traced at all. The central veterinary lab is understaffed and ill equipped. Under these circumstances, the distribution of animals without appropriate follow-up services or support results in low value for money and wastage as a significant proportion of the animals die for lack of proper attention. The distribution of vaccines, even if it could be more effective, may not by itself be a cost- effective means of achieving the desired outcome--improved livestock health. The NAADS impact evaluation supports these findings about technical inefficiency in MAAIF development projects. NAADS has succeeded by being demand-driven and decentralized. The wastage and leakage associated with centrally managed programs is reduced because most NAADS activities and procurement occur locally. Members of NAADS had fairly equal access to advisory services regardless of wealth ranking, and they had greater access to information on various aspects of the farming enterprise than the non-NAADS households. Yet the impact of NAADS could have been even greater if its activities had been complemented by rural services and other xvii public investments, which could perhaps have been more equitably distributed than grants and private goods. NAADS alone will not substitute for insufficient finance, deteriorated roads, or expensive inputs. If complementary investments and actions are undertaken, NAADS will remain a powerful means of sustaining rural development in Uganda. Against this background, the following recommended actions would increase the effectiveness and technical efficiency of agricultural sector expenditure in Uganda: x Faced with the prospects of declining budget allocations in the future, the topmost priority for MAAIF is to ensure that its limited resources are used as efficiently and effectively as possible. MAAIF should target the highest priorities and seek institutional arrangements to more effectively deliver services through decentralized implementation arrangements. In line with the widely accepted principles of good governance, MAAIF should ensure transparency, accountability and participation in its service delivery in an effort to maximize the efficiency and effectiveness of its expenditures. x MAAIF should ensure that problems with launching projects are at the top of the agenda for national stakeholders. Rates of return to development projects will increase significantly if Parliament ratifies loans rapidly, government counterpart funds are released in a timely and predictable way (budget execution), and public officials improve their procurement and fiduciary capacity. Remedies for these problems have to be found at the national level, and MAAIF needs to play a critical role in this process. x The appraisal of infrastructure investments should be done in a participatory manner with local governments and local stakeholders. Land tenure issues must be clarified before projects are launched. Technical designs for communal roads and bridges should take into account connectivity with feeder and national roads. Where possible, procurement should be shifted from the central to district level, as in NAADS, and local government procurement staff should receive training. These actions will reduce unit costs, minimize wastage, improve the quality of work, and lead to the timely delivery of construction work. x In some instances, central procurement has merits and may be more cost effective. For example, central procurement of vehicles and other products procured internationally allows MAAIF to exploit economies of scale and effectively deal with import tax credits and other importation issues. For goods and services procured domestically, however, procurement must shift gradually to local governments. Compared to central procurement, local procurement usually has lower unit costs, wastage, and leakage. Local government should receive technical and financial assistance to strengthen procurement and fiduciary capacity and ensure the integrity of financial management. This support should be provided not only by MAAIF but also by national programs. xviii x During project implementation, wastage can be reduced through better coordination of activities and allocation of adequate operating funds for supervision by local production department staff. Local technical personnel are overburdened by multiple projects and lack funds to do their job. More resources should be set aside for operating expenses of local technical staff in development project budgets. x Individual projects should be better integrated into the strategy for agricultural growth and development. Public expenditures should address the roots, not the results, of market failures. It is inefficient to disseminate technologies in certain geographic areas without a coherent, sustainable approach to ensure farmers' access to those technologies once subsidies end. A larger share of public expenditure needs to be allocated to core public goods. The subsidization of private goods (inputs) must be considered only as a temporary measure to overcome market failures. The delivery of increased budgets for inputs and other goods should be directed at resource-poor farmers, which is not currently the case, and should not displace but engage the private sector. Budget processes The strategic improvement of public spending on agriculture must begin by making greater use of the Sector Budget Framework Paper (SBFP). The format and content of the SBFP are largely dictated by the guidelines in the Budget Call Circular issued by MoFPED. It is time to reassess the information currently required for the SBFP as outlined in the Circular to determine if it is essential and presented in the most effective format for making informed decisions on budget allocations. Except for 2006/07, there has been no feedback to the Sector Working Group (or MAAIF) from MoFPED on the content and quality of the SBFP, nor are there suggestions for improvement in future years. Such feedback should become a regular feature of the budget preparation process. Detailed recommendations for improving the SBFP guidelines are summarized in Annex 1. The credibility of MAAIF's Development Sector Investment Plan (DSIP) should be strengthened by giving greater attention to: (i) the criteria used for prioritization; (ii) the expected outcomes; (iii) detailed explanations of expenditure estimates; and (iv) linking the investment plans more closely to anticipated ceilings in the Medium-Term Expenditure Framework (MTEF), indicating how plans would change if MTEF ceilings were increased or decreased. There is an urgent need for MAAIF to update its strategy and investment plan for 2008/09 onwards to provide a continued link between policy, planning, budget preparation, and negotiations. MAAIF should also consider whether a three-year or five-year investment plan is best suited to its purposes. A five-year plan would provide a longer horizon for investment decisions and reduce the perceived need for frequent revisions, but changing priorities and events might make the estimates for outer years unreliable. Budget execution, the binding constraint for implementing projects, should take into account the strong seasonality of agriculture (and thus resource requirements) and the large costs of untimely, unpredictable counterpart financing xix in development projects. Within the overall cash flow, MoFPED should cater for the relatively small but particular cash flow requirements of agencies such as MAAIF and NAADS to improve their operational effectiveness. Substantial benefits will arise from creating a results-based system for monitoring and evaluation (M&E) of public expenditures. Such a system would make provisions for regular program monitoring and also for evaluating the impacts of major interventions. Currently MAAIF focuses on regular monitoring of project implementation. Such monitoring is necessary to follow the progress and achievement of targets; it should cover not only budget but also off-budget expenditure, which is likely to account for 10­20 percent of the sector expenditure. It is also critical to ensure that monitoring reports are used inside and outside of MAAIF to reward good performance (or invoke sanctions for poor performance), to address inefficiencies, or to reallocate resources between different priority areas. Rigorous impact evaluation should be done for most development projects. The budget for impact evaluation needs to be sufficient and reflected in project budgets. Results-based M&E is indispensable to good national management and policy planning. The regular monitoring system provides very little information about the real impact of public programs. The absence of a results-based M&E system makes it easy for interest groups to manipulate public debates and thus very difficult to change current approaches to public expenditures and agricultural policy in Uganda. Impact evaluations can range from Public Expenditure Tracking Surveys (PETS) to impact evaluations such as the one done for NAADS, both of which have been presented here. Once these evaluations become the norm, policy makers and planners will be well equipped to guide budget allocations across sectors and address operational constraints in agriculture programs. xx 1. Introduction 1. This Agriculture Public Expenditure Review (AgPER) comprehensively reviews public expenditures on agriculture in Uganda and analyzes their efficiency and effectiveness. Its genesis lies in Agriculture Sector Working Group (A­SWG) discussions, especially during the budget process, which raised concerns about the seemingly low budget allocations to the sector and the failure to align limited resources with recognized priorities in the sector. To address these concerns, the A­SWG resolved to undertake this AgPER. Aside from providing a better understanding of the nature and composition of agricultural pubic expenditures in Uganda, the Review would specifically analyze their efficiency and effectiveness with a view to identifying the types of expenditures that would promote pro-poor growth. 2. The AgPER is thus a joint product of the Ministry of Agriculture, Animal Industry, and Fisheries (MAAIF), the World Bank, and the Department for International Development (DfID). It was carried out under the overall guidance of the A­SWG and managed by a technical subcommittee of the Working Group, chaired by the Permanent Secretary, MAAIF. The AgPER's main objectives were to help the Government of Uganda better align future expenditures to its priorities, assess the robustness of the agricultural public expenditure management system to maximize the effectiveness of scarce resources, and make the case for the need and absorptive capacity for additional expenditure targeted at high-priority areas. 3. The rationale for the AgPER is that very little analytical information is available to ensure that public expenditures are prioritized to support the objectives of sustaining agricultural growth and reducing poverty. Such information is particularly crucial in a context where the national budget is increasingly stressed (by limited progress in generating domestic revenue, the demands of unforeseen crises, and other government obligations) and it is vital to manage aid flows prudently. The annual budget cycle is inevitably accompanied by complaints about insufficient allocations from sector ministries and implementing agencies. The Ministry of Finance, Planning, and Economic Development (MoFPED) has challenged line ministries to produce value-for- money analyses of current expenditures as a basis for considering increased budgetary allocations. Because of the weak information base, the annual sector Budget Framework Papers offer insufficient information on expenditures in the previous year and on what those expenditures achieved. Without this information, it is difficult to develop and justify a more educated budget allocation for the coming year. With few exceptions, annual budgets and Medium-Term Expenditure Framework (MTEF) ceilings present only incremental changes, irrespective of emerging sector priorities. 4. This shortcoming is evident to MAAIF, which is deeply concerned about the lack of resources as well as the limited potential for realigning expenditures to high- priority activities set forth in its Development Strategy and Investment Plan (DSIP). The bulk of MAAIF's budget is already allocated to specific activities, either for the core autonomous organizations--the National Agricultural Research Organization (NARO) and National Agricultural Advisory Services (NAADS)--or for activities to which the government has committed under various projects. MAAIF has a strong interest in 1 conducting a comprehensive AgPER to tailor its future expenditures to its priorities, improve the quality of public service delivery, and make a compelling case for appropriate levels of funding for agriculture. 5. Demand for the AgPER comes from several other quarters as well. The AgPER is part of the broader annual Public Expenditure Review, a core diagnostic undertaken collaboratively by the Government of Uganda and the World Bank. That diagnostic identifies potential allocative efficiencies in public expenditures that would create the space for high-priority investments to spur fiscally sustainable economic growth that benefits the poor. 6. Additional justification for the AgPER comes from the New Partnership for Africa's Development (NEPAD), under which Africa's Heads of State committed to allocate 10 percent of their national budgets to agriculture each year. Although prescriptive allocations need to be approached with caution--the quality of expenditure is equally if not more important than the quantity--under NEPAD's Comprehensive Africa Agriculture Development Programme (CAADP), one of the initial tasks for each country is to undertake a public expenditure review that documents the level, composition, and quality of expenditures in the agricultural sector. Uganda is among the first five countries to conduct such a review 7. DfID and the World Bank have established a global program to analyze and disseminate evidence on the levels and composition of public expenditure that stimulate pro-poor agricultural growth. Uganda was selected as one of the case study countries for this program primarily because the government--MoFPED as well as MAAIF--expressed demand for the information it could provide. The information would also be useful to local development partners, especially DfID and the World Bank, and the Regional Strategic Analysis and Knowledge Support System (ReSAKSS) for Eastern and Central Africa, which is mandated by the Common Market for Eastern and Southern Africa (COMESA) to monitor government spending on agriculture in the region. 8. This report synthesizes the findings from several studies that are key building blocks for the AgPER. The analysis was carried out in three phases. Phase I, conducted by Oxford Policy Management (OPM), focused on the levels and patterns of expenditure in the agricultural sector. Phase I findings and analysis have been updated and expanded by the task team to reflect more recent data. Phase II (also carried out by OPM) reviewed the budget process and performance. Phase III studied the efficiency and effectiveness of agricultural public expenditures using a variety of tools. These included a public expenditure tracking study, carried out by the Economic Policy Research Centre (EPRC), and a household survey, conducted by the International Food Policy Research Institute (IFPRI), to analyze the incidence of key public expenditures and evaluate the impact of NAADS, which accounts for the largest share of agricultural public expenditures in Uganda. 9. The report is structured as follows. Section 2 analyzes trends in nominal and real sector budgets. It highlights the current and projected importance of agricultural sector expenditure in the national budget and gross domestic product (GDP). An analysis 2 of agricultural price distortions indicates the extent to which the sector benefits from supportive policies. The allocative efficiency of public spending is evaluated to determine whether scarce public resources are allocated in ways that reflect government priorities for agriculture, with the proper combination of recurrent and capital expenditures. 10. Section 3 describes budget planning and implementation at the national and local level and presents policy recommendations to improve those processes. Section 4 focuses on the technical efficiency of public spending. By tracking resource flows and analyzing the unit costs of goods and service delivery, Section 4 sheds light on whether public resources are used efficiently and which actions could improve efficiency. The concluding section summarizes the major findings and policy recommendations. 3 2. Level and Composition of Agricultural Sector Expenditure in Uganda 2.1. Trends in agricultural sector expenditure 11. Public resources are expended on agriculture by a number of agencies at the national and local level. At the national level, MAAIF is the lead ministry responsible for agricultural expenditure, along with four autonomous organizations: (i) NARO, (ii) the NAADS Secretariat, (iii) the Uganda Cotton Development Organization (UCDO), and (iv) the Uganda Coffee Development Agency (UCDA). At the local level, District Agricultural Extension, NAADS, and (through nonsectoral conditional grants) programs such as those under the Plan for Modernization of Agriculture (PMA) also manage agricultural expenditures. 12. The approved budget for Uganda's agricultural sector grew by 46 percent from 2001/02 to 2008/09, driven by a threefold rise in spending for NAADS through the districts.4 Cumulative spending over the same period was UShs 1,346 billion (Figure 1). In real terms, the approved budget for the sector was stagnant over this period by (in fact lower by about 2 percent), largely because of a sharp decline in the early 2000s. Between 2004/05 and 2008/09, however, the budget grew by 38 percent in real terms.5 Figure 1: Trends in nominal and real approved budgets for agriculture, 2001/02­2008/09 (UShs billions) 240 Nominal sector budget Real sector budget 220 200 180 USHs billion 160 140 120 100 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 2008/09 Source: MoFPED and MAAIF. 4 Throughout this report, data for 2008/09 are derived from MoFPED (2008): Draft Estimates of Revenue and Expenditure (Recurrent and Development). Kampala. 5 The real sector expenditure is estimated as the nominal sector expenditure adjusted for inflation. 4 13. The funds that MoFPED released to the agricultural sector were usually smaller than the budget approved by Parliament. During 2001/02­2005/06, the released budget for agriculture was 10 percent lower than the approved budget. The cumulative growth apparent in the approved budget (7 percent) was belied by the cumulative decline in the released budget (15 percent) (Figure 2).6 In real terms, the rise in spending is also less spectacular for the released budget (2 percent) compared to the approved budget (28 percent). Figure 2: Released versus approved budgets for the agricultural sector, 2001/02­2005/06 (UShs billions) 180 160 Nominal approved budget 140 120 Nominal released budget 100 UShs bill 80 Real approved budget 60 Real released 40 budget 20 0 2001/02 2002/03 2003/04 2004/05 2005/06 Source: MoFPED and MAAIF. 14. MAAIF, NARO, and the NAADS Secretariat receive the largest share of the sector budget, although their share has declined very significantly since 2001/02. These agencies received 89 percent of the budget in 2001/02 compared to 55 percent in 2008/09 (Figure 3). As budget allocations were increasingly decentralized, the share of sector budget going to NAADS at the district level rose from only 2 percent in 2001/02 to 36 percent in 2008/09. According to the MTEF, 42 percent of the sector budget is expected to be allocated directly to the sub-national level in 2008/09, a share that is projected to increase to 51 percent by 2012/13. 6 As of this writing, no data are available on the released budget after 2005/06. 5 Figure 3: Structure of the approved budget for agriculture, 2001/02­2008/09 (UShs billions) 250 Non-Sectoral Conditional Grant 200 NAADS (districts) 150 District Agricultural Extension UShs bill UCDA 100 UCDO 50 MAAIF 0 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 2008/09 Note: MAAIF budget includes the budgets of NARO and the NAADS Secretariat. Source: MoFPED and MAAIF. 15. Over time, the approved budget for agriculture in relative terms has declined from a level that was already quite moderate. In 2008/09, agriculture's share in the national budget and GDP is estimated at 3.8 percent and 1.6 percent, respectively, lower than in 2001/02 (Table 1). The released sector budget has declined as well, leaving agriculture with only 1.2 percent of GDP and 4.8 percent of national expenditure in 2005/06.7 This small budget for agriculture is a long way from the 10 percent share that Uganda pledged to allocate under the Maputo Declaration of CAADP. 16. In the medium run, national expenditures on agriculture are projected to continue declining. According to the MTEF, the agricultural sector's share in the national budget will be 3.2 percent in 2012/13, compared with 3.8 percent in 2008/09 and 4.6 percent over 2001/02­2007/08 (Table 2). The budget for agriculture as a share of GDP will also decline from 1.6 percent in 2008/09 to 1.4 percent in 2012/13. Uganda plans to shift resources to other priorities for sustaining its long-term economic development, notably to roads, education, and health, which are also expected to help agriculture and rural areas more broadly, with better access to markets, opportunities for developing agro-processing industries, and better human capital (Figure 4). 17. In light of this outlook, it is imperative that the shrinking national budget for agriculture is used as effectively and efficiently as possible at all levels. Strategically, it will increasingly become more critical for MAAIF to work in collaboration with local 7 The share of agricultural sector expenditure in agricultural GDP also decreased during the period under review. In 2003/04, the agricultural budget amounted to 9 percent of agricultural GDP, but in 2008/08 it fell to 6.8 percent (based on the World Development Indicators data on agricultural value added). 6 governments, which are expected to take increasing responsibility for implementing public programs as decentralization continues. Table 1: Shares of the agricultural sector budget in the national budget and GDP, 2001/02­2008/09 (percent) 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 2008/09 Approved agriculture budget 5.7 5.6 4.1 3.8 4.4 4.6 4.3 3.8 as share of national budget Released agriculture budget as share of 8.2 5.5 3.8 3.3 4.8 n/a n/a n/a national budget Approved agriculture budget n/a n/a 1.5 1.4 1.5 1.7 1.6 1.6 as share of GDP Released agriculture budget as share of n/a n/a 1.4 1.2 1.2 n/a n/a n/a GDP Source: MoFPED and MAAIF. Table 2: Medium-Term Expenditure Framework for the agricultural sector budget, 2008/09­2012/13 (UShs billions) Average Ministry/agency 2001/02­ 2008/09 2009/10 2010/11 2011/12 2012/13 2007/08 MAAIF 122.6 122.0 114.5 130.1 131.1 134.4 UCDO 1.5 5.7 7.0 4.2 5.4 6.6 UCDA 2.3 0.9 4.6 2.2 2.7 2.8 District Agricultural 6.3 7.2 9.2 14.2 14.2 14.2 Extension NAADS (districts) 21.2 81.2 98.7 103.7 108.7 113.7 Nonsectoral conditional grants 6.5 6.2 13.2 18.8 18.7 18.7 TOTAL BUDGET 160.4 223.2 247.2 273.2 280.8 290.5 Agriculture budget as a 4.6% 3.8% 3.7% 3.5% 3.4% 3.2% percentage of national budget Agriculture budget as a n/a 1.6% 1.5% 1.6% 1.3% 1.4% percentage of GDP Note: MAAIF budget includes NARO and NAADS Secretariat. Source: MoFPED and MAAIF. 7 Figure 4: Agriculture's share of the national budget in relation to other sectors, 2007/08­2012/13 (UShs billions) 10,000 Other sectors 9,000 Interest 8,000 7,000 Public Administration 6,000 Security UShs bill 5,000 Water & Environment 4,000 Roads & Works 3,000 Health 2,000 Education 1,000 Agriculture - 2007/08 2008/09 2009/10 2010/11 2011/12 2012/13 Source: MoFPED/MTEF (2008). 18. The national expenditure on agriculture is slightly larger when analyzed based on the United Nations Classification of Functions of Government (COFOG), but it is still far from the 10 percent committed under CAADP. The COFOG classification includes budgetary allocations for forestry, water for production, and issues related to agricultural land, which are directed to the Ministry of Water, Lands, and Environment (MWLE) and the National Forest Authority (NFA). Table 3 shows the approved and released budgets for agriculture based on the COFOG classification. In nominal terms, the approved budget in 2005/06 was the same as it was in 2001/02, after declining by around 10 percent in 2003/04 and 2004/05--similar to the trend in MAAIF's approved budget (Figure 1).8 As a share of the approved national budget, allocations to agriculture based on the COFOG classification declined continually from 8.1 percent in 2001/02 to 5.7 percent in 2005/06. The released budget for agriculture as a percentage of the national budget fell from 6.9 percent to 5.4 percent over the same period. 8 The data for approved and released budgets using the COFOG definition of agriculture is constructed only until 2005/06 based on data available. 8 Table 3: Expenditures on agriculture based on the COFOG classification, 2001/02­2005/06 (UShs billions) 2001/02 2002/03 2003/04 2004/05 2005/06 Budget Release Budget Release Budget Release Budget Release Budget Release CENTRAL GOVERNMENT MAAIF* Recurrent 7.45 5.82 7.85 6.58 6.64 7.22 11.46 10.71 14.49 13.55 Development 127.85 66.81 126.16 82.08 90.66 58.25 85.06 39.37 106.07 60.71 Other COFOG Lands, Forestry, and Water-for-Production Recurrent 1.51 1.12 1.84 1.10 1.38 1.12 0.19 0.16 0.24 0.21 Development 63.71 39.58 56.92 31.29 66.46 57.34 48.59 20.83 33.30 19.90 Agencies DDA 0.50 0.50 0.16 0.16 0.24 0.24 0.22 0.22 0.22 0.22 UCDO 1.26 1.26 0.95 0.95 1.77 1.77 1.98 1.98 1.03 1.03 UCDA 2.12 1.85 1.89 1.76 2.78 2.25 2.61 2.69 2.76 3.12 NFA 0.00 0.00 0.00 0.00 1.03 1.03 10.76 10.76 13.25 13.25 LOCAL GOVERNMENT Extension Wage 2.47 2.16 3.06 2.83 3.06 3.19 3.06 3.89 3.08 3.85 Nonwage 3.00 2.90 2.92 2.73 2.92 2.81 2.92 2.92 2.92 2.78 NAADS (Districts)** 2.57 2.42 5.66 9.32 14.27 13.75 16.02 15.13 27.46 24.87 NSCG*** 6.07 4.44 5.76 4.09 5.96 4.20 6.38 4.01 6.18 4.22 LGDP**** - 1.62 - 1.06 - 1.69 - 1.42 - 0.67 TOTAL 218.51 130.48 213.17 143.95 197.17 154.86 189.25 114.09 211.00 148.38 Government of Uganda expenditure 2,686 1,895 2,768 2,720 3,107 3,128 3,380 3,369 3,716 2,760 COFOG agriculture budget as % of national budget 8.1 6.9 7.7 5.3 6.3 5.0 5.6 3.4 5.7 5.4 Note: * Includes NARO and NAADS Secretariat. ** From 2003/04 an amount appears in this Vote under Donor Development. This amount cannot be accessed by the NAADS program, and yet in at least one year it was reported as fully spent. *** NSCG data were made available by the PMA Secretariat and analyzed by the AgPER Task Team to identify COFOG-relevant agricultural expenditure. **** LGDP figures for COFOG-relevant agricultural expenditures are based on the AgPER Task Team's analysis of data provided by the LGDP Co-ordination Unit. Source: MoFPED Aid Liaison Department database; information collected by the AgPER Task Team from MoLG, PMA Secretariat, and NFA; DDA, NFA, UCDO, and UCDA annual reports; MoFPED, various MTEF projections. 19. Aside from the expenditure already described, there is considerable off- budget expenditure on agriculture, both centrally and locally, but Uganda lacks a mechanism for consistently monitoring these funds. The expenditures are fragmented and it is difficult to obtain information on what is being spent and for what purpose. Table 4 provides an example of development assistance provided by the United States Agency for International Development (USAID) and the Swedish International Development Agency (SIDA), two important players in the agricultural sector in Uganda who were the only ones willing or able to share the information with the AgPER study 9 team.9 Neither agency channels funds to agriculture through the national budget. Both agencies operate completely independently and maintain only ad hoc links to MAAIF or its semiautonomous agencies. Table 4 shows that funds provided to agriculture by just these two off-budget development partners added the equivalent of 10­20 percent of the released budget for agriculture during the period under review. Aside from USAID and SIDA, other development partners providing off-budget support for agriculture include GTZ, the Food and Agriculture Organization of the United Nations (FAO), the United Nations Development Program, the Government of France, and the Japan International Cooperation Agency (JICA).10 Clearly it is challenging for MAAIF to perform its key functions of planning, coordinating, and prioritizing investments to achieve sectoral development goals if agencies providing significant sums of money to the sector operate independently, in many instances without any links or coordination with MAAIF. Table 4: Off-budget disbursements to agriculture (based on the COFOG classification) by two donors, 2000/01­2006/07 (UShs billions) 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 USAID 15.16 10.32 15.29 23.22 20.86 19.55 13.77 SIDA 1.75 1.32 2.24 0.81 0.68 1.33 0.00 Total 16.91 11.64 17.53 24.03 21.53 20.88 13.77 COFOG total n/a 130.48 143.95 154.86 114.09 148.38 n/a. USAID / SIDA as % of 8.9% 12.3% 15.5% 19.9% 14.1% COFOG total Note: From the data provided by the two agencies, it was not possible to review the items of expenditure to make sure that they complied with the COFOG definition of agriculture. After discussions with the agencies' program officers, it was decided to include all USAID funding classified as "Agriculture" and none classified as "Environment," whereas 75 percent of SIDA's "rural development/agriculture (natural resources and environment) disbursements" were included. Source: USAID and Embassy of Sweden, Kampala (May 2007). 20. At the same time, in addition to the public expenditures noted above, nongovernmental organizations (NGOs) have become an important presence in agriculture, particularly in recent years. Many of these organizations provide free agricultural inputs to the northern districts, mainly to internally displaced people. Such inputs include livestock, seed, plant cuttings, pesticides, veterinary drugs, farm tools, and crop processing equipment.11 Based on data collected by FAO in a survey of about 25 agencies providing agricultural inputs to roughly 300,000 households in northern districts during the first season of 2007, it is estimated that the value of inputs supplied amounted to almost UShs 9 billion.12 Given that agricultural inputs are also supplied in the second season (though to fewer farm households), the annual value of agricultural inputs provided by NGOs and other humanitarian agencies is estimated to be UShs 14.4 billion. This sum is equivalent to almost 10 percent of the COFOG-based expenditures on 9 SIDA no longer supports agriculture directly, but its funding was important at the turn of the century. 10 Some of the projects of these donors are also on-budget. 11 As well as vouchers to purchase such items; the vouchers are earned through participation in rural works programs. 12 Food Security Group (2007). 10 agriculture in 2005/06. Considerable care has to be taken, however, to avoid double- counting in aggregating donors' budget and off-budget expenditures, because a high proportion of the national budget is provided by donors. Once again, neither MAAIF nor the District Production and Marketing Directorates have much knowledge of these programs or activities, adding to the challenges of sector coordination and harmonization. 21. How does Uganda's spending on agriculture compare to spending by other countries? An international comparison shows that Uganda spends relatively less than other countries (Table 5), when spending is measured as the share of agricultural budget in GDP (adjusted by the size of the sector). Uganda spends about as much as other Sub- Saharan African countries but less than middle- and high-income countries. This finding is consistent with Uganda's narrower fiscal capacities and its greater need to support competing public investments in infrastructure and social sectors compared to higher- income countries. 22. International experience also illustrates that lower public spending does not always translate into lower agricultural competitiveness. It is well known that in the United States and the European Union, for example, expensive public support in the form of huge farm subsidies has failed to keep farmers competitive in world markets, whereas lower spending in Brazil, Australia, and New Zealand has not prevented their farmers from being highly competitive. Three conditions must be met to ensure that public expenditures in agriculture are effective. First, public expenditures on agriculture must be supported through an enabling environment in which agricultural prices are subject to few distortions. It is counterproductive to raise the public expenditure on agriculture when farm-gate prices are depressed. Second, the allocative efficiency of public expenditures should be high. Spending that does not contribute (or contribute as much as other types of spending) to major objectives for agriculture, such as economic growth, private sector investment, and poverty reduction, is allocatively inefficient or unproductive. Third, the technical efficiency of public expenditures should be high. Technical efficiency in the public sector involves making the best use of inputs to provide outputs in the form of public services. In other words, technical efficiency means doing things well and allocative efficiency means doing the right things. When these three conditions are met, it is highly probable that even small budgets will be well-positioned to achieve national objectives for agriculture. The sections and chapters that follow provide a detailed examination of each of these conditions in Uganda. 11 Table 5: International comparison of budgets for agriculture, average for 2002­04 Income group and country Agriculture as Agriculture Agriculture budget as percentage of budget as percentage of GDP, GDP percentage of adjusted to the size of GDP the agricultural sector in each country A C C/A High-income countries Australia 3.0% 0.31% 0.10 Canada 2.3% 0.51% 0.22 EU 2.3% 0.65% 0.28 USA 1.6% 0.73% 0.46 Middle-income countries Turkey 13.0% 2.0% 0.15 Mexico 4.0% 0.7% 0.18 Venezuela 5.0% 0.5% 0.12 China 15.0% 1.2% 0.08 Brazil 9.3% 0.7% 0.08 Russia* 6.0% 0.95% 0.16 Ukraine 11.6% 1.3% 0.11 Low-income countries Uganda 32% 1.5% 0.05 Tanzania 45% 1.2% 0.03 Ethiopia** 44% 2.7% 0.06 Kenya 29% 1.3% 0.04 Note: * Data for Russia are for 2003. ** Data for Ethiopia are for 2004/05. To make the data for Ethiopia comparable to data for other countries, the AgPER Task Team excluded transfers under the vulnerability and food security program and expenditures on rural energy, mining, federal roads, and water supply. Source: Estimates derived from: Organization for Economic Cooperation and Development (2004): Agricultural Policies in OECD Countries: At a Glance. Paris; Tangermann, S. (2006): OECD Work on Agricultural Policies in Brazil, China, India, and South Africa. Presentation at Rural Week of the World Bank, Washington, D.C., February 27, 2006; Zorya, S. (2006): Improving Agricultural Fiscal Policy in Ukraine. World Bank ECSSD Working Paper 36970, Washington, DC; World Bank (2006): Ethiopia Agricultural and Rural Development: Public Expenditure Review for 1997-98 and 2005-06. Washington, DC; and Zorya, S. (2008): Rapid Assessment of the 2008/09 Agriculture Public Expenditure in Tanzania. Washington, DC: World Bank. 2.2. Agricultural policy environment 23. Uganda has successfully addressed a key requisite for efficient public spending by largely eliminating agricultural price distortions. Uganda's agricultural policies provide a conducive environment in which even scarce public resources can make a difference. The recent World Bank study on agricultural incentives in Uganda 12 shows the shift from agricultural taxation to agricultural support in recent years. 13 The Direct Rate of Assistance (DRA) to agriculture was 1 percent during 2001­2004, with zero support to exportable products and 13 percent support to importable products through import tariffs (Table 6).14 For exportable products, this absence of support represents a large shift from the high taxation (depressed prices) prevalent between 1961 and 1995. The level of support for importable products represents a reduction in taxation of poor consumers that has only small negative effects on farmers, because Uganda's farmers produce only small volumes of these importable products. Table 6: Distortion indicators for Ugandan agriculture, five-year averages 1961­ 1966­ 1971­ 1976­ 1981­ 1986­ 1991­ 1996­ 2001­ COMMODITY 65 70 75 80 85 90 95 2000 04 Coffee ­0.15 ­0.39 ­0.64 ­0.89 ­0.73 ­0.74 ­0.20 ­0.02 ­0.01 Cotton ­0.13 ­0.22 ­0.53 ­0.80 ­0.47 ­0.51 ­0.04 0.00 0.00 Exportables DRA ­0.11 ­0.28 ­0.59 ­0.88 ­0.66 ­0.66 ­0.08 ­0.01 0.00 Rice 0.14 0.21 0.52 0.49 0.48 0.36 0.07 0.13 0.18 Importables DRA 0.16 0.22 0.52 0.81 0.48 0.54 0.14 0.14 0.13 Changing status commodities Maize ­0.01 0.08 0.10 0.17 0.00 ­0.13 ­0.04 0.07 0.00 Beans 0.11 ­0.07 0.00 0.00 0.00 0.00 ­0.04 0.04 0.00 DRA agriculture 0.00 ­0.04 ­0.08 ­0.14 ­0.07 ­0.07 ­0.01 0.02 0.01 DRA nonagriculture 0.08 0.09 0.13 0.15 0.13 0.13 0.08 0.10 0.08 TRA agriculture ­0.08 ­0.13 ­0.21 ­0.29 ­0.20 ­0.19 ­0.09 ­0.09 ­0.07 Source: Matthews et al. (2006). 24. The aggregate effect of direct distortions on the agricultural sector depends not just on the size of agricultural policy interventions but also on the size of distortions generated by policies in sectors outside agriculture. The Total Rate of Assistance (TRA) to agriculture measures the size of direct distortions in agriculture relative to those in other sectors. The latter are captured by the nominal direct rate of assistance to nonagricultural production. The higher the assistance to nonagricultural production, the more other sectors are in a position to attract resources away from agriculture, adding to the discrimination against this sector or reducing the value of any positive assistance that may be granted. 13 Matthews, A., P. Claquin, and J. Opolot (2006): Distortions to Agricultural Incentives in Uganda. World Bank Agricultural Distortions Research Project. Washington, DC. 14 The methodology for estimating DRA resembles the widely used OECD methodology for "rate of protection." During 2006­07, the World Bank Agricultural Distortions Research Project (which prepared reports for many countries, including Uganda) adjusted the original OECD Producer Support Estimates to the Bank's methodology for producer assistance, and the new term "rate of assistance" was introduced to distinguish between Bank and OECD estimates. 13 Figure 5: Agricultural sector Total Rate of Assistance, Uganda, 1961­2004 0.6 Post-Independence Collapse Recovery Post-liberalization 0.4 0.2 0.0 -0.2 -0.4 -0.6 1961 1963 1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 NRA non-agriculture DRA agriculture TRA agriculture Source: Matthews et al. (2006). 25. Various policy measures were included in computing the nonagricultural rate of assistance: customs duties, export taxes (applied on copper, hides, and skins in some years up to 1977), the import commission and withholding tax, and the differential application of sales tax and VAT. Other nontariff barriers were not included in the absence of specific information; these may have been important in earlier decades but have been phased out since market liberalization. Table 6 and Figure 5 illustrate the notable decrease in nonagricultural rate of assistance and corresponding improvement in the TRA to agriculture. As a result, during 2001­04 some small positive protection (1­2 percent) of the agricultural sector arose from direct policies alone (due to the continued protection of importables after government interventions were abolished on exportables). Even so, this level of protection remains smaller than the positive protection of the nonagricultural sector, which averages 8 percent. Thus at present a small negative bias remains in incentives for agricultural production in Uganda. This small bias does not have a large negative impact on spending in agriculture, however, so the current policy environment for agriculture in Uganda may fairly be described as enabling public expenditure to have a lasting impact. 2.3. Economic composition of the sector budget 26. What determines the allocative efficiency of public expenditures on agriculture? Spending is allocated efficiently when it is used to support the "right" things--in this instance, the priorities that are most capable of spurring pro-poor growth. The analysis of allocative efficiency examines the economic and functional composition 14 of the budget for agriculture.15 The economic composition is presented first for MAAIF and then for the overall budget for agriculture. 27. In determining the economic composition of public expenditures on agriculture, expenditures are first classified into recurrent and development expenditures. Recurrent expenditures are further divided into "wage bill" (employee salaries, allowances, and other wage costs) and "nonwage" recurrent expenditures. Because "development" and "capital" expenditures no longer synonymous, they are decomposed into "current" and "capital" expenditures. 28. The development budget constitutes the largest share of MAAIF's budget. In 2001/02, its share was 96 percent and remains substantial at 76 percent in 2008/09 (Figure 6). On average, the share of development spending in MAAIF's budget was 88 percent during the period under review. Figure 6: MAAIF recurrent and development budgets, 2001/02­2008/09 (UShs billions) 100 90 80 70 60 Development UShs bill expenditure 50 40 30 Recurrent 20 expenditure 10 0 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 2008/09 Source: MoFPED (various years). 29. The structure of the recurrent budget for MAAIF is presented in Table 7. The recurrent budget consists of employee costs (the wage bill), use of goods and services, grants to other organizations, and domestic arrears. MAAIF's recurrent budget fluctuated significantly because of substantial variation in grants and in allocations to pay domestic arrears. On average, these two items constitute 55 percent of recurrent budget allocations, with the remainder covering the more common items in a recurrent budget, such as wages and recurrent operational costs of MAAIF staff. 15 Given the lack of detailed data on released budgets, the analysis of economic and functional composition relies only on data for approved budgets. 15 Table 7: MAAIF recurrent budget allocation, 2001/02­2008/09 (UShs billions) Year Wage bill Use of goods and Grants Including Domestic Total services grants as a arrears budget wage subvention 2001/02 1.57 2.17 1.56 0.03 3.18 8.48 2002/03 2.22 1.48 0.56 0.03 0.75 5.01 2003/04 2.22 1.93 6.25 2.77 0.17 10.57 2004/05 2.17 1.11 4.76 1.99 0.00 8.04 2005/06 2.47 1.94 3.48 2.03 2.25 10.13 2006/07 2.49 2.26 3.13 1.55 1.46 9.35 2007/08 2.53 4.01 3.49 1.89 4.38 14.41 2008/09 2.60 4.30 4.98 1.94 4.87 16.76 Source: MoFPED (various years). 30. What grants are made by MAAIF? Table 8 lists all grants and shows the importance of these subventions to autonomous institutions. On average, these subventions made up about 90 percent of the grants over the period under review, with half being wage subventions. Grants have grown larger since 2003/04. Large grants are channeled mainly by the following programs: x Crop Production Department, to finance the NAADS Secretariat before it was allocated a separate Vote. x Directorate of Animal Resources, to support the National Animal Genetic Resource Centre and Databank. x Department of Animal Production, to finance the Dairy Development Authority. x Department of Farm Production, to fund the PMA Secretariat and, in 2006/07, a contribution to the Inter-Governmental Authority on Drought and Development. x Department of Livestock Health and Entomology, to fund the Coordinating Office for the Control of Trypanosomes in Uganda. Table 8: Allocation to grants in the MAAIF approved recurrent budget, 2001/02­2008/09 (UShs billions) 2001/02­ 2003/04­ 2007/08 2008/09 2002/03 2006/07 1. Other grants 0.24 0.22 0.13 0.13 2. Contributions to international organizations 0.17 0.20 0.19 0.19 3. Grants to central government ministries 0.73 0.15 0.03 0.27 4. Contributions to autonomous organizations 0.01 1.80 1.25 2.46 5. Wage subventions to autonomous organizations 0.03 1.96 1.89 1.94 Total grants 1.18 4.33 3.49 4.98 Source: MoFPED (various years). 31. Domestic arrears are another share of the recurrent budget that is not available for MAAIF to pursue its activities. Domestic arrears are debt liabilities to 16 various entities that have provided goods and services to MAAIF in previous years, such as input suppliers and contractors. In 2001/02, domestic arrears constituted 38 percent of MAAIF's recurrent budget (Table 7). Although the arrears were fully repaid by 2004/05, they began to grow again in 2005/06. In 2008/09, MAAIF had to use 29 percent of its recurrent budget to pay debts. The rising accumulation of domestic arrears demands that serious attention be given to paying bills on time. Otherwise the Ministry's future budgets will be utterly consumed by debt payments. 32. After the deduction of grants and domestic arrears, how was the remaining recurrent budget spent? "Headquarters" has typically absorbed the lion's share of the approved recurrent budget (Table 9), with a 34 percent share in 2008/09, down from 40 percent in 2005/06. The allocation to Headquarters is high partly because Headquarters finances some expenses incurred by all departments, such as telecommunications, electricity, water, and security guards (25 percent of the Headquarters budget). Other major costs (42 percent) include travel (inland and abroad), fuel, and vehicle maintenance; one reason for this is that MAAIF is located in Entebbe and many meetings taken place in Kampala. Yet the high Headquarters allocation is not fully explained by any of these facts. It remains important to reduce Headquarters costs with the objective of shifting more resources to other departments so that MAAIF's technical staff can do their work. The crop programs had usually received one-quarter of the recurrent budget but now receive less, as funds have been shifted to fisheries. The share of the livestock subsector doubled between 2005/06 and 2008/09, when it absorbed 42 percent of MAAIF's recurrent budget. Allocations to the Department of Planning have fallen in recent years. Table 9: Percentage distribution of MAAIF's recurrent budget by department (excluding grants and domestic arrears), 2005/06­2008/09 2005/06 2006/07 2007/08 2008/09 01 Headquarters 39.6 40.6 32.4 33.9 02 Directorate of Crop Resources 0.9 0.9 0.6 0.6 03 Farm Development Department 7.5 7.2 5.1 4.8 04 Crop Protection Department 17.0 16.8 12.4 11.1 05 Crop Production Department 4.4 3.6 3.2 3.0 Subtotal crops 29.8 28.4 21.2 19.5 06 Directorate for Animal Resources 1.4 1.2 1.1 1.1 07 Animal Production Department 6.4 7.1 4.9 7.5 08 Livestock Health and Entomology 7.6 7.0 5.1 4.8 09 Fisheries Department 8.9 8.7 30.8 28.9 Subtotal livestock 24.5 24.0 41.9 42.4 10 Department of Planning 6.1 6.9 4.5 4.3 Total recurrent budget (UShs billions) 4.40 4.76 6.54 6.90 Source: MoFPED (various years). 33. How has MAAIF's recurrent budget been distributed between wage and nonwage expenditures? Until 2007/08, each type of expenditure accounted for about half of the recurrent budget. The most important nonwage recurrent expenses were travel, fuel, and vehicle maintenance. Starting in 2007/08, nonwage recurrent expenditures grew to almost two-thirds of MAAIF's recurrent budget, owing to the large allocation (UShs 17 1.6 billion per annum) to the Fisheries Department (Table 10). However, going beyond the average for MAAIF, the breakdown across different spending units shows that the ratio of wage to nonwage recurrent expenditures is very unevenly distributed. While the ratio is in a comfortable range for MAAIF Headquarters, and for the past two years for the Fisheries Department, for all other technical departments/units, including core functions such as pest and disease control, the ratio is very high, indicating that most of these departments are severely constrained in terms of operating funds, which prevents them from being effective. Table 10: Main parts of MAAIF's recurrent budget (excluding grants and domestic arrears), 2005/06­2008/09 (UShs billions) 2005/06 2006/07 2007/08 2008/09 Wage bill 2.47 2.49 2.53 2.60 Salaries 96.5% 96.6% 96.3% 93.3% Allowances and other employee costs 3.5% 3.4% 3.7% 6.7% Nonwage recurrent 1.94 2.26 4.01 4.30 Goods and services 3.0% 3.0% 42.1% 43.4% Travel 20.4% 22.3% 12.3% 13.5% Fuel and lubricants 10.8% 9.5% 5.5% 6.1% Vehicle maintenance 11.3% 9.1% 5.9% 5.3% Recurrent budget 4.40 4.76 6.54 6.90 Source: MoFPED (various years). 34. What are the trends in MAAIF's development budget? As noted, between 2005/06 and 2008/09, development expenditure accounted for about 83 percent of MAAIF expenditures. MAAIF has executed 24 development projects. The seven principal projects constituted 82 percent of the development budget. 16 Donors financed 85 percent of MAAIF's development budget, as government allocations diminished from 20 percent in 2005/06 to 13 percent in 2007/08 (Table 11). The main donors are African Development Bank (ADB), World Bank, International Fund for Agricultural Development (IFAD), the European Union, the Danish International Development Agency (DANIDA), and the United Nations Development Program. When one adds the development expenditure of NARO and NAADS, however, the share of donor finance in the agricultural development budget declines and the share of government financing rises to 49 percent in 2007/08 and (even higher) 65 percent in 2008/09. 16 These projects are the North-West Smallholder Agricultural Development Project (11.6 percent); National Livestock Productivity Improvement Project (19.7 percent); Livestock Disease Control (6.6 percent); Support to Fisheries Development (17.1 percent); Farm Income Enhancement Project (8.5 percent); Vegetable Oil Development Project (15.6 percent); and Farming in Tsetse Areas of East Africa (3.6 percent). 18 Table 11: Sources of funding for the development budget, 2005/06­2008/09 (UShs billions) 2005/06 2006/07 2007/08 2008/09 Average Share, % MAAIF Donor financing 51.21 51.89 63.02 44.17 52.57 84.2% Government financing 7.28 12.98 9.67 9.63 9.89 15.8% MAAIF development budget 58.49 64.87 72.69 53.79 62.46 100.0% NARO Donor financing 21.19 19.22 20.21 54.5% Government financing 17.20 17.20 17.2 45.5% NARO development budget n/a n/a 38.39 36.44 37.42 100.0% NAADS Donor financing 1.47 1.47 2.3% Government financing 54.08 92.54 73.31 97.7% NAADS development budget n/a n/a 55.55 92.54 74.05 100.0% TOTAL AGRICULTURAL SECTOR BUDGET Donor financing 85.68 63.39 74.25 42.7% Government financing 80.95 119.37 100.40 57.7% Total development budget n/a n/a 166.63 182.77 173.93 100.0% Note: The development budget may deviate slightly from the official estimate because different exchange rates were used to convert foreign financing from US dollars and Euros to Ugandan shillings. The year- average market exchange rates reported by Uganda Central Bank were used for this AgPER. Source: MoFPED (various years). 35. What share of the development budget is allocated to wages, nonwage recurrent, and capital outlays? These data were not readily available but were collected through rigorous analysis of documents and information for each MAAIF development project. Table 12 indicates that nonwage recurrent expenditures made up 65 percent of MAAIF's development budget over the study period. The share of capital outlays ranged from 19 percent in 2006/07 to 25 percent in 2008/09. The time has long since passed when "capital" and "development" expenditures were synonymous, and Table 12 shows that in practice the separation of "recurrent" and "development" spending is not being properly applied. The wage bill accounted for the remaining 11 percent of expenditures in the development budget. Table 12: The economic composition of MAAIF's development budget, 2005/06­ 2008/09 (UShs billions) 2005/06 2006/07 2007/08 2008/09 Average Share, (%) Wage bill 6.41 8.30 7.70 5.30 6.93 11.1% Nonwage recurrent 41.29 39.16 45.58 35.30 40.33 64.6% Capital outlays 17.17 11.03 19.41 13.23 15.21 24.3% Development budget 64.87 57.45 69.58 52.33 60.86 100.0% Source: AgPER Task Team estimate, based on MoFPED (various years). 36. The detailed disaggregation of MAAIF's development budget is presented in Table 13. The wage bill makes up 11 percent of the development budget, with 19 allowances dominating. Allowances are difficult to control and target to the right priorities. Attention must be given to this probably unsustainable source of growth in the wage bill. Table 13: The detailed economic composition of MAAIF's development budget, 2005/06­2008/09 (UShs billions) 2005/06 2006/07 2007/08 2008/09 Average Share, (%) Wage bill 6.41 8.30 7.70 5.30 6.93 11.1% Salaries 0.90 0.99 1.16 0.80 0.96 1.5% Allowances 5.51 7.21 6.44 4.50 5.97 9.3% Nonwage recurrent 41.30 39.16 45.58 35.30 40.33 64.6% Goods and services 20.39 18.47 30.63 28.14 24.41 39.1% Medical and veterinary 3.04 1.72 0.80 0.68 1.56 2.5% supplies Capital outlays 17.17 11.03 19.41 13.23 15.21 24.3% Nonresidential buildings 8.16 4.47 10.03 4.75 6.85 11.0% Other structures (markets) n/a 2.51 5.93 5.30 4.58 7.3% Machinery and equipment 2.21 0.25 0.28 0.02 0.70 1.1% Livestock 0.64 0.30 n/a n/a 0.23 0.4% Irrigation n/a 0.28 0.20 0.01 0.12 0.2% Roads and bridges 4.98 2.70 0.90 0.69 2.31 3.7% Land 1.18 n/a 2.00 1.15 1.15 1.8% Development budget 64.87 58.49 72.69 53.83 62.47 100.0% Source: AgPER Task Team estimate, based on MoFPED (various years). 37. The most substantial share of the nonwage recurrent budget was allocated to goods and services. On average, goods and services made up 61 percent of the nonwage recurrent and 39 percent of MAAIF's development budget (Table 13). According to project documents, the category "goods and services" comprises seed, fertilizer, pesticide, and other agricultural inputs.17 In most cases these are hidden input subsidies. Their share grew from 49 percent of the nonwage recurrent budget in 2005/06 to 80 percent in 2008/09. The subsidized inputs were distributed through various projects and increasingly through NAADS. In 2008/09, NAADS received an additional UShs 37 billion "for inputs to the small-scale farmers who cannot afford to purchase the necessary inputs."18 Input delivery has been supply driven, however. Little has been done to strengthen private networks of agro-dealers and improve subsidy targeting (through vouchers, for example). Consequently, as currently implemented, most programs do not meet the requirements for "market-supporting smart subsidies" described in Box 1, especially because they have tended to benefit the wealthiest rural households.19 As allocations for subsidized inputs grow, the fiscal burden and economic distortions 17 The information is, however, insufficient to disaggregate this category into specific input subcategories. 18 The Budget Speech for the FY 2008/09 dated to June 12, 2008. 19 See the incidence analysis of various types of public expenditure in Section 4.6. 20 (wastage, leakage, and the displacement of input sales from the private to the public sector) will grow as well, unless the input delivery mechanism remains unchanged.20 Box 1: What are the conditions for "smart" input subsidies? There are compelling rationales for implementing "market smart" subsidy programs when markets do not function properly. Voucher systems have proven more effective for providing inputs and less likely to distort input markets than the direct subsidies and centralized input procurement and distribution systems used intensively in the past. The benefits of a smart subsidy include increased agricultural output, the promotion of private input markets, and increased adoption of new technologies by poor farmers, all of which ultimately result in sustained poverty reduction. But achieving benefits depends greatly on how the program is designed and implemented. The experience from several African countries, especially Malawi, Tanzania, and Zambia, provides practical guidelines for maximizing the effectiveness of input subsidies in meeting the objectives of improving food security, alleviating hunger, and increasing equity. To be "market smart," input subsidies should: (i) be directed at poor farmers to encourage incremental input use by people would not otherwise use inputs; (ii) not displace existing commercial sales; (iii) use vouchers, matching grants, or other instruments and strengthen existing private distribution systems; and (iv) be introduced for a limited period, with a clear schedule for phasing out once their purpose has been achieved. Source: Morris, M., V.A. Kelly, R.J. Kopicki, and D. Byerlee (2007): Fertilizer Use in African Agriculture: Lessons Learned and Good Practice Guidelines. Washington, DC: World Bank, and World Bank, World Development Report 2008: Agriculture for Development (Washington, DC: World Bank, 2007). 38. The share of capital outlays remained stable but relatively low at 24 percent during the study period. Notably, the capital investment budget was smaller than the budget for goods and services. The largest infrastructure elements are livestock and wholesale markets, rural communal roads and bridges, and machinery and equipment (Table 13). Investments in irrigation to boost agricultural production and reduce farmers' dependence on rainfall remain extremely low. This trend should be reversed to spur agricultural growth in Uganda. 39. What is the economic composition of MAAIF's consolidated recurrent and development expenditures? Table 14 shows wage, nonwage recurrent, and capital expenditures across the "recurrent" and "development" budget. When MAAIF's recurrent budget is added, the share of the wage bill and nonwage recurrent expenses grows to 14 percent and 64 percent, respectively, while the share of capital outlays falls to 22 percent. Figure 7 illustrates the growing trend for nonwage recurrent expenditures and the declining trend for capital expenditures during the period under review. 20 Section 4 presents many examples of low technical efficiency of input and vaccine delivery in selected districts of Uganda. 21 Table 14: Economic composition of MAAIF's budget (excluding grants and domestic arrears), 2005/06­2008/09 (UShs billions) 2005/06 2006/07 2007/08 2008/09 REC DEV TOT REC DEV TOT REC DEV TOT REC DEV TOT Wage bill 2.5 6.4 8.9 2.5 8.3 10.8 2.5 7.7 10.2 2.6 5.3 7.9 Nonwage 1.9 41.3 43.2 2.3 39.2 41.5 4.0 45.6 49.6 4.3 35.3 39.6 current Capital 0.0 17.2 17.2 0.0 11.0 11.0 0.0 19.4 19.4 0.0 13.2 13.2 outlays MAAIF 4.4 64.9 69.3 4.8 58.5 63.3 6.5 72.7 79.2 6.9 53.8 60.7 budget Source: AgPER Task Team estimate, based on MoFPED (various years). Figure 7: Economic composition of MAAIF's budget (excluding grants and domestic arrears), 2005/06­2008/09 (UShs billions) 90 80 Capital outlays 70 60 50 Ushs bill Non-wage recurrent 40 30 20 Wage bill 10 - 2005/06 2006/07 2007/08 2008/09 Source: AgPER Task Team estimate, based on MoFPED (various years). 40. The share of capital outlays becomes even smaller in the context of the recurrent and development budgets of the autonomous organizations. Most of the budget allocations to NARO and NAADS go to pay wages, finance operational expenses, and increasingly to provide goods and inputs through NAADS and other development projects, as reported previously. As a result, capital outlays constitute only 8.5 percent of the agriculture budget, with the wage bill absorbing 31.5 percent and nonwage recurrent expenses 60.1 percent (Figure 8). 22 Figure 8: Economic composition of the agricultural sector budget, 2005/06­ 2008/09 (UShs billions) 250 200 Capital outlays 150 UShs bill Non-wage recurrent 100 50 Wage bill - 2005/06 2006/07 2007/08 2008/09 Note: The sector budget comprises MAAIF, NARO, UCDA, UCDO, NAADS Secretariat, NAADS districts, and district extension. It is assumed that 60 percent of the NARO and NAADS Secretariat development budget is allocated to wages, with the rest allocated to nonwage recurrent expenditures. For NAADS districts, it is assumed that about 30 percent of the development budget goes to wages, with the rest going to nonwage recurrent spending. Source: AgPER Task Team estimate, based on MoFPED (various years). 41. Is the present economic composition of sector expenditure appropriate to enhance pro-poor agricultural growth in Uganda? There are several reasons to believe that the current economic composition of expenditures requires improvement. First, too little capital funding is allocated to construct markets, link remote areas to road networks, rehabilitate and expand irrigation systems, upgrade veterinary and sanitary and phytosanitary laboratories, and improve access to livestock breeds. The share of capital outlays in the sector budget fell to 6 percent in 2008/09 from an already low level of 11 percent in 2005/06. Although there is a significant need for capital to ease the binding constraints in the agricultural sector, as outlined in the Country Economic Memorandum (2007), a greater share of spending is needed to finance capital investments. 42. A second reason for improving the current economic composition is the rapidly rising share of the budget devoted to providing subsidized inputs. Accessible supplies of inputs are integral to raising agricultural productivity, but supply-driven subsidies that do not provide benefits for the poorest21 and do not support the development of private input markets are likely to fail and crowd out much desirable capital investment at the same time. 21 See the incidence analysis of public expenditures in Section 4.6. 23 43. The third problem with the allocative efficiency of the agriculture budget is the unbalanced wage and operational expenditure between MAAIF Headquarters and departments. Currently, 35 percent of MAAIF's recurrent budget goes to MAAIF Headquarters. As mentioned, this large share is partially explained by higher wage allocations for senior staff, high transport costs to Kampala from Entebbe, and other recurrent expenses paid from Headquarters for services used by all MAAIF departments, but it is necessary to find ways to reduce this imbalance in favor of the operational or technical departments whose effectiveness is constrained by the lack of nonwage operational funds. 2.4. Functional composition of the sector budget 44. The second part of the analysis of allocative efficiency is the functional composition of sector expenditure. Public funds should be allocated to obtain the highest social payoffs for growth and poverty reduction. The right mix of spending is important, as demonstrated empirically by Fan et al. (2004) of IFPRI.22 Using district- and household-level data for 1992, 1995, and 1999, the authors estimated the effects of different types of public expenditure on agricultural growth and poverty reduction in Uganda. They found that all types of public spending reduced poverty while increasing agricultural production. They also found that the production and poverty reduction gains differed greatly by type of expenditure. Public spending on agricultural advisory services and research had the highest return to labor productivity and to poverty reduction, followed closely by investments in feeder roads. Investments in education ranked third in terms of effects on productivity and poverty reduction, whereas investments in health had the smallest impact. 45. Another example of how important it is to determine the right mix of expenditures comes from Latin America. Using data from 10 Latin American and Caribbean countries between 1985 and 2000, López (2005) investigated how the mix of public expenditures influenced agricultural per capita income, controlling for trade openness and the share of the nonagricultural sector in per capita GDP.23 The major finding was that the structure or composition of expenditures is more important for per capita agricultural growth than the size of the expenditure. A reallocation of 10 percentage points of public expenditure from subsidies to public goods would increase per capita agricultural income by an average of 2.3 percent, without increasing total expenditure.24 The impacts of expenditure mix are significant, because they capture the positive effects of providing more public goods and reducing the distortions created by subsidies. In contrast, a larger public expenditure (without any change in the composition 22 Fan, S., X. Zhang, and N. Rao (2004): Public Expenditure, Growth, and Poverty Reduction in Rural Uganda. DSG Discussion Paper 4, Washington, DC: International Food Policy Research Institute. 23 López, R. (2005): Why Governments Should Stop Non-Social Subsidies: Measuring the Consequences for Rural Latin America. University of Maryland, College Park, revised version, February 4, 2005. 24 In that study, public goods expenditures included those on technology generation and transfer, soil conservation, sanitary and phytosanitary protection, communications and information services, rural infrastructure, and social services such as education and health. Expenditures on private goods included commodity-specific subsidies, marketing assistance and promotion, subsidized credit, and irrigation. 24 of expenditures) was much less effective in raising per capita agricultural incomes. A 10 percent expansion of government outlays yielded, on average, only 0.6 percent growth in agriculture income. 46. In Uganda, the largest share of expenditures in the agricultural sector is allocated to two areas with high potential for enhancing pro-poor growth: advisory services and agricultural research. As shown in Figure 9, these core public goods received about 57 percent of the approved sector budget from 2005/06 to 2007/08, followed by physical infrastructure (which received 15 percent). Other functions received small allocations--especially plant pest and livestock disease control--despite their tremendous importance for supporting other government programs that distribute cattle and goats to farm households. Figure 9: Functional composition of the agricultural sector budget, 2006/07 Physical infrastructure Promotion Processing & 15% 2% marketing 3% Research 19% Seed capacity development 5% Water capacity building 4% Institutional development 2% Planning & policy 2% Livestock & fish Advisory services regulatory services Plant pest and 38% Livestock diseases 2% diseases 7% 1% Note: Physical infrastructure includes wholesale/livestock markets, village community roads and bridges, irrigation, fish ponds, and other capital expenditure managed by the agricultural ministries, agencies, and local governments. Source: Agricultural Sector Budget Framework Paper 2007/08, Section 4. 47. Identifying unproductive spending is harder in Uganda now than it was in the past, because reforms have eliminated the most egregious instances. The privatization of state-owned enterprises and parastatals means that less public money is spent on products that private enterprises could provide more efficiently. The government also abolished most nationwide subsidies on food and inputs, although subsidized inputs and livestock are increasingly channeled through development projects and NAADS, as indicated previously. 25 48. To identify unproductive public spending on agriculture in Uganda now, it is necessary to focus on whether the composition of public spending is in line with development priorities, whether the economic composition of spending is appropriate, and whether the government is using the right mechanisms (delivery instruments) to do the right things (technical efficiency). The economic composition of public agricultural expenditures was analyzed in Section 2.3, and their technical efficiency will be analyzed in Section 4. Here, the analysis will focus on the extent to which current public spending is aligned with the government's priorities. 49. How well is the functional composition of expenditures aligned with the government's priorities for agriculture? The Government of Uganda describes its priorities for the agricultural sector in the MAAIF Development Strategy and Investment Plan, prepared in 2006. The DSIP presents the justification for a strategy to achieve sustained growth in agricultural productivity and poverty reduction and estimates the costs of high-priority investments to achieve those objectives. The DSIP is supposed to be used to prioritize and define spending plans during the budgetary process each year. Table 15 shows the extent to which MAAIF's annual budgets25 (including expenditures for development, recurrent costs, and district grants, which fall within the MAAIF portfolio) are aligned with projections for the three years covered in the DSIP. Clearly advisory services have been accorded higher priority in practice than planned, whereas allocations to research have remained in line with DSIP projections. Together, advisory services and research receive 59 percent of available funds, compared with the DSIP projection of 48 percent. Both of these subsectors are accorded high priority in the Poverty Eradication Action Plan (PEAP). 25 MAAIF's recurrent budget was allocated on a pro-rata basis, in accordance with the proportion of other funds allocated to each priority area. This approach was used because of the difficulty of allocating recurrent funds to individual priority areas. 26 Table 15: Proportion of MAAIF budget allocated to DSIP priority areas, compared with DSIP projections, 2005/06 to 2007/08 (percent) Budget allocations Average DSIP over the 2005/06 2006/07 2007/08 average period Research 17 19 23 20 19 Advisory services 30 45 41 39 29 Livestock disease 9 7 4 7 6 Plant pests and diseases 1 1 0 1 5 Livestock and fish regulatory services 2 2 2 2 5 Planning and policy 2 2 1 2 1 Institutional development 4 1 0 2 9 Water capacity building 3 4 4 4 10 Seed capacity development 9 3 3 5 8 Processing and marketing 7 2 2 3 3 Physical infrastructure 12 14 18 15 5 Promotion 3 1 1 2 1 Source: MAAIF internal documents and discussions with MAAIF staff; MAAIF (2006). 50. With respect to other priority areas identified in the DSIP, notable discrepancies appear between actual average budget allocations over the three years and the DSIP projections. The most obvious difference is in the high proportion of funds budgeted for physical infrastructure compared with the proportion originally planned in the DSIP--15 percent compared with 5 percent. This discrepancy may be explained partly by delays in implementing the infrastructure component of development projects in previous years and a sudden increase in allocations to this activity in the period under review. Yet whatever the reasons for this discrepancy, the capital expenditure is desirable and should be increased from current low levels in terms of both actual spending and planned spending in the DSIP. 51. Worryingly, the emphasis given in the PEAP to disease control has not been reflected in the proportion of funds allocated to this activity in MAAIF's annual budgets. Plant pest and disease control has received less than 1 percent of total resources, compared with the 5 percent originally planned, while the proportion of funding allocated to livestock disease control has fallen in each of the three years, even if the average is close to that projected in the DSIP. Other significantly underfinanced functions include investments in livestock and fish regulatory services, water capacity building, and institutional development. These are all core public goods and essential investments at Uganda's current stage of development. The private sector will not invest in most of them, but they are important to catalyze private investments in agricultural production and agribusiness. Core public goods are also critical for ensuring that the geographically focused interventions of development projects are sustainable in the sense that local farmers still have access to markets and services after the development projects end. Last but not least, public investments in public goods are necessary to complement other interventions. Section 4 shows how the small and declining budget for controlling livestock diseases, for example, prevents the Veterinary Department from detecting and controlling the most infectious and dangerous diseases in a timely way, with the result 27 that resources have been wasted in promoting vaccines with limited effectiveness. Failure to invest in public goods has also diminished the impact of NAADS by burdening it with an additional function--input distribution--which lies outside its mandate to provide advisory services. 52. Finally, the share of budget allocated to each priority area varies considerably from one year to the next. The approved budgets are less evenly balanced than the planned budgets presented in the DSIP, with a greater concentration of resources on fewer priorities: Five priority areas received 86 percent of MAAIF budget allocations over the three-year period, compared with 75 percent in the DSIP. Based on this information, it may be concluded that the DSIP has not been used to draw up subsector budgets, which detracts from the value of the document and the effort involved in its preparation. Preparation of the second DSIP, covering 2008/09­2010/11, should include an analysis of future budget requirements by program, based on a realistic assessment of the resources that will be available. The objectives and work plan for each priority area should also clearly define its functions, specify budgets for these functions, and ensure that these functions are fully aligned with national development strategies and programs, such as PEAP and PMA. 28 3. Budget Process and Performance 53. The effectiveness of public expenditures depends, among other things, on budget preparation and implementation. Parliamentary approval of the sector budget is no guarantee that all approved funds will be released and spent. If funds are released at unpredictable times, or if their release is concentrated in the last quarter of the financial year, it is difficult to implement programs smoothly and procure goods and services at the lowest cost. In other words, poor budget preparation and performance can undermine even the best allocation of funds. 54. This section explains the budget preparation process at the central and local level, analyzes budget execution and control, and describes how budget performance is monitored. It also provides recommendations for improving budget preparation, performance, monitoring, and evaluation. 3.1. Budget preparation 55. The main stages in formulating the national budget are described in Box 2. For agriculture, the process, at least in principle, lasts nine months--from the national consultative meeting scheduled in October, at which MTEF ceilings are announced, to the reading of the Budget in June the following year. Box 2: Steps in preparing the national budget October: Draft Budget Ceilings x MoFPED distributes the Budget Call Circular (BCC) to all ministries and agencies with inter- and intrasector MTEF allocations. x MoFPED hosts a "Budget (Framework) Consultative Workshop." November­December: Preparation of Sector Working Group Reports x Sector Working Groups use indicative budget ceilings to arrive at intersector allocations and prepare Sector Budget Framework Paper (SBFP). January: Preliminary Estimates x SBFPs discussed with MoFPED during ministerial consultations. x Ministries and agencies prepare draft budget estimates on this basis. March: National BFP to Cabinet and Parliament x MoFPED compiles SBFPs into a National Budget Framework Paper (NBFP), presented to the Cabinet. x When the NBFP is considered and approved, it is submitted to Parliament. April­May: Parliament and Public Expenditure Review x The Budget Committee of Parliament discusses the NBFP and presents recommendations to the President and MoFPED. x The national PER meeting is held, at which the NBFP is discussed. June: Finalization of Budget x On the basis of Parliamentary/PER recommendations, the proposed Budget and MTEF is amended. x The Budget is read. Source: Adapted from Williamson, T. (2003): Targets and Results in Public Expenditure Management: Uganda Case Study. ODI Working Paper 205, London: Overseas Development Institute. 29 56. The MTEF should provide a reliable, rolling three-year guide to sector and subsector budget allocations. In practice, from year to year major changes are made to the MTEF ceilings in total and in the allocations to individual Sector Votes. Such changes make medium-term planning difficult to implement and undermine the predictability of funding in the two later years. Table 16 shows the agricultural sector MTEF and budget ceiling for 2007/08 as stated in annual National Budget Framework Papers, Budget Speeches, and Approved Budget Estimates, from the time it was first announced in the Budget Framework Paper in March 2005 to the Budget Framework Paper in March 2007. Table 16 shows the substantial variation in the planning figure and the individual sector components. Table 16: Evolution of 2007/08 budget ceiling projections for agriculture, March 2005 to March 2007 (UShs billions) Recurrent Development Date Document Wage Nonwage Domestic Donor Total National Budget March 2005 6.78 29.15 137.76 173.69 Framework Paper June 2005 Budget Speech 6.52 21.74 70.22 131.44 229.92 Approved Estimates for 3rd quarter, 2005 6.58 29.15 133.76 1.94 171.43 2005/06 National Budget March 2006 6.03 31.71 53.43 85.22 176.39 Framework Paper Approved Estimates for 3rd quarter, 2006 6.52 23.24 70.22 131.44 231.42 2006/07 December 2006 Budget Call Circular 6.29 20.84 71.75 53.65 152.53 National Budget March 2007 6.29 18.14 78.75 81.68 184.86 Framework Paper Note: The "development" figure in the 2005 National BFP is an "integrated ceiling" combining domestic and donor funds. MoFPED has not maintained this practice. 57. There are also frequent modifications to the MTEF and budget ceilings as sectors prepare their Budget Framework Papers and later when the proposed budget is considered at the National Budget Workshop and by the Cabinet and Parliament. Table 17 illustrates changes to budget ceilings for the various agriculture Votes when the 2006/07 budget was under preparation. 58. The earlier adjustments reflect revised estimates of domestic revenue and donor commitments and revised national policy priorities (such as energy generation and hosting the Commonwealth Heads of Government Meeting) expected to affect national and sector budget ceilings. Later adjustments are frequently related to new policy initiatives and political directives. They tend to be accommodated by adjustments in sector and subsector allocations. 30 Table 17: Evolution of the 2006/07 budget ceiling projections for agriculture by Vote (UShs billions) RECURRENT DEVELOPMENT TOTAL Vote Agency Date Exclud- Includ- Wage Nonwage GoU Donor ing donor ing donor 05/06 Approved Estimate 2.49 4.93 34.23 38.08 41.65 79.73 Nov. 05 MTEF 2.37 6.79 26.28 60.18 35.45 95.63 010 MAAIF Apr. 06 NBFP 2.38 5.53 8.86 57.74 16.08 73.81 Jun. 06 Budget Speech 2.41 5.53 7.08 48.88 15.03 63.91 05/06 Approved Estimate 13.36 7.4 6.31 20.76 27.07 Nov. 05 MTEF 10.58 12.2 5.53 22.78 28.32 142 NARO Apr. 06 NBFP 2.84 9.2 5.31 12.04 17.35 Jun. 06 Budget Speech 2.84 17.2 5.35 20.04 25.39 05/06 Approved Estimate 3.21 3.2 6.41 6.41 District Nov. 05 MTEF 3.21 3.15 6.36 6.36 501­ Exten- 580 Apr. 06 NBFP 3.08 2.87 5.95 5.95 sion Jun. 06 Budget Speech 3.88 3.16 7.04 7.04 05/06 Approved Estimate 27.1 2.76 27.1 29.85 501­ NAADS Nov. 05 MTEF 34.75 2.56 34.75 37.31 580 Districts Apr. 06 NBFP 24.75 2.47 24.75 27.21 Jun. 06 Budget Speech 37.13 1.53 37.13 38.66 05/06 Approved Estimate 3.77 2.44 6.21 6.21 NAADS Nov. 05 MTEF 3.77 2.44 6.21 6.21 152 Secre- tariat Apr. 06 NBFP 3.77 8.42 12.19 12.19 Jun. 06 Budget Speech 4.46 5.35 9.81 9.81 05/06 Approved Estimate Nov. 05 MTEF 155 UCDO Apr. 06 NBFP Jun. 06 Budget Speech 1.2 1.2 1.2 05/06 Approved Estimate Nov. 05 MTEF 160 UCDA Apr. 06 NBFP Jun. 06 Budget Speech 0.58 0.58 0.58 05/06 Approved Estimate 5.7 25.26 71.17 47.15 102.13 149.27 TOTAL Nov. 05 MTEF 5.58 24.29 75.67 68.27 105.55 173.83 Apr. 06 NBFP 5.46 15.01 51.23 65.52 71.01 136.51 Jun. 06 Budget Speech 6.29 17.77 66.76 55.76 90.83 146.59 Source: SBFP; NBFP; Budget Speech. 31 59. The first stage in the annual budget process for agriculture is for the Agriculture Sector Working Group to prepare the Agriculture Sector Budget Framework Paper (SBFP). Next, the Budget Call Circular is distributed and the National Consultative Workshop is held. At the Workshop, the main budget priorities and MTEF ceilings are announced at the sector and Vote level. The proposed terms of reference, composition, and method of work for the Sector Working Group are set forth in the Budget Call Circular for 2005/06 (Box 3). The Permanent Secretary of MAAIF chairs the Sector Working Group, which meets soon after the Budget Call Circular is distributed in October and perhaps once or twice in December when the Budget Framework Paper is prepared. Until recently, the Working Group met rarely, if at all, outside of this October­December period. 60. The SBFP is drafted by the Secretariat of the Sector Working Group (MAAIF's Agricultural Planning Department). In some years, a small "technical drafting committee," consisting of staff from several departments in MAAIF and MoFPED, has been appointed to draft the Budget Framework Paper. The bulk of the task, however, falls on the shoulders of the Agricultural Planning Department. Inputs to the Budget Framework Paper are provided by each of the technical units in the Ministry and its semiautonomous agencies; representatives of these units are invited to attend meetings of the Sector Working Group. The final draft of the SBFP is presented at a meeting of the Ministry's Top Policy Management group,26 in which alterations to the document and final changes to the proposed budget allocations are agreed. 61. Since the enactment of the 2001 Budget Act, all SBFPs have to be submitted to MoFPED not later than December 31; there is no leeway for slippage, as was the case in earlier years.27 MAAIF does not begin preparing the Budget Framework Paper until receiving the Budget Call Circular and the National Consultative Workshop has been held. Delays in either or both of these events further curtail preparation time. As more and more staff members take their annual leave as December progresses, considerable pressure is placed on a very small number of the staff. MAAIF could ease this problem by starting to prepare its budget earlier. Much documentation could be prepared in advance, particularly the current year's performance evaluation28 and outturn and cases for new initiatives, even in the absence of MTEF and budget ceilings for the next year. MoFPED could assist by releasing the Budget Call Circular at a set time each year, even if the National Consultative Workshop is delayed. 62. The SBFP's format and content are dictated largely by the Budget Call Circular. The guidance provided in the Circular would be improved, however, by assessing whether the information it requires is essential and presented in the most effective format for making informed decisions on budget allocations. Except for 2006/07, MoFPED has received no feedback or suggestions from the Sector Working 26 Chaired by the Minister of MAAIF. 27 Although in 2006/07 delays by MoFPED in sending out the Budget Call Circular caused the date of submission to be fixed for January 15, 2007. 28 Much of this material is also required for the quarterly and annual Budget Performance Reports (see later). 32 Group (or MAAIF) on improving the content and quality of the SBFP. It is recommended that this kind of feedback should become a regular feature of the budget preparation process. Other recommendations for improving SBFP guidelines are summarized in Annex 1. Box 3: The Agriculture Sector Working Group Proposed terms of reference for the Working Group x Review sector strategies and investment programs. x Review and recommend projects for submission to the PMA Secretariat and Development Committee, in line with sector plans. x Prepare Agricultural Sector Budget Framework Paper (SBFP) as a basis for compiling the sector's annual budget and for enabling MTEF to evolve over time. x Provide the main forum for the sector-wide approach to planning and budgeting for agriculture. x On the basis of sector expenditure and investment plans and SBFP, identify policy issues for consideration and action by Ministry Top Policy Management. x Provide information for the Joint Government of Uganda Donor Reviews. x Monitor budget implementation with respect to the aims and objectives set forth in the BFP. Composition of the Working Group x MAAIF (including PMA Secretariat). x MAAIF Agencies (NARO, NAADS, UCDA, UCDO, DDA, Animal Genetic Resource Centre, COCTU). x Donor Sub-Group on Agriculture. x MoFPED. x Representation from private sector. x Representation from nongovernmental organizations (NGO Forum). x Representation from civil society. x Representation from local governments (ULAA). x Representation from farmers' associations and the Agricultural Council of Uganda. Proposed method of work for the Working Group The Working Group will carry out its activities through meetings, retreats, and workshops as well as consultative visits to local governments, as appropriate, as follows. x During the SBFP preparation process: - At least once immediately following the conclusion of the National Budget Conference. - At least once during November to review the draft SBFP. - At least once during December to review and approve the final SBFP. x Quarterly after the SBFP is finalized. x Extraordinary/emergency meetings as necessary. Financing of the Sector Working Group Funding for the Working Group's activities should be budgeted from within the sector MTEF ceilings. Source: MoFPED (2005). 63. The credibility of the DSIP--MAAIF's strategy and investment plan-- should be strengthened. Its credibility can be improved by paying greater attention to: (i) the criteria used for prioritization; (ii) the expected outcomes; (iii) detailed explanations of expenditure estimates; and (iv) linking investment plans more closely to anticipated MTEF ceilings and indicating how the plans would change if MTEF ceilings were increased or decreased. It is urgent for MAAIF to update its strategy and investment 33 plan for 2008/09 and beyond to continue providing a link between policy, planning, budget preparation, and negotiations. MAAIF should consider whether a three-year or five-year investment plan is best suited to its purposes. A five-year plan would provide a longer time horizon for investment decisions and reduce the perceived need for frequent revisions, but changing priorities and events might reduce the reliability of estimates for the later years. 64. How is the budget prepared at the local government level? A single Budget Framework Paper is prepared for each district by the District Technical Planning Committee, with contributions from the District Production Department. In principle, the District Budget Framework Paper essentially coordinates plans submitted by lower levels of government and developed in a participatory manner. In practice this participatory process is weak, rarely involving more than a minority of the local population. Moreover, many subcounties lack the capacity to prepare such plans. 65. There is only a short period between the Regional Local Government Budget Framework Paper Workshop, when Indicative Planning Figures are announced, and the date for submitting the District Budget Framework Paper. In 2005, for example, the Regional Workshop for Mukono District was held on December 19­20 and the Budget Framework Paper had to be submitted by January 20, 2006. The MoFPED and line ministries comment on the Budget Framework Paper in March­April, and then it is examined by various District Council committees before being read and approved by the full District Council in June. As at the national level, the Indicative Planning Figures for individual districts are frequently adjusted as the budget is prepared. 66. At the district level, funds for production activities come from various sources. These sources include the NAADS District Grant, the District Agriculture Extension Grant, and the parts of the Non-Sectoral Conditional Grant (NSCG) and the Local Government Development Program (LGDP) that are apportioned to production activities, together with any projects funded from local tax revenue. In some districts, the Production Department receives extra funding from the 10 percent pooled from conditional grants under the Fiscal Decentralization Strategy. Each grant channeled from the Central Government is subject to various conditions. Decisions on how the NAADS District Fund is allocated for the district and subcounties are predetermined by the NAADS Secretariat. The district only distributes funds as scheduled by the Secretariat. Decisions on the actual extension activities to be financed by the grant are made by farmers in fora at the subcounty level. 67. In principle, the subcounties are supposed to decide how 65 percent of NSCG funds should be used. The district's role is to check conformity with the guidelines (for example, to ensure that proposed investments are of the acceptable "public good" type). In practice, many decisions on project selection are made by the District Technical Committee, based on recommendations from various Departments. The Production Department's scope for flexibility in planning is limited to decisions on allocating funds from the NSCG, the LGDP, the allocation from the 10 percent flexibility fund under the Fiscal Decentralization Strategy, and locally financed projects. 34 68. Local political pressures cause more plans to be developed than can be realistically financed, so the plans inflate the projected level of local revenue. Failure to implement projects is then blamed on shortfalls in local revenue collection. Aside from funds that are controlled directly by local governments, many production activities in the districts are supported by off-budget projects financed by donors and/or NGOs. In some districts, this is an important source of finance, but at the moment it does not appear to be taken fully into account by districts or central government agencies in planning the allocation of their own funds. 3.2. Budget execution 69. The relationship between planned (approved) and actual (released) expenditure is an indicator of the effectiveness of the budget in allowing departments and programs to plan their activities and deliver public services for the year, as expressed in policy statements, output commitments, and work plans. Typically, in assessing aggregate budget performance, the original approved budget allocations are taken as the measure of "planned" expenditure" as, in principle, this should be the basis used by ministries, departments, and agencies in deciding on their programs of work. 70. During the period under review, the released budgets deviated significantly from the approved budgets and were typically lower than planned (recall Figure 2). These deviations can arise from at least four sources: x The extent to which budget allocations misjudged requirements in the first place. x Readjustments in budget allocations after the "approved" allocations have been announced, owing to: The need for MoFPED to accommodate fresh calls on the budget from other Votes. Effective negotiations to increase the ministries' and agencies' budgets in line with perceived requirements. x Budgeted funds ("approved" or "revised") are not released due to revenue shortfalls or further unforeseen calls on available funds. x The untimely release of funds required for seasonally determined uses. 71. The outturns (deviations) differed for recurrent and development expenditure. Usually outturns were larger for development expenditures financed by donors. In the case of recurrent expenditures, Table 18 shows the approved and revised recurrent budget allocations for MAAIF and released recurrent expenditures for 2000/01­ 2005/06. Released expenditures were lower than the approved budget allocations in three of the six years, but higher in the other three. Recurrent budget allocations were revised upwards from approved allocations in all years, but--with the exception of 2003/04-- released expenditures were lower than revised budget allocations. 35 Table 18: Approved, revised, and released recurrent expenditures for MAAIF, 2000/01­2005/06 (UShs billions) 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 Approved 4.71 7.36 6.01 4.88 9.54 11.41 Revised 4.73 7.61 6.34 5.89 12.19 12.19 Released 4.01 6.05 6.22 7.83 11.32 10.31 Source: MoFPED Approved Budget Estimates and Auditor General's Reports. 72. Table 19 examines the released recurrent expenditure for MAAIF programs (departments) as a percentage of the approved recurrent allocations. Variations of more than 15 percent between approved and released allocations are shaded. In three of the past six years, the total released recurrent expenditures have varied by more than 15 percent from the total approved recurrent allocations. In two years (2003/4 and 2004/05), total released and total approved expenditures differed because the released allocations for Headquarters and the Crop Protection Department were more than double the approved amount. In 2001/02, the released expenditure was well below the approved budget allocation, with most programs showing large downward deviations from approved budget expenditures. Table 19: Released recurrent expenditures as a percentage of approved recurrent budget allocations for MAAIF programs, 2000/01­2005/06 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 01 Headquarters 88.1 80.2 122.1 241.7 206.7 81.4 02 Directorate of Crop Resources 83.9 77.0 94.4 85.0 96.2 75.8 03 Farm Development Department 89.2 78.1 100.2 93.7 103.4 97.1 04 Crop Protection Department 75.7 113.2 87.6 235.3 219.0 91.0 05 Crop Production Department 83.6 74.2 80.2 79.0 99.7 97.2 06 Directorate of Animal Resources 92.1 93.9 95.1 91.0 97.6 100.0 07 Animal Production Department 90.7 74.1 90.6 89.0 98.6 100.0 08 Livestock Health and Entomology 78.9 73.4 84.4 84.7 96.5 100.3 09 Fisheries Resources Department 82.5 89.8 77.7 91.0 94.0 74.9 10 Department of Planning 82.9 72.5 84.1 84.0 95.8 96.5 Weighted average 85.2 82.1 103.5 160.5 118.7 90.4 Source: MoFPED Approved Budget Estimates and Auditor General's Reports. 73. Do any revised expenditures more closely match released recurrent allocations? Some do, although most do not. There was a closer match in Crop Protection in 2003/04 and 2004/05, indicating that the budget allocation was revised to include recruitment of crop inspectors (Table 20). But Headquarters massively overspent in 2003/04 and then underspent in 2004/05 and 2005/06. In fact, the most noticeable feature of Table 20 is that virtually all programs underspend every year. It remains unclear whether this behavior is voluntary or an obligatory response to shortfalls in MoFPED releases as a result of the cash budget system and unforeseen calls on the aggregate budget. 36 Table 20: Released recurrent expenditure as a percentage of revised recurrent budget allocations for MAAIF programs, 2000/01­2005/06 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 01 Headquarters 87.3 79.7 110.4 212.9 84.1 68.6 02 Directorate of Crop Resources 83.9 77.0 94.4 85.0 93.0 75.8 03 Farm Development Department 89.2 78.1 91.7 93.7 96.7 97.1 04 Crop Protection Department 72.9 81.5 89.3 94.5 95.2 91.0 05 Crop Production Department 83.6 69.4 80.6 79.0 97.0 97.2 06 Directorate of Animal Resources 92.1 93.9 95.1 91.0 90.3 101.8 07 Animal Production Department 90.7 75.5 90.6 99.9 92.2 102.5 08 Livestock Health and Entomology 80.7 72.9 84.1 84.7 96.8 94.7 09 Fisheries Resources Department 82.5 89.8 77.4 91.0 99.1 74.9 10 Department of Planning 86.0 72.5 84.1 84.0 96.9 96.5 Weighted average 84.8 79.5 98.1 133.0 92.9 84.6 Source: MoFPED Approved Budget Estimates and Auditor General's Reports. 74. The shortfall was greater in development expenditures than in recurrent expenditures. According to Table 21, the average shortfall of the development allocation was 44.3 percent, compared with only 4 percent for the recurrent allocation, even after excluding the 60 percent overspending in 2003/04 (see Table 19). There is also a noticeable difference in outturn for donor-financed development funds allocated to MAAIF and for funds financed by the Government of Uganda. As illustrated in Table 21, over the period under review, donor-financed development disbursements/releases were 57.7 percent below the budget allocation, compared with a shortfall of 19.5 percent for government-financed development funds. Table 21: Disbursement/release of MAAIF of donor- and government-financed development funds as a percentage deviation from development budget allocations, 2001/02­2005/06 2001/02 2002/03 2003/04 2004/05 2005/06 Average Donor ­64.3 ­43.4 ­49.0 ­66.9 ­65.6 ­57.7 GoU ­23.5 ­21.1 ­10.0 ­26.5 ­17.1 ­19.5 Total ­48.4 ­35.4 ­35.1 ­57.9 ­47.2 ­44.3 Source: MoFPED Aid Liaison Department databank. 75. The situation has not improved in recent years. Individual development projects provided information on approved and released budgets at the project level. Table 22 shows the outturn of the development expenditure across the projects that made up 64 percent of MAAIF's development expenditure during 2005/06­2007/08. The underperformance of donor-financed funds significantly exceeded that of government- financed funds. Reasons for the shortfalls in government funds were described earlier. 37 Implementation delays, procurement problems, and weak coordination of central and local governments are the major reasons for shortfalls in donor funds.29 Table 22: Approved and released development budgets in the selected MAAIF development projects, 2005/06­2007/08 2005/06 2006/07 2007/08 North-West Smallholder Agriculture Development Project Donor financing (US$ millions) Approved budget 6.31 3.52 1.73 Released budget 3.05 3.00 14.3 Share of released budget in approved budget (%) 48% 85% 829% GoU financing (UShs millions) Approved budget 472.0 156.0 156.0 Released budget 406.9 130.0 127.3 Share of released budget in approved budget (%) 86% 83% 82% National Livestock Productivity Improvement Project Donor financing (US$ millions) Approved budget 6.49 5.35 7.93 Released budget 1.64 8.38 3.13 Share of released budget in approved budget (%) 25% 157% 39% GoU financing (UShs millions) Approved budget 824.0 500.0 450.0 Released budget 448.7 477.0 408.7 Share of released budget in approved budget (%) 54% 95% 91% Fishery Development Project Donor financing (US$ millions) Approved budget 3.25 5.28 8.96 Released budget 0.47 2.94 1.36 Share of released budget in approved budget (%) 14% 56% 15% GoU financing (UShs millions) Approved budget 1,356.6 425.0 395.0 Released budget 657.1 500.0 260.0 Share of released budget in approved budget (%) 48% 118% 66% Vegetable Oil Development Project Donor financing (US$ millions) Approved budget 5.42 4.09 4.33 Released budget 0.72 1.63 0.76 Share of released budget in approved budget (%) 13% 40% 18% GoU financing (UShs millions) Approved budget 1,986.4 250.0 2,250.0 Released budget 732.3 291.0 1,915.1 Share of released budget in approved budget (%) 37% 116% 85% Source: MoFPED Approved Budgets of the development projects (various years). 76. In addition to the shortfalls in allocations, the effectiveness of the public expenditures was weakened by uncertainty in seasonal cash flows. The development budget typically is released in four equal installments in the first month of each quarter. 29 These reasons are examined and presented in detail in Section 4. 38 However, the cash budget system restricts quarterly fund releases to cash flow availability. While this practice helps to ensure macroeconomic stability, it may destabilize activities at the central and local government levels. Because wage payments and Poverty Action Fund-related expenditures have first call on available funds, nonwage recurrent expenditures and government development expenditures on non-PAF activities tend to be affected the most. Figure 10: NAADS (Districts) budgeted and actual cash flows, 2005/06 (UShs billions) 1,000,000 900,000 Budgeted cashflow Actually released 800,000 700,000 600,000 UShs mill 500,000 400,000 300,000 200,000 100,000 - ry y ay ly er r r t ch ne r il us be be be ar pr Ju ob ua M ar Ju ug nu em em em A M br ct A Ja O ov ec pt Fe Se D N Source: NAADS Secretariat (2007). 77. Cash flow within NAADS provides a good example of how seasonal requirements can affect expenditures. NAADS District expenditure is classified as a PAF activity and, as such, releases made on a regular quarterly basis are usually 98­99 percent of budget allocations. The seasonal nature of many agricultural operations, however, means that NAADS District expenditures peak in the first and third quarters of the financial year, when crops are planted (see Figure 10 above). One effect of cash budget ceilings and limited cash releases to MAAIF is that some anticipated donor project expenditures cannot be undertaken. At the national level, it used to be the case that some donor projects required the government to provide counterpart funds to pay for various items before donor funds were disbursed. This restriction appears to have been relaxed by most, if not all, development partners.30 It is necessary that, within the overall cash flow available, MoFPED should cater for the relatively small but particular cash flow requirements of agencies such as NAADS to improve their operational effectiveness. 30 Typically, however, the donor will refuse to reimburse the project for any items that should have been funded by the Government of Uganda. 39 3.3. Budget performance monitoring 78. MAAIF's Policy Department has a Monitoring and Evaluation (M&E) Division staffed by six officers, supplemented by ten others from technical departments, who are required to monitor and evaluate all MAAIF activities. These officials monitor each department, and each one is also responsible for the M&E of a development project. The M&E Division is also supposed to undertake field visits to monitor activities at the district level, but insufficient funds have been allocated to do this. It is planned to pool the M&E funds now available to each individual project to increase the funds available to the M&E Division. 79. As noted earlier, lists of outputs, indicators, and activities are now produced for the SBFP. Similar information is provided in the quarterly Budget Performance Reports that all Ministers responsible for a Vote in Appropriation-in-Aid are required to make to Parliament under the Budget Act (2001). These reports are expected to explain how funds have been spent and to reflect specific data on value for money from the expenditure involved. The M&E Division of the Policy Department also prepares an annual Performance Monitoring and Evaluation Report. 80. At the moment no use appears to be made of these various reports within or outside MAAIF to reward good performance (or invoke sanctions for poor performance), to measure the effectiveness of MAAIF's public expenditure, or to reallocate resources between priority areas. Furthermore, no feedback appears to be provided to MAAIF on the quality or usefulness of any of these reports. 81. Within MAAIF, the Sector Working Group would be the appropriate body to carry out regular quarterly appraisals of the effectiveness of budget expenditures across MAAIF and its agencies. The terms of reference for the Sector Working Group already include responsibility to monitor the performance of budget expenditures. Implementing this responsibility would give the Sector Working Group a year-round agenda and help to increase its effectiveness. Yet such a system can be instituted effectively only if it receives strong ministerial support and is seen to be implemented and supported by the Permanent Secretary and senior colleagues. It would also require, and lead to, substantial improvement in the quality of the reporting system, with a critical examination of the reasons for variations between planned and actual outputs and attention paid to the cost-effectiveness of different programs and resource allocations. MoFPED could also cause performance monitoring to be taken more seriously in MAAIF by insisting on a more rigorous analysis and presentation of sector performance in the SBFPs. 82. The monitoring and especially the evaluation of development projects also require more attention. Despite the fact that most development projects are financed by donors, impact evaluations are rare. Even if an impact evaluation is done, it presents outputs rather than outcomes, yielding little information about the value for money of the investments and few lessons, if any, to prepare new projects. 40 83. At the local level, on paper, performance monitoring is much more rigorous at the district level than in MAAIF. Each department prepares a quarterly work plan with quantifiable targets and also a quarterly plan and budget implementation report. This information is reviewed by the District Council to measure the effectiveness of each department, with an emphasis on physical and political accountability and value for money. Different types of reviews and monitoring are carried out, including: x Quarterly reviews for sectors receiving PAF funding. x Annual reviews as part of the Budget Framework Paper preparation process. x Annual reviews undertaken after the end of the financial year. x LGDP annual assessment of minimum conditions and performance measures. 84. In theory, quarterly monitoring and review is expected to be done by the central government (for activities under PAF) and by the districts using the monitoring and accountability grant. However, funding shortages sometimes constrain this exercise. At the subcounty level, for example, operational budgets are highly constrained, so field supervision is often not undertaken, and the contribution that subcounties are required to make to NAADS funding, for example, is often not possible. 85. Monitoring and evaluation constitute one of the principal roles of the Directorate of Production and Marketing, while politicians also use the funds of the Directorate (including those of the NSCG) to monitor the progress of projects in the district. The preparation of the Budget Framework Paper involves a review not only of the year for which the budget is under implementation but also the previous two to three years. The down-side to this exercise is that neither MoFPED nor MAAIF provides feedback to the districts. 86. At the end of every financial year, districts carry out an annual review with the Annual Report as the main output. The quality of the reports varies from district to district. Poor reports list activities that were carried out with little or no quantification of results. Some reports, however, such as the report for Bushenyi during 2004/05, compare targets with actual achievements, explain why there were deviations from planned targets, and make recommendations for remedial action. 87. Finally, the annual assessment of minimum conditions and performance measures is another form of review that is undertaken in local governments. The assessment results in rewards (in the form of increased funding) for good performance and sanctions (reduction in the LGDP grant) for poor performance. The assessment does not cover production sector indicators, however, so it cannot be used to improve performance in the sector. 41 4. Efficiency and Effectiveness of Agricultural Sector Expenditures 88. A review of how a sector spends its resources and what it gets for those resources--in other words, value for money--is very important for a low-income country with limited resources and high aid flows. Such an analysis is not straightforward. It requires careful analytical work and systematic collection of key data over a certain period. The discussion below presents the process and findings from such a review of the largest agricultural development projects between 2005/06 and 2007/08. 89. The review of public expenditures is more complex for agriculture than for other sectors. In the education sector, for instance, almost all resources follow the same path from the central budget to schools: Inputs such as teachers, classrooms, and textbooks are "bought," and outputs such as number of graduates with a minimum level of proficiency are "produced." In contrast, agricultural projects vary in how they are implemented and how resources, financial and physical, flow between various layers of government. This is the first difficulty. The second difficulty arises from the absence of consistently collected data on sector outcomes, except for overall agricultural production and agricultural value added. Inputs cannot be linked easily with outputs/outcomes because of the time lag in agricultural production, multiple outputs/outcomes, and attribution problems.31 Impact evaluation is the most appropriate tool for examining those issues, and such an evaluation was undertaken for the NAADS program, which accounts for the largest share of Uganda's agricultural sector budget. Presented later in this section, the NAADS impact analysis (conducted by IFPRI, the NAADS Secretariat, and other institutions) is an example of how these evaluations can be done.32 This section also presents a technical efficiency analysis of public spending on agriculture for a range of other development projects. Although it is not as comprehensive as the impact evaluation done specifically for NAADS, the technical efficiency analysis does identify important constraints on project implementation and offer potential remedies. 90. This section begins with a discussion of the methodology and the analytical approach adopted. Public Expenditure Tracking Surveys (PETS) have been used in Uganda since 1996 for the health and education sectors, but this report present the first known application of PETS in agriculture in Uganda. 33 After a brief description of the 31 Attribution is an issue in all sectors, but agriculture in Sub-Saharan Africa is especially affected by numerous exogenous factors, such as highly variable weather and changes in terms of trade, which make it difficult, for example, to link trends in overall agricultural growth with either particular or total public expenditures. Agricultural output may decrease following a drought even if public expenditure rises, and we may erroneously identify a negative relationship between spending and agricultural growth. Econometric techniques must be used to answer this question, but the lack of consistent time-series data in African agriculture makes such estimates very approximate, even if they are possible. 32 Benin, S., E. Nkonya, G. Okecho, J. Rdriamamonjy, E. Kato, G. Lubade, M. Kyotalimye, and F. Byekwaso (2008): Impact Evaluation and Returns to Investment of the NAADS Program in Uganda. October 27, 2008. 33 Uganda first used PETS in 1996 to explain poor performance and identify leakages in the education sector. Since then, PETS has become the standard tool for measuring the effectiveness of public sector 42 sampling strategy and typical implementation arrangements for MAAIF development projects, the report identifies major cross-cutting and project-specific problems with the implementation of development projects. The analysis distinguishes between problems that are within and beyond the control of MAAIF and local governments. The impact evaluation for NAADS follows. The concluding section reviews the main findings and discusses policy recommendations. 4.1. Methodology and analytical approach 91. Step 1: Define the result chain suitable for PETS in agriculture. To improve efficiency in agriculture--in other words, to get more outputs and outcomes without spending more--one needs to define "inputs," "outputs," and "outcomes." Figure 11 shows how these terms can be used for agriculture and how they are related. Figure 11: Relationship between inputs, outputs, and outcomes in agriculture Source: AgPER Task Team. 92. Outcomes are what stakeholders should be interested in achieving, inputs are the means to getting there, and outputs are milestones along the way. The evaluation delivery and corruption in developing countries. The stock-taking study of PETS in Sub-Saharan Africa, prepared by Bernard Gauthier in September 2006, provides detailed information, including best practices and proposals for future surveys. 43 of outcomes of most agricultural public expenditure is critically important but also difficult, and generally such information is not available in Uganda. There is a time lag in agricultural production, and it is challenging to link specific inputs directly to marginal value of product. For example, when a vendor is contracted and undertakes civil work (inputs) to construct a wholesale market (output), it is important to have information about how traders/farmers use the market and the market's effect on transaction costs (outcomes). When farmers use cattle vaccine (input), they expect that cattle will be healthier (output) and that healthier cattle will be more productive (outcome). When farmers apply fertilizers provided by a project, they expect crop yield to rise (output) and farm income to increase (outcome). Each of these result chains requires a rigorous impact evaluation comparing targeted and nontargeted households, and the evaluation must cover a sufficient timeframe to allow outcomes to be realized and properly documented. 93. This kind of rigorous impact evaluation is presented for the NAADS program later in this report. Similar detailed evaluations of additional programs were not feasible, given the time and resources available. Instead, the PETS methodology was used to identify sources of inefficiency in the delivery of public services. The PETS methodology was complemented by focus group discussions with project beneficiaries, service providers, agricultural staff in the local government and MAAIF, project management unit staff, and political leaders to obtain qualitative feedback on service quality and project performance. This combination of quantitative (PETS) and qualitative (focus group) information provides an understanding of the factors affecting the quality of service provision and makes it possible to evaluate the operational efficiency of project implementation. For evaluating NAADS, the quantitative results of the impact evaluation were combined with qualitative assessments in selected districts. 94. Step 2: Set up the operational framework for analyzing technical efficiency. For operational purposes, it is important to distinguish between waste, leakage, and inefficiency.34 Waste and leakage can be dealt with in a different way from inefficiency--for example, by imposing stronger control systems, developing better incentives against malpractice, and improving accountability. Public service inefficiency, which stems from poor choices related to policy, technology, or management, is harder to detect, with assessment demanding more knowledge of the service in question. 95. Inefficiency can arise because changing circumstances could cause the optimal choice or combination of inputs or the appropriate type of output to change over time. For example, a government can be doing the right thing but using the wrong mechanism (delivery instrument)--resulting in inefficiency. In contrast, waste occurs when a completely unnecessary and avoidable cost is borne by the public sector. Waste cannot be defended with integrity. It can arise from many sources, such as weaknesses in the system, weak capacity, low levels of management accountability, improper planning, and corruption. Examples of waste include: 34 This AgPER uses the framework for technical efficiency analysis proposed in the Uganda Public Expenditure Review (2007), World Bank, Washington, DC. 44 x Duplication of administrative functions, either between MAAIF and local governments or between MAAIF and other ministries/agencies. x Unnecessary delays and contract disputes in project implementation, which lead to cost overruns. x Improper appraisal and feasibility work, which lead to delays and cost overruns. x Poor asset maintenance, which leads to early replacement of physical capital such as irrigation, market, and road infrastructure. 96. Leakages occur when released funds are not spent on the inputs for which they were intended. One way to look for leakage is to track the flow of financial and physical resources. This is the methodology used in PETS, which tracks the flow of funds between different layers of government. For instance, the procured cattle do not end up with the designated beneficiary but with someone else. Leakages in public service increase the cost (input) of achieving a given level of service (output) and result in a deteriorating measure of efficiency. But since leakage is not a consequence of choices in policy, technology, or management, for operational purposes leakage should be detected separately. 97. Step 3: Select the indicators of efficiency. The most simple but highly useful indicator of efficiency is unit cost (the cost of producing one unit of output). Other things being equal, reducing unit cost will improve the efficiency of production. While knowing the level of unit costs is important, it is also necessary to know whether the current level of unit costs is appropriate and to know which factors most affect variations in unit cost across different projects. There are several ways to assess whether the level of current unit costs is appropriate. For instance, unit costs for procured goods can be compared with market prices. Unit costs of centrally and locally prepared items may also be compared. For civil work such as building roads, unit costs can be compared with other projects and guidelines from the Ministry of Infrastructure. 98. Step 4: Analyze the findings and develop hypotheses of what drives inefficiency. It is one thing to identify the factors that mechanically drive unit costs up or down. It is quite another thing to identify the underlying causes of these factors and prescribe remedial actions at the local, MAAIF, and national government levels. 4.2. Sample selection and procedures for field work 99. The complex structure of agricultural projects in Uganda demands that the sampling strategy be considered carefully. For this analysis, development projects were selected if their budgets exceeded 3 percent of MAAIF's development budget.35 Table 23 lists six projects that meet this criterion.36 Together they account for 74.1 35 The Farm Income Enhancement Project also falls within this category, because it accounts for 8 percent of MAAIF's development budget. Because this project was just launched, little activity has occurred and no results have been achieved, so it is not included in the analysis. 36 During the period under review, MAAIF executed 24 development projects. 45 percent of MAAIF's development budget and thus offer appropriate scope for the analysis. Table 23: Shares of selected projects in MAAIF's development budget, 2005/06­ 2007/08 Project Share of development budget (%) 1. Farming in Tsetse Areas of East Africa (FITCA) 3.6 2. North-West Smallholder Agricultural Development Project (NWSADP) 11.6 3. Livestock Disease Control Project (LDCP) 6.6 4. National Livestock Productivity Improvement Project (NLPIP) 19.7 5. Support to Fisheries Development Project (SFDP) 17.1 6. Vegetable Oil Development Project (VODP) 15.6 Total 74.1 Source: Policy Document for MAAIF (June 2007). 100. Geographically, the analysis covers much of the country. Projects were analyzed in Mbarara, Kiruhura, and Bushenyi Districts (western Uganda); Mukono and Wakiso Districts (central Uganda); Arua, Nebbi, Yumbe, Koboko, Moyo, and Adjumani Districts (northwestern Uganda); and Palissa, Budaka, Mbale, Tororo, Busia, and Soroti Districts (eastern Uganda). Subcounties within districts were randomly selected. The districts covered in each project were:37 x Livestock Disease Control Project: Kiruhura and Bushenyi. x National Livestock Productivity Improvement Project: Pallisa, Bushenyi, Soroti, Kiruhura, and Mbarara. x Vegetable Oil Development Project: Pallisa, Soroti, and Tororo. x North-West Smallholder Agriculture Development Project: Arua, Nebbi, Koboko, Yumbe, Moyo, and Adjumani. x Farming in Tsetse Areas of East Africa: Pallisa, Soroti, and Tororo. x Support to Fisheries Development: Lake Victoria. 101. The analysis drew on a range of primary and secondary information, beginning with preliminary interviews with government officials, both at MAAIF and in local governments, and moving on to include project documents from MAAIF and local governments, government reports and budgets, various policy documents, and fieldwork and interviews at the national, district, and subcounty level. 102. The analysis was initiated by meeting the respective technical staff of MAAIF and heads of various units, departments, projects, and subject matter specialists in the sector. 37 Although other agricultural projects are implemented at the district level, such as the World Food Programme's Agricultural and Marketing Support Project, the Support for Irrigation Project, PMA activities, and agricultural extension, this report focuses on the projects presented in the main text. 46 This approach was taken to create ownership of the process and enable the staff get acquainted with the tools and methods of the study. 103. In each district, interviews were conducted at headquarters with technical officers and at lower local government levels. These key informant interviews aimed to elicit respondents' perceptions of service delivery in their districts. In addition, group discussions were conducted in the Production Department in all districts, because it is the main coordinating department. These interviews sought to elicit technical perceptions of delivery mechanisms under the projects. Similarly, field visits were made to learn about beneficiaries' perceptions of delivery of planned outputs in their localities and assess how beneficiaries had been affected. Key project documents such as budgets, quarterly and annual reports, final accounts, and work plans were reviewed to study the planned activities, unit costs, target groups, and timeframe, among other things. 4.3. Implementation arrangements of MAAIF development projects 104. The resource flows in agricultural development projects are complex. Funds originate from various sources--donors, central government, and local governments. Each development project has a Project Management Unit (PMU), usually located in Entebbe but sometimes in regional centers such as Arua, as in NWSADP. The PMU receives funds from donors and the central government, and part of these funds is transferred directly to the districts via a special account. The production departments in the districts are responsible for overseeing the activities of subcounties and service providers and for procuring small contracts. The districts also cofinance projects from their local budgets.38 Figure 12 depicts the flow of funds and project implementation arrangements. 105. Large procurements are centralized at MAAIF, however. For instance, about 83 percent of funds released from ADB for NWSADP between 2005/06 and 2007/08 were used by the National Project Coordination Unit. For NLPIP, this figure was about 60 percent. All funds for SFDP and LDCP were used centrally. It is argued that local governments lack capacity in procurement and in controlling the use of funds. 38 In this report, the terms "district" and "local government" are used interchangeably. 47 Figure 12: Implementation arrangement for most MAAIF agricultural development projects Source: AgPER Task Team based on MAAIF project documents. 106. Public works, procured centrally, are supposed to be overseen by local governments. As the subsequent analysis will show, weak coordination between the central and local governments and inadequate incentives for local monitoring caused many contracts to be implemented inefficiently. 107. The complex implementation arrangements of most projects studied here posed several challenges for conducting PETS. A PETS analysis usually aims to examine flows of funds and materials from central government to local service providers, via regional and local governments, to identify resource use and leakages. In the projects studied here, however, only a small share of funds is transferred directly to local governments (Figure 12), and the bulk is spent at the central level to procure goods and public works. Procured goods are transferred to the districts, but most public works, such as road and market construction, bypass district budgets. For this reason, the analysis tracks only tangible resource flows between central and local governments and looks separately at the delivery of centrally procured public works. 4.4. Analysis of technical efficiency 108. The major components of the MAAIF projects under review can be classified into two groups. The first group consists of investments in infrastructure, such as the construction and/or rehabilitation of rural roads, bridges, livestock markets, and fish ponds. A full value-for-money analysis was done for this group, because field assessments and interviews with local government and beneficiary communities made it 48 possible to evaluate the progress and results of public works. The second group consists of the supply of goods (mainly inputs), advisory services, and trainings to farmers. As indicated previously, the outputs and outcomes of those interventions cannot be measured. The analysis is limited to an assessment of the quality of delivery mechanisms and the strategic alignment of project interventions with overall agricultural policy. 109. Before proceeding with the analysis, it is important to understand how the broader environment enables or impedes project performance. The prerequisites for launching projects quickly and smoothly are often beyond the immediate influence of MAAIF, including ratification by Parliament, the provision of counterpart funds, and overall procurement and fiduciary capacity in the public sector. If foreign loans are ratified without delay, and if counterpart funds are provided in full and on time, MAAIF and local governments are "enabled" to deliver goods and services efficiently to beneficiaries. If the opposite is true, a project's effectiveness is likely to be hampered from the very beginning. 110. In Uganda, the environment is not entirely conducive for implementing agricultural projects. It takes about one-and-a-half years for Parliament to ratify a loan. Delayed ratification reduces project benefits. For NLPIP, for example, a year's delay in project implementation is estimated to reduce the economic rate of return by 6 percent (from 23 percent to 18 percent). The untimely release of counterpart funds by MoFPED reduces the quality of project implementation. As indicated previously, outturns of government funds ranged from 10 percent to 50 percent for the period under review (recall Table 21 and Table 22). After Parliament ratifies a project, one year is typically needed to establish a procurement and management unit that meets domestic and local requirements and open special project accounts, especially if a project includes more than one Ministry.39 Most of these issues are relevant to all projects, not just agricultural projects. Yet even if these problems are beyond MAAIF's immediate control, MAAIF can still raise them with MFEDP and propose concrete remedies. 111. An equally important concern is the lack of security in some areas of Uganda. For example, insecurity in Karamoja and Acholi Districts made it difficult to implement NLPIP. Insecurity in Northern Uganda has also made VODP less effective. This reality should be reflected when evaluating the performance of agricultural projects. 4.4.1. Infrastructure-related components 112. Poor infrastructure still limits farmers' access to markets and capacity to diversify. As Balat et al. (2008) demonstrate, increased trade in cash crops (coffee, tea, 39 No clear rules govern the opening of special project accounts at the district level when two or more ministries are involved in implementing a project. To open such an account, three signatures are usually required--the district chief, accountant, and Production Department staff. In the Farm Income Enhancement Project, involving MAAIF and MWLE, the presence of an Agricultural Department representative and Environment Department representative increases the number of staff involved from three to four. This project experienced a delay of one year as it clarified project procedures and opened special project accounts. 49 cotton, and fruits) has a strong positive effect on reducing poverty among Ugandan farmers. High marketing costs caused by poor rural infrastructure have left many Ugandan farmers with little choice but to produce staple food crops for domestic consumption and avoid commercial agriculture.40 The NAADS impact evaluation (Section 4.5) also confirms that poor rural infrastructure significantly constrains commercial opportunities for NAADS participants. 113. Despite the urgent need for infrastructure, the share of capital investment in the agricultural sector budget is low, estimated at 8.5 percent (Figure 8). Although capital investments constitute a significant share of MAAIF's budget (24 percent; Table 13), far more capital expenditure is needed to make a difference in rural areas. It will not be effective to scale up these expenditures, however, unless they currently demonstrate good value for money. Unfortunately, this is not the case for most of the MAAIF projects with infrastructure components reviewed here. Urgent action is required to improve the value for money spent on infrastructure under these projects, given its critical importance to rural growth in Uganda. 114. Delays in delivering infrastructure have been significant in the projects under review. The rural infrastructure component of NWSADP was delayed by five years. Only 10 percent of large projects were built on time, to specification, and within the agreed budget. Only 2 contracts of 25 issued for rehabilitating roads and building bridges were nearing completion in 2009, implying that 92 percent of contracts under NWSADP were at risk as the project prepared to close that same year (Table 24). Only 2 markets of 15 were nearing completion, implying that 86 percent of market construction contracts were also likely to default. If the project were to end in 2009 as planned and the remaining works remained unfinished, 56 percent of the project budget earmarked for infrastructure (UShs 23 billion) would have been utterly wasted. The situation in SFDP is similar: The rehabilitation of fish ponds has been slow, with much wastage. 115. Delays have various causes, mainly arising from improper appraisal and feasibility studies, weak coordination of implementation between central and local governments, ineffective procurement, and problems related to land tenure. Improper appraisals and feasibility studies have also caused cost overruns. In NWSADP, for example, the cost of building Moinya Bridge on Itikikwa River along Obilokongo-Moinya-Ayiri Road in Adjumani District exceeded the cost specified in the original contract by 69 percent, and it took four times longer than planned. Katu Bridge on Lurujo-Nyai Road in Arua District has--so far--taken three times longer to build than anticipated and posted an overrun of 27 percent. 40 Balat, J., I. Brambilla, and G. Porto (2008): Realizing the Gains from Trade: Export Crops, Marketing Costs, and Poverty. World Bank Policy Research Working Paper 4488, Washington, DC. 50 Table 24: Contract performance within NWSADP Contracts Contracts Contracts Risk-free Commencement Nature of project activity in nearing at risk contracts date number completion (%) (%) Road rehabilitation and 25 Nov. 14, 2006 2 92 8 bridge construction Market construction 2 (Adjumani 15 Jun. 12, 2007 86 14 District) Construction of water 6 Feb. 1, 2006 1 83 17 supply systems Rehabilitation of DFIs 2 Jul. 30, 2007 1 50 50 Construction of DFIs 2 Dec. 5, 2006 0 100 0 Source: Monitoring and Evaluation Reports, Project Office (2007/08). 116. Appraisals for bridge construction were flawed by numerous omissions in technical design and connectivity. Esia Bridge on Odu-Kwoma Road in Adjumani District, for instance, was not included in the original plan but proved important for connecting roads. Technical faults included omitting gabion boxes, underestimating the width of rivers, and failing to identify rivers across roads (as in Omuriba-Kangoo, Kova- Odramacaka, Leju-itia, Odima-Kololo, and Menjo-Yoro in Arua District). Technical inefficiency was exhibited by road contractors who failed to follow specifications for the thickness of the gravel layer or compacting, as indicated by the heavy vegetation growing through most roads. Aberi-Zombo Road in Nebbi District, Mongoya-Oya and Majale- Kilaji Roads in Yumbe, and Arivu-Jayia-Opia Road in Arua exemplify this waste and limited value for money. 117. Often improper appraisal has been caused by weak coordination of project preparation with the local government and local communities. Involving local governments in the implementation of centrally procured public works increases the effectiveness of implementation, but engagement with local counterparts needs to begin earlier, at the project appraisal stage. Local counterparts' participation in project appraisal creates the sense of ownership that fosters more effective implementation. The same recommendation is valid for delivering centrally procured goods such as vaccines and cattle, as discussed later. 118. During implementation, wastage can also be reduced through better coordination of activities and allocation of adequate operating funds for supervision by local production department staff. Local technical staff members are overburdened by multiple projects and lack operating funds. Funds for supervising infrastructure projects are often included in vendor contracts but not transferred (or transferred too late) to local governments. Under SFDP, district fishery departments obtained motorcycles but not the corresponding operational funds, leaving insufficient resources to mobilize farmers, prepare contracts, and supervise work. 119. Public works on private and community land were delayed by unclear land tenure before the project began. Most markets built under NWSADP are being constructed on private land, and disagreements have arisen in some instances. For example, the priest in Yumbe objected to the water tank being built on mission land, and 51 subcounty authorities were drawn into a dispute with the owners of land for Keri market in Koboko and Nyaravur market in Nebbi. 120. The resulting cumulative wastage is high in the three MAAIF projects with infrastructure components. In NWSADP, NLPIP, and SFDP, capital expenditures accounted for 43.4 percent of the budget (Table 25), 81.2 percent of MAAIF's capital budget, and 20.4 percent of MAAIF's development budget. Table 25: Percentage of capital expenditures (approved budgets) of NWSADP, NLPIP, and SFDP in the budget of these projects, MAAIF's capital budget, and MAAIF's development budget, 2005/06­2008/09 2005/06 2006/07 2007/08 2008/09 Total Share of capital spending in three projects in 42.1 32.5 48.6 53.5 43.4 total budget of these projects (%) Share of capital spending of three projects in 79.1 91.2 85.8 80.9 81.2 MAAIF capital expenditure (%) Share of capital spending of three projects in 21.0 17.1 22.9 19.8 20.4 MAAIF development budget (%) Source: AgPER Task Team estimate, based on project documents and approved budgets for MAAIF. 121. Not only were most of the resources in the infrastructure components at a high risk of being wasted--the unit costs of most work were inflated. The unit costs of constructing roads were much higher under NWSADP than other programs. For example, unit costs in NWSADP ranged from UShs 30.9 million to UShs 38.5 million per kilometer, while the unit costs of road improvements under the Area Agricultural Modernization Program (AAMP) were UShs 8­20 million per kilometer. Although the costs of building and rehabilitating roads under NWSADP were estimated by MOTWC engineers, who conducted the surveys and participated in evaluating the bids, more realistic appraisals are necessary to plan implementation accordingly. One might argue that roads constructed under NWSDAP are not directly comparable with those constructed under AAMP, but field assessments show that the terrain was not any hillier under NWSADP roads than under AAMP roads. 4.4.2. Project components related to the delivery of inputs and other goods 122. About 42 percent of MAAIF's development budget and a similar share of the agriculture budget went to supply goods and services (Section 2.3). The share of goods and services in MAAIF's nonwage recurrent expenditures rose from 52 percent in 2005/06 to 82 percent in 2008/09 (Table 13). How well were inputs and other goods actually delivered? 123. There is evidence that goods procured locally cost less and were less prone to wastage and leakage than goods procured centrally. The superior performance of local procurement and goods distribution is explained by lower transaction costs and an increased sense of ownership. The unit costs of vaccines procured centrally by LDCP and NLPIP exceeded market prices; for example, lumpy skin vaccine procured under these projects cost UShs 1,300 versus the market price of UShs 1,200, and brucellosis vaccine 52 cost UShs 1,000 versus the market price of UShs 800. The high prices of these vaccines partially explain their low demand among farmers. 124. Prices of cattle procured centrally under NLPIP also exceeded market prices. According to the 2005 quarterly project report, MAAIF procured Boer goats at a unit cost that was nearly four times higher than it would have been at the local level (UShs 892,000 versus UShs 250,000). Even the local goats procured by MAAIF cost 1.5 times more than goats procured locally (UShs 70,000 versus UShs 50,000). The unit costs of centrally procured cattle under NLPIP exceeded those under NAADS, PMA, LGDP, and Northern Uganda Social Action Fund. Based on data for June 2007/08, if all local goats (13,087) procured centrally under NLPIP were procured by districts, the savings (or losses avoided) would have reached UShs 262 million, or 30 percent of the budget for central procurement of local goats. In contrast, the unit costs of oxen and spray drugs procured at the district level by FITCA were close to reported market prices. 125. The delivery of centrally procured animals was prone to wastage. Under NLPIP, the poor health of many goats reportedly made potential beneficiaries in Kiruhura and Soroti Districts reluctant to accept them. According to project documents, of 14,239 animals distributed by mid-2008, 998 adult animals (7 percent) died and 235 abortions (2 percent) occurred. In Nyakayojo Subcounty of Mbarara District, 6 of 16 goats given to farmers (38 percent of distributed goats) died.41 Casualties were worse when animals were supplied during the dry season and no facilities were available to keep them, partly because district staff had not been involved. Under FITCA, 6 of 31 tsetse-tolerant heifers (19 percent) supplied to farmers in Palissa District for animal traction reportedly died in the first months. The death rate in Budaka was reportedly 29 percent. 126. Serious wastage was also observed in supplying centrally procured vaccines. Farmers usually are reluctant to vaccinate animals unless a major disease outbreak is in progress. They are even more reluctant when they have to pay for vaccine to be administered, especially when it costs more than it does in the local market. This additional cost, according to respondents in Kiruhura and Bushenyi, led few farmers to use lumpy skin vaccine supplied through NLPIP. In Kiruhura, Mbarara, and Bushenyi, low use of 320,000 doses of brucellosis vaccine, intended to be supplied on a cost- recovery basis, caused MAAIF to vaccinate animals free of charge to avoid losses from expired vaccine. Wastage was also caused by a lack of communication between MAAIF and the districts. Districts often received vaccine in quantities that exceeded the capacity of the cold chain and thus could not be stored properly. 127. Records of transfers of physical and financial resources from central management units to the districts often do not match, which could be indicative of wastage, diversion, or leakages. The source of the problem is unclear. It could be poor record keeping (which implies low accountability), diversion of resources away from the intended areas or beneficiaries, corruption, resource misallocation, improper accounting, or incomplete information from PMUs or local governments. Table 26 shows 41 Unpublished 2007/08 subcounty monitoring report. 53 the mismatch of records under NWSADP: Arua District reported receiving more funds than the PMU reportedly transferred. No discrepancies were found for transfers to Nebbi District, but there was a 5 percent shortfall for Koboko District. Under NLPIP, the differences are more pronounced. Three districts reported receiving less funding than reported by the PMU: Mbarara (22 percent), Soroti (2 percent), and Palissa (22 percent). In Bushenyi District, the district authority reported receiving no funds, though it is not clear if this discrepancy arose from improper accounting or leakage (Table 27). Table 26: Financial resources transferred by the NWSADP Project Management Unit and received by districts (UShs millions) 2006/07 2007/08 Total Reported financial transfers from PMU to districts Arua 96.5 302.2 398.6 Koboko 46.1 113.9 160.0 Nebbi 205.7 323.2 528.9 Financial transfers reported received by districts Arua 313.9 282.6 596.5 Koboko 32.4 120.9 153.3 Nebbi 205.7 323.2 528.9 Source: NWSADP Project Office and districts. Table 27: Financial resources transferred by NLPIP Project Management Unit and received by districts (UShs millions) 2006/07 2007/08 Total Reported financial transfers from PMU to districts Mbarara 136.3 14.0 150.3 Bushenyi 22.8 0 22.8 Soroti 48.5 30.2 78.7 Palissa 50.2 24.0 74.2 Financial transfers reported received by districts Mbarara 114.4 0 114.4 Bushenyi 0 0 0 Soroti 48.5 29.0 77.5 Palissa 50.2 7.4 57.6 Source: NLPIP Project Office and districts. 128. Leakage was also reported when centrally procured goods were transferred to districts. Under LDCP, the leakages were quite high (Table 28). There was no leakage of goats supplied by NLPIP and 4 percent leakage of heifers/zebu cattle (Table 29). 54 Table 28: Physical resources reportedly transferred by the LDCP Project Management Unit and received by districts (000 doses of vaccine) Bushenyi Kiruhura Mbarara Budaka Soroti Min. Dist. Min. Dist. Min. Dist. Min. Dist. Min. Dist. Rabies 7.5 11.5 20.5 11.0 1.0 Foot and mouth 43.1 32.9 321.0 9.5 disease Newcastle disease 103.0 53.0 55.0 5.0 50.0 50.0 20.0 28.8 Lumpy skin 9.1 11.0 24.0 25.0 0.7 3.9 disease CBPP 0.1 0.0 130.0 0.0 1.5 5.0 Brucella S19 60.6 30.5 43.5 20.0 25.0 1.4 4.0 1.6 Total 223.4 138.9 594.0 50.0 25.0 24.1 50.0 56.0 24.0 34.4 Source: LCDP Project Office and districts. Table 29: Physical resources reportedly transferred by NLPIP Project Management Unit and received by districts (numbers of animals) Goats Heifers/zebu cattle Cattle supplied by PMU to district Mbarara 755 Bushenyi 400 Kiruhura 700 Soroti 1,595 Palissa 1,423 Cattle received by district Mbarara 755 Bushenyi 400 Kiruhura 700 Soroti 1,560 Palissa 1,353 Source: NLPIP Project Office and districts. 129. Despite the leakage found under some projects, project procedures for identifying potential beneficiaries took the geographic distribution of poverty into account, in an effort to reduce leakage. Most projects used MoFPED's formula for distributing goods and services supplied through poverty alleviation funds to local governments. The formula weights districts by population and poverty level (an example of how the formula was used to distribute goats across NLPIP districts is shown in Annex 2). It is unclear, however, how beneficiaries were selected at the community level. 130. The last but not least weakness identified in the projects reviewed here is the lack of strategic direction and of complementary public expenditure. Stand-alone projects should nevertheless be part of a coherent agricultural development strategy. The success of many projects depends on complementary public expenditures and policies. If a project introduces improved seed and fertilizer to farmers, for example, farmers will need to be able to obtain those inputs in the market even after the project ends. If access to those inputs is not addressed systematically by improving roads and strengthening the network of agro-dealers, the project usually does not result in the sustained use of technologies by farmers, and the resources invested are simply wasted. 55 131. Distribution of vaccines or live animals is ineffective without complementary measures for effective disease control. Even if wastage from vaccine distribution and utilization can be reduced, vaccination alone is unlikely to improve animal health substantially. Similarly, distributing livestock without addressing the prevalent diseases will result in low value for money and wastage, because animals often die on delivery. Livestock disease control has not been a priority for government spending. Only 7 percent of Uganda's agriculture budget is allocated to controlling livestock diseases (Table 15, Section 2.4), and even that small sum is irregular and unpredictable. MAAIF can detect and control only one-sixth of the dangerous diseases listed by the World Health Organization and only those caused by ticks and tsetse. Diseases caused by viruses cannot be traced at all. The central veterinary lab is understaffed and ill equipped. Samples are sent overseas, usually to Italy, for testing, and results are not available for two to three months, which is a very long time to deal effectively with dangerous diseases. The weak capacity of MAAIF and district veterinary departments to detect and eradicate diseases leads to poor livestock health and consequently low livestock productivity. Under these circumstances, the distribution of animals or even vaccines, even if it can be done more effectively, may not by itself be a cost-effective means of achieving the desired outcome--improved livestock health. 132. Another drawback in choosing the proper low-cost intervention to address livestock diseases is the lack of a central mandate to coordinate the response to disease outbreaks. Livestock disease control has been shifted to the districts, with MAAIF providing technical and financial assistance. Diseases do not have borders, however. District veterinary officers use animal checkpoints along known routes to deter illegal movement of animals. Those who move the animals should have permits and trading licenses. The checkpoints control disease better then vaccination, but when disease outbreaks occur, neighboring districts do not necessarily limit the movement of animals, and the outbreaks spread. When interviewed, local government staff who are directly involved in managing outbreaks proposed that the response to disease outbreaks should be centrally coordinated to ensure that all parties comply with control measures. 133. A final example of sources of waste relates to microfinance. Almost all projects have sought to improve farmers' access to credit through Savings and Credit Cooperatives (SACCOs). NWSADP planned for farmers to obtain credit from SACCOs to purchase improved seed and fertilizer, introduced by the project during the first year. The lack of viable SACCOs and other microfinance institutions in most project areas, however, delayed the project's production enhancement component by five years. SACCOs alone cannot solve farmers' credit problem, however. Complementary work is required on collateral, regulations, and macroeconomic improvements to ensure that smallholders can obtain low-interest loans from SACCOs. Even if a SACCO is established in an area, smallholder farmers do not automatically gain access to credit. On the contrary, observations and interviews in rural areas show that most microcredit goes to better-off households with diversified activities and off-farm income. And as in past experiences with supply-driven rural finance, the low rates of repayment (55­59 percent in NWSADP and 57­69 percent in AAMP) raise similar concerns about the sustainability of SACCOs. When preparing and appraising project activities, it is vital to have realistic expectations of what SACCOs or other rural financial institutions may and may not offer. 56 Funds will be wasted if project components are designed based on the assumption that potential beneficiaries will have easy, automatic access to credit, use it effectively and be able (and willing) to repay. 4.5. The impact of NAADS 134. The National Agricultural Advisory Services (NAADS) program is a 25-year program with an initial phase of 7 years.42 Its goal is to increase market-oriented agricultural production by empowering farmers to demand and control agricultural advisory and information services. The specific objectives of NAADS are to: x Increase the effectiveness, efficiency, and sustainability (including financing, private sector participation, farmer responsiveness, deepening decentralization, and gender sensitivity) of the extension delivery service. x Increase and sustain farmers' access to knowledge (education), information, and communication to farmers. x Increase and sustain farmers' access to effective, efficient, and productivity- enhancing technologies. x Create and strengthen linkages and coordination within the overall extension services. x Align extension to government policy, particularly privatization, liberalization, decentralization, and democratization. 135. NAADS receives the highest share of the agriculture budget. It accounted for 36 percent of the average sector budget during 2005/06­2008/09, rising from 30 percent in 2005/06 to 46 percent in 2008/09 (Table 2 and Table 3). At the beginning of implementation, the NAADS Secretariat handled many procurement and other functions on behalf of the districts and subcounties, but over time spending has shifted from the Secretariat to local governments, as districts and subcounties assumed their functions and as more districts were added to the program. Whereas only 59 percent of the NAADS budget was managed at the district level in 2005/06, in 2008/09 this share was programmed to reach 83 percent. 136. Investments in NAADS have brought high economic returns. The overall benefit-cost ratio was estimated at 2.5 in the impact evaluation, indicating that every shilling invested in NAADS yielded about 2.5 shillings of net social benefits. These findings are supported by qualitative findings that a higher proportion of NAADS farmers feel their food security, wealth, and overall wellbeing has improved over the past four years relative to the non-NAADS members. 137. Since most NAADS activities are now locally designed and executed, the districts receive financial transfers instead of centrally procured goods and services. 42 As mentioned, this subsection is based on a recent IFRPI impact evaluation, complemented by AgPER Task Team findings in the districts visited. 57 Ending the central procurement and transfer of goods has reduced the scope for leakages. Decentralized procurement, however, has not eliminated all concerns about whether NAADS delivers value for money. In several instances, inputs have reportedly been provided at a higher cost than the prevailing market price, and their quality has also been reported to be poor. These issues were associated with the procurement process and are now addressed by entrusting procurement directly to the communities. With substantial resources now going into inputs, the concern with value for money will continue, calling for closer monitoring. 138. Districts usually report that they have received the financial resources for NAADS as planned. There are some minor discrepancies--mostly reports of higher transfers than shown in the records at the Secretariat (Table 30).43 The scope for leakages associated with central management, however, has been eliminated: Funds are transferred directly from the Treasury to the districts, not through the NAADS Secretariat, and are to be used only for expenditures related to NAADS. Table 30: Flow of financial resources from the NAADS Secretariat to districts, 2005/06­2006/07 (UShs millions) 2005/06 2006/07 TOTAL Reported financial transfers from the Secretariat to districts Arua 1,246.0 1,092.0 2,338.0 Mbarara 427.0 622.0 1,049.0 Bushenyi 891.0 1,277.0 2,168.0 Pallisa 224.1 224.1 Financial transfers reported received by districts Arua 1,307.8 1,092.0 2,400.0 Mbarara 667.8 622.0 1,289.1 Bushenyi 823.6 1,283.8 2,107.4 Pallisa 224.1 224.1 Source: Secretariat and district NAADS offices. 139. Farmer training generally covered such skills as developing a group constitution and bylaws, leadership, growth and development, planning, record keeping, and savings mobilization. More groups were trained in subcounties where NAADS first started than in subcounties where the program started later or was not implemented. About 90 percent of farmers found the training to be very useful or useful, although more groups in the initial NAADS subcounties reported that it was very useful or useful, compared to groups in subcounties where the program started later. These findings suggest that NAADS is gradually strengthening farmers' skills and institutional capacity to improve natural resource management, agricultural productivity, and marketing. NAADS service providers, compared to others, were rated very highly on their training methods and performance. 43 The small deviations between funds reported transferred and received might be explained by delays in reflecting the full accounts at the district level or by improper placement of funds between fiscal years. 58 140. Farmers generally felt that their sense of empowerment, especially in relation to technical public officers, was stronger. This response is very encouraging, because technical public officers are well placed to address farmers' agricultural production problems and concerns. But technical public officers also received the worst performance rating in terms of their response to farmers' requests. This rating raises concern that the ability of advisory services to meet farmers' demands is weakening as NAADS expands to more districts and subcounties. It is critical to provide sufficient resources to maintain or increase advisory services for farmers already participating in the program and for new farmers who will enter the program. 141. The evaluation findings indicate that NAADS has been successful in increasing farmers' capacity to demand improved crop varieties, crop management practices, livestock breeds, postharvest practices, and marketing information, indicating that the demand-driven approach is working. On the other hand, participation in NAADS is associated with a lower probability that farmers will demand improved soil fertility and agro-forestry practices, suggesting that either farmers have less capacity to demand these technologies and/or that NAADS is not paying sufficient attention to them. NAADS spent relatively few resources on demonstrations of soil fertility management practices, compared to, for example, acquiring improved planting material. To ensure sustainable productivity, NAADS must increase farmers' capacity to demand soil fertility management practices. In the initial stages of demand-driven approaches, it may be necessary to proactively provide advisory services on soil fertility and agro-forestry practices to build farmers' capacity to demand them. 142. Consistent with the positive impact on capacity strengthening, demand for advisory services, and adoption of improved technologies, NAADS has had a significant impact on crop productivity. The gross value of crop output per acre was 29 percent higher for farmers participating in the NAADS program than for nonparticipants. The impact of the NAADS program on livestock productivity is surprising: A decline of about 27­45 percent in the value of gross livestock output per unit of animal was found among NAADS participants compared to nonparticipants. This finding requires further analysis, because it is not clear how an association with NAADS would make livestock less productive. 143. NAADS participants were associated with a 42­53 percent average increase in per capita agricultural income compared to nonparticipants. The impact of the program on total agricultural income was more favorable than its impact on income from crops and livestock, owing to the additional substantial income from participants' high- value agricultural activities, such as beekeeping and aquaculture. This finding demonstrates the program's effectiveness in diversifying incomes and promoting more high-value activities. The results also show that compared to nonparticipants, a significantly larger proportion of NAADS participants perceived their situation to have improved; nonparticipants felt that it was unchanged or worse. For example, 41­58 percent of NAADS participants perceived that their average wealth, ability to meet basic needs, overall wellbeing, access to adequate food, and the nutritional quality of their food had improved between 2000 and 2004 and between 2004 and 2007, compared to 27­44 percent of their nonparticipating counterparts. These results are consistent with the 59 positive impacts of NAADS on the adoption of improved technologies and agricultural productivity and income, and they suggest that NAADS has helped farmers improve their households' standard of living. 144. The impact evaluation clearly indicates that NAADS could have had a greater impact if its activities had been complemented by other rural services and public investments. For example, NAADS participants increased their sales of agricultural output only marginally. Crop sales were 6 percent higher for NAADS participants compared to nonparticipants, and sales of all agricultural produce were only 4 percent higher. The impact on sales of livestock products was negligible. The limited impact on sales results from the failure to invest in infrastructure and market development, as mentioned throughout this report. 145. Another finding of the impact evaluation supports the call for increased public investments to complement NAADS: The program has had its greatest impact in better-off areas. Income growth was highest in central and western Uganda, where the per capita income of NAADS participants rose by 65­165 percent between 2004 and 2007 compared to that of nonparticipants. The better-off regions are characterized by higher quality infrastructure and better access to such services as finance and transport. The experience of NAADS strongly confirms the importance of a coherent strategy for "packaging" public expenditure to support sustainable agricultural development. 4.6. Incidence of agricultural public expenditure 146. One measure of the effectiveness of public expenditures is the extent to which they are reaching the targeted beneficiaries among the rural population. This kind of benefit incidence analysis is particularly helpful in assessing the extent to which public services are pro-poor, which is a key objective of government and development policy. Using the household survey data that were also used to estimate the impact of NAADS, the incidence of some public services in Uganda is examined here. The analysis begins by examining the incidence of public expenditure by wealth quintile. It is then complemented by as assessment of financial resources received from all sources, governmental and nongovernmental. The final sections assess the incidence of access to information and advisory services, infrastructure, and other public services (in other words, core public goods). 4.6.1. Incidence of public support and nongovernmental credits 147. Are the various forms of public support pro-poor? As noted in Section 2.3, expenditures on inputs and other private goods constitute the largest share of agricultural expenditure in Uganda. A key rationale for subsidizing inputs and other private goods is to help farmers, especially poor farmers, obtain new technologies and inputs. Yet the results of the 2007 household survey show that the opposite has occurred. The Integrated Support to Farmer Groups (ISFG) grants, as well as non-ISFG NAADS support, have been concentrated mostly among wealthier households (Figure 13). Overall, Uganda's public support to agriculture, which is dominated by expenditures on private goods such as inputs, appears to have been regressive. 60 148. The wealthiest quintile of farming households also benefited the most from credit and other forms of assistance.44 Total support from all sources--including credit as well as assistance from nongovernmental and governmental sources--was highest among the most affluent farmers (Figure 14). The proportion of households receiving any kind of support also increased with wealth, although the differences among wealth groups were less marked (Figure 15). Overall, between 40 and 50 percent of households received support in each of the top four quintiles, and the data show no gender bias. 149. Households participating in NAADS showed greater access to support in every quintile, although the mean total support per capita is similar among NAADS and non-NAADS participants across all quintiles, particularly when averaged across households getting support from any source. This finding suggests that support provided to NAADS households is more extensive and more equitable than support supplied to other households (Figure 16 and Figure 17). On per capita basis, however, the households that could obtain credit benefited the most from grants and subsidies. This finding holds true for NAADS members as well (Figure 18). 44 Wealth is highly correlated with household consumption and income. 61 Figure 13: Incidence of public agricultural expenditures by household wealth quintile 60 40 20 0 Incidence of Govt. Support per capita by Source : 2007 1 2 3 4 5 NAADS (non-isfg) Support (USh) ISFG Support (USh) Other Govt. Support (USh) Source: AgPER Task Team, based on household survey. Figure 14: Incidence of total support by Figure 15: Proportion of households wealth quintile receiving support by wealth quintile Incidence of Credit, Grants & Subsidies over Wealth Quintiles : 2007 Incidence of Support over Wealth Quintiles : 2007 .5 300 .4 200 .3 .2 100 .1 0 0 1 2 3 4 5 1 2 3 4 5 Source: AgPER Task Team, based on household survey. 62 Figure 16: Incidence of support for Figure 17: Amount of support received by NAADS and non-NAADS participants by NAADS and non-NAADS participants by wealth quintile wealth quintile Incidence of Support over Wealth Quintiles : 2007 Incidence of Credit, Grants & Subsidies over Wealth Quintiles : 2007 Incidence for Naads Non-Members (Left) and Naads Members (Right) Incidence for Naads Non-Members (Left) and Naads Members (Right) 500 .8 400 .6 300 .4 200 .2 100 0 0 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 0 1 0 1 Source: AgPER Task Team, based on household survey. Figure 18: Incidence of credits and public expenditure by wealth quintile Incidence of Credit, Grants & Subsidies over W ealth Quintiles : 2007 3,000 2,000 1,000 0 1 2 3 4 5 Total Credit Support per capita (USh) Total Grant Support per capita (USh) Total Subsidy Support per capita (USh) Source: AgPER Task Team based on household survey. 63 4.6.2. Incidence of agricultural advisory services 150. The primary outputs of NAADS--advisory services and access to information--appear to be supplied more equitably than credits, grants, and subsidies. Across all households, the proportion that had received agricultural advisory services in the previous year (2006­07) was higher among wealthier households (Figure 19). Even so, a comparison of NAADS members and nonmembers reveals that NAADS members received significantly more advisory services and that those services were more evenly distributed across wealth quintiles (Figure 20). Among NAADS members, regardless of wealth quintile, the incidence of advisory services was unrelated to the level of education, while it was related to education for non-NAADS households. Advisory services were not supplied so equitably among male- and female-headed households, however. Households headed by men received more agricultural advisory services in 2006­07. Figure 19: Incidence of Agricultural Figure 20: Incidence of Advisory Services Advisory Services by wealth quintile by NAADS membership by wealth quintile Incidence of Agricultural Advisory Services Across Welath Quintiles : 2007 Incidence of Agricultural Advisory Services Across Welath Quintiles : 2007 Incidence for Naads Non-Members (Left) and Naads Members (Right) .5 .8 .4 .6 .3 .4 .2 .2 .1 0 1 2 3 4 5 1 2 3 4 5 0 1 2 3 4 5 0 1 Source: AgPER Task Team, based on household survey. 151. Access to crop market information and production practices was evenly distributed across wealth quintiles. Both NAADS and non-NAADS members accessed information on crop markets and production practices irrespective of wealth quintile. As a group, NAADS members had somewhat better access compared to nonmembers (Figure 21). 152. The more equitable distribution of information and advisory services provided by NAADS has increased the impact and effectiveness of this public expenditure. As shown in Figure 9, about 37 percent of total agriculture spending was allocated to the agricultural advisory services through NAADS. The high rates of return to NAADS investments described in Section 4.5 and the more equitable incidence of its advisory services suggest that the implementation arrangements have been successful. 64 Figure 21: Access to information on crop production methods and practices by wealth quintile Incidence of Information on Improved Crop Production Methods over W ealth Quintiles : 2007 Incidence for Naads Non-Members (Left) and Naads Members (Right) .8 .6 .4 .2 0 1 2 3 4 5 1 2 3 4 5 0 1 Improved Crop Production Technology Improved Crop Production Practice Source: AgPER Task Team, based on household survey. 4.6.3. Access to infrastructure and other services 153. The incidence of other core public services varies for social and agricultural services. Table 31 shows household access to various public and private services, measured as the distance from the homestead to the service point. Social services such as health, education, and drinking water are quite equitably distributed, whereas access to agricultural services is not. The poorest wealth quintile lives farthest from most agricultural and complementary services, especially extension officers, financial institutions, and primary consumption markets. 65 Table 31: Incidence of public and private services across wealth (per capita) quintiles, per capita in 2007 Median distance to service (km), 2007 Service Quintile 1 Quintile 2 Quintile 3 Quintile 4 Quintile 5 All-weather road 1.5 1.5 1.2 1.0 1.6 Seasonal road 0.4 0.2 0.3 0.2 0.2 Primary consumption market 1.6 0.8 0.8 0.8 0.5 Secondary (sellers) market 4.0 3.1 3.0 3.0 3.0 Livestock market 7.6 7.0 7.0 6.4 6.0 Slaughter facility 5.0 4.8 4.0 3.8 4.0 Input shop 8.0 7.0 6.4 6.0 6.0 Vet drugstore 7.0 6.4 6.4 5.0 4.8 Vet clinic 6.4 6.0 6.0 4.8 5.0 Livestock disease facility 7.8 5.0 6.4 4.8 4.8 Primary school 1.5 1.0 1.0 1.0 1.0 Public clinic 4.0 3.2 3.0 3.0 3.0 Private clinic 2.0 2.0 1.5 1.6 1.5 Human drinking water source 0.5 0.5 0.5 0.5 0.5 Livestock drinking water source 1.0 0.8 0.5 0.6 0.5 Fuel supply source 0.5 0.5 0.5 0.5 0.3 Bank 22.2 19.2 20.0 16.0 14.2 Microfinance institution 14.7 8.0 7.0 6.5 5.0 Extension/agricultural office 7.5 7.0 6.4 5.0 4.8 Source: AgPER Task Team, based on household survey. 66 5. Conclusions and Policy Recommendations 5.1. Agricultural sector budget: recent developments 154. Between 2001/02 and 2008/09, agricultural sector expenditures can be divided into two distinct phases. During the first phase, from 2001/02 to 2003/04, the budget for agriculture fell sharply in both nominal and real terms. It began to recover after 2004/05. In nominal terms, cumulative growth in the sector budget was 46 percent higher in 2008/09 than in 2001/02. Despite this seemingly spectacular increase, however, in real terms the 2008/09 sector budget was about the same as it was in 2001/02. Real expenditures were 38 percent higher in 2008/09 than at their low point in 2004/05. As a share of the national budget, the agricultural sector budget had fallen to 3.8 percent in 2008/09 from 5.7 percent in 2001/02. Finally, in terms of GDP, the sector budget remained stable, albeit low, at 1.6 percent. These are the statistics for the approved budget. The released budget was on average 10 percent lower, reducing the share of sector expenditure in GDP to a very low 1.2 percent. 155. The agricultural sector budget is not much larger when it is defined according to the United Nations classification of functions of government. Under the COFOG classification, recommended by CAADP/NEPAD for comparing agricultural expenditures across African countries, the sector budget as a percentage of the national budget is about 1 percent higher (for example, increasing to 5.4 percent for the released budget in 2005/06). Donors account for substantial off-budget spending, although information on these expenditures is fragmented, difficult to obtain, and largely ignored by planners and policy makers. For two donors for whom data are available, off-budget spending is estimated to account for 10­20 percent of the total sector budget. Overall, even with off-budget funds, the agricultural sector budget in Uganda is still about twice as low as necessary to meet Uganda's Maputo Declaration pledge to allocate 10 percent of the national budget to agriculture under CAADP. In addition to these public expenditures, significant volumes of expenditures by nongovernmental organizations, estimated conservatively at about 10 percent of the current budget amount, pose additional coordination and harmonization problems. 156. Is the agricultural sector expenditure expected to grow in the near future? It is not projected to grow but to keep declining, according to the MTEF. In 2012/13, the agricultural expenditure is expected to be 3.2 percent of national expenditure, compared to 3.8 percent in 2008/09 and 4.6 percent in 2001/02. Even though budget will shift to other areas, notably infrastructure and social development, these shifts will indirectly benefit agriculture. Within the agricultural sector budget itself, resources will continue to shift from the national to local level, mainly to NAADS in the districts. Given this outlook, it is more critical than ever to ensure that scarce budgetary resources are used in a highly effective, efficient way at all levels. It is also vital for MAAIF to improve collaboration with local governments, because they will increasingly be responsible for implementing agricultural programs. 157. How does Uganda's agricultural sector expenditure compare to that of other countries? As a share of GDP, public spending on agriculture in Uganda (adjusted by the 67 size of the sector) is low compared to spending in high- and middle-income countries but comparable to spending in selected Sub-Saharan African countries. Although additional public spending is required for agriculture in Uganda, two important points must be borne in mind. First, the current lower level of spending is consistent with Uganda's narrower fiscal capacities and its greater need to invest in infrastructure and social development. Second, international experience shows that lower public spending does not necessarily mean lower agricultural competitiveness. Vastly expensive farm subsidies in the United States and European Union have failed to keep farmers competitive internationally, while the small public expenditure on agriculture in Brazil, Australia, and New Zealand has not prevented farmers from being highly competitive on world markets. 158. Three conditions must be met to maximize the impact of public spending on agriculture in Uganda. These conditions will be even more critical if spending begins to rise again. First, public expenditures on agriculture must be supported through an enabling environment in which agricultural prices are subject to few distortions. It is counterproductive to raise the public expenditure on agriculture when farm-gate prices are depressed. Second, the mix of spending should be efficient (allocative efficiency). Spending that does not contribute (or does not contribute as much) to growth and poverty reduction, relative to alternative goals, is allocatively inefficient or unproductive. Third, technical efficiency should be high. Technical efficiency in the public sector involves making the best use of inputs to provide outputs in the form of public services. (Put simply, technical efficiency is doing things well, and allocative efficiency is doing the right things.) When those three conditions are met, even small budgets will be well positioned to generate economic growth and encourage more private investment. It is also very likely that when these conditions are met, the scaling up of agricultural sector expenditure in Uganda will bring the highest rates of return. 5.2. Agricultural price distortions 159. Uganda has successfully addressed one important component of the efficiency of public expenditure: Agricultural price distortions have been largely eliminated, and most farm-gate prices are at the level of reference border prices adjusted for marketing costs. In 2001­04, the rate of assistance to agriculture was estimated at 1 percent, with no taxation to exportables and about 13 percent support to importables such as rice. The nonagricultural rate of assistance has also notably decreased, reducing the prices of farm inputs and stimulating resource flows to agriculture. The policy environment in Uganda can be described as quite conducive for public expenditure to have a lasting impact. 5.3. Allocative efficiency of public expenditure 160. The allocative efficiency of public expenditure can and must be improved. The first drawback is the small share of capital expenditure in the sector budget. An economic decomposition of public spending illustrates that in Uganda "development expenditure" is not synonymous with "capital expenditure," as usually assumed. Development expenditure makes up about 80 percent of the agriculture budget but is heavily oriented to nonwage recurrent expenditures rather than to capital expenditures. 68 The share of capital outlays in the 2008/09 approved budget was 6 percent, down from an already low level of 11 percent in 2005/06. Too little is invested in rural community roads, bridges, and wholesale and livestock markets. Only a minor share of the budget is allocated to irrigation and to veterinary, sanitary, and phytosanitary laboratories and equipment, despite their considerable importance for raising agricultural productivity. Furthermore, much technical inefficiency in capital expenditures reduces the impact of even that small capital outlay to a very low level. 161. Uganda's low level of capital expenditure is bad news, because the condition of rural infrastructure strongly affects agricultural trade, farm diversification, and ultimately poverty. The recent study by Balat et al. (2008) in Uganda demonstrates that greater trade in cash crops (coffee, tea, cotton, and fruits) has a strong effect on poverty reduction. High marketing costs caused by poor rural infrastructure prevent many Ugandan farmers from participating in markets, leaving them little choice but to grow staple food crops for domestic consumption.45 162. The second drawback of agricultural sector expenditures is that they are increasingly dominated by nonwage recurrent expenditures, mainly subsidized inputs and other goods.46 Between 2005/06 and 2008/09, nonwage recurrent spending, mainly for farm inputs, comprised 65 percent of MAAIF's development budget on average. The share of nonwage recurrent spending in MAAIF's total budget grew from 49 percent in 2005/06 to 80 percent in 2008/09. Inputs are distributed by many development projects and increasingly by NAADS. In 2008/09, NAADS received about UShs 37 billion "for inputs to the small-scale farmers who cannot afford to purchase the necessary inputs." 163. Making inputs accessible is integral to raising agricultural productivity, but the current approach of distributing inputs is unlikely to achieve this objective. The current approach favors the wealthiest farmers, who can perhaps already afford these inputs. It does not strengthen private input suppliers, improve the targeting of subsidies (through the use of vouchers, for example), or invest in infrastructure to ensure that farmers can obtain inputs easily. Most input distribution and promotion is simply supply driven. It fails to meet the requirements of market-supporting smart subsidies. As allocations for subsidized inputs grow, the fiscal burden and economic distortions (waste and leakages, as well as displacement of input sales by private sector) will also grow, if the mechanism for delivering inputs remains unchanged. 164. The third problem with the allocative efficiency of the agricultural sector budget is the share of wage and operational expenditures going to MAAIF Headquarters relative to other MAAIF departments. MAAIF Headquarters absorbs 35 percent of MAAIF's recurrent budget. Although this large share is partially explained 45 Balat, J., I. Brambilla, and G. Porto (2008): Realizing the Gains from Trade: Export Crops, Marketing Costs, and Poverty. World Bank Policy Research Working Paper 4488, Washington, DC. 46 The analysis in Section 4.6 clearly shows the regressive nature of subsidies and fairly equal distribution of core public goods. 69 by higher wage allocations for senior staff, high transport costs to Kampala from Entebbe, and other recurrent expenses paid from Headquarter for services used by all MAAIF departments, Headquarters operating costs must be reduced and more operating funds shifted to technical staff in other departments of MAAIF, where recurrent expenditures are dominated by wages, leaving little operational funding for the technical staff to be effective. 165. Another way of assessing allocative efficiency is to analyze the functional composition of public expenditures. At first glance, the functional composition is quite efficient, given that the largest share of the budget is allocated to areas with the highest potential for enhancing pro-poor growth, namely advisory services and agricultural research.47 These categories absorbed 57 percent of the sector budget between 2005/06 and 2007/08. Even so, many other core public goods remain underfinanced, undermining the potential impact of research and advisory services. The critically underinvested areas are water capacity building, rural infrastructure, livestock and plant pest and disease control, regulatory services, and institutional development. These are core public goods. The private sector will not invest in most of them, but they are essential for catalyzing private investment in agricultural production and agribusiness. Investments in these public goods are also critical for ensuring that the geographically focused interventions of development projects can be sustained, in the sense that farmers retain access to markets and services even after development projects have ended and that technologies promoted by NAADS are available at the lowest possible cost. 166. Given these drawbacks, how can the allocative efficiency of public expenditure be improved? More specifically, how can Uganda increase its expenditure on critical public goods without undermining fiscal sustainability? In the short run, it might be not possible to shift substantial resources to critical functions within the existing budget, because these resources are committed to ongoing development projects. Resources for underfinanced critical public goods need to come from additional sources, including donor funds. In the medium to longer run, however, development resources must be shifted strategically from private to public goods and from nonwage recurrent to capital expenditures. Furthermore, the technical efficiency of agricultural spending must be improved, as indicated in the next section, but such improvements will increase the allocative efficiency of public spending and its impact on agricultural growth. 5.4. Technical efficiency of public expenditure 167. The technical efficiency of most MAAIF development projects is low. Technical efficiency must improve if MAAIF is to make a convincing case for substantially scaling up funding for agriculture. Not all of the blame for technical inefficiency rests with MAAIF. Many factors--neither specific to agriculture nor under MAAIF's immediate control--prevent projects from being implemented efficiently, including delayed budget ratification by Parliament, untimely and insufficient release of 47 The impact of NAADS expenditure is reinforced by the program's fair provision of services to all NAADS members, irrespective of their wealth ranking. 70 government counterpart funds, the weak procurement and fiduciary capacity of public officials, and insecurity in northern Uganda. These constraints cause several years to pass before a project can be launched effectively in Uganda, and they adversely affect a project's viability. The transaction costs of implementing a project rise when counterpart funds are constantly delayed and the procurement of goods and works is cancelled. MAAIF should keep raising these important issues and actively seek concerted remedies at the national and regional level. 168. Despite the many factors outside MAAIF's control, the remaining constraints on technical efficiency are significant and must be addressed by MAAIF and local government. Most MAAIF development projects with infrastructure components are characterized by delays of three to five years in building and rehabilitating rural infrastructure. Aside from the above-mentioned causes, these delays have been caused by improper appraisal and feasibility work, poor coordination of project preparation and implementation between MAAIF and local governments, inadequate operating budgets for local technical staff, ineffective project procurement, and problems related to land tenure. The results are high cost overruns, low-quality work, and other kinds of wastage. Given that capital expenditures account for such a large share of the development projects reviewed here (81 percent of MAAIF's capital budget), MAAIF must take immediate action to devise remedies based on these lessons. These lessons should also be considered when new investment projects are prepared. 169. Several other lessons on technical inefficiencies have emerged from assessments of how goods and services are being delivered to farmers and frontline service providers. First, the unit costs of goods procured at the central level were usually 20­50 percent higher than comparable market prices or unit costs at the local level. Second, the delivery of centrally procured goods was prone to wastage. The projects provided unhealthy livestock. A significant number of animals (ranging from 7 to 28 percent across interventions) died in the first months after delivery. Vaccine was wasted because it was overpriced and could not be stored. Records of transfers of financial and physical resources from central offices to districts often do not match, suggesting wastage, diversion, or leakage, although the precise source of the problem remains unclear. Possibilities include poor record keeping, indicating low accountability, diversion from intended beneficiaries or areas, corruption, resource misallocation, improper accounting, or simply insufficient information from project implementation units). Leakages were more pronounced for in-kind transfers than financial transfers, suggesting that inputs/goods are easier to redirect from the intended beneficiaries or purpose. These costs are lower for NAADS, because most procurement is being decentralized and the process is more participatory. 170. The impact of individual projects is considerably circumscribed by the lack of a strategic approach for using public expenditures to support agriculture. One example is the current strategy for improving farmers' access to inputs, which focuses on input distribution, complemented by training. This strategy fails to resolve the real causes of farmers' poor access to inputs, including high transaction costs (caused by poor rural infrastructure), slow progress in microfinance development, and the weak technical and 71 business capacity of private agro-dealers. Although too little capital is invested in rural areas, even that small expenditure can be improved significantly. 171. Other examples of how impact is reduced by the absence of a more strategic vision is the distribution of animals and the promotion of animal vaccines without investments in complementary disease control measures. The small share of the agricultural sector budget allocated to control livestock diseases in Uganda means that MAAIF's capacity to detect and control dangerous diseases is extremely limited. Under these circumstances, the distribution of animals is likely to result in low value for money and wastage. Similarly, distribution of vaccines, even if it is improved, is unlikely to improve livestock health sustainably. 172. The NAADS impact evaluation supports these findings about technical inefficiency in MAAIF development projects. NAADS has succeeded by being demand-driven and decentralized. Much of the wastage and leakage of centrally managed programs is avoided because most NAADS activities and procurement occur locally. Although NAADS has substantially improved the incomes of the program participants and shows a high economic rate of return, its impact would have been greater if its activities had been supported by complementary investments in rural services such as finance and roads. If these investments are made, they can reduce the cost of inputs, make inputs more accessible, and ensure that NAADS remains a powerful means of sustaining rural development. 173. Against this background, the following recommended actions would increase the effectiveness and technical efficiency of agricultural sector expenditure in Uganda: x Faced with the prospects of declining budget allocations in the future, the topmost priority for MAAIF is to ensure that its limited resources are used as efficiently and effectively as possible. MAAIF should target the highest priorities and seek institutional arrangements to deliver services more effectively through decentralized implementation arrangements. In line with the widely accepted principles of good governance, MAAIF should ensure transparency, accountability, and participation in its service delivery in an effort to maximize the efficiency and effectiveness of its expenditures. x MAAIF must ensure that problems with launching projects are at the top of the agenda for national stakeholders. Rates of return to development projects will increase significantly if Parliament ratifies loans rapidly, government counterpart funds are released in a timely and predictable way (budget execution), and public officials improve their procurement and fiduciary capacity. Remedies for these problems must be found at the national level, and MAAIF must play a critical role in this process. x The appraisal of infrastructure investments should be done in a participatory manner with local governments and local stakeholders. Land tenure issues must be clarified before projects are launched. Technical designs for communal roads and bridges should take into account connectivity with feeder 72 and national roads. Where possible, procurement should be shifted from the central to district level, as in NAADS. Procurement staff in local governments should receive training. These actions will reduce unit costs, minimize wastage, improve the quality of work, and lead to the timely delivery of construction work. x In some instances central procurement has merits. For example, it may be more cost effective to procure vehicles and other products internationally. Central procurement of such goods allows MAAIF to exploit economies of scale and effectively deal with import tax credits and other importation issues. Procurement of domestically procured goods and services must shift gradually to local governments, however. Compared to central procurement, local procurement usually has lower unit costs, wastage, and leakage. Local government should receive technical and financial assistance to strengthen procurement and fiduciary capacity and ensure the integrity of financial management. This support should be provided not only by MAAIF but also by national programs. x During project implementation, wastage can be reduced through better coordination of activities and the allocation of adequate operating funds for supervision by local production department staff. Local technical staff members are overburdened by multiple projects and lack funds to do their job. More resources should be set aside for operating expenses of local technical staff in development project budgets. x Individual projects should be better integrated into the strategy for agricultural growth and development. Public expenditures should address the roots, not the results, of market failures. It is inefficient to disseminate technologies in certain geographic areas without a coherent, sustainable approach to ensure farmers' access to those technologies once subsidies end. A larger share of public expenditure needs to be allocated to core public goods. The subsidization of private goods (inputs) must be considered only as a temporary measure to overcome market failures. The delivery of increased budgets for inputs and other goods should be directed at resource-poor farmers, which is currently not the case, and it should not displace but engage the private sector. 5.5. Budget processes 174. The strategic improvement of public spending on agriculture must begin by making greater use of the Sector Budget Framework Paper. The format and content of the SBFP are largely dictated by the guidelines in the Budget Call Circular, issued by MoFPED. It is time to reassess the information currently required for the SBFP as outlined in the Circular to determine if it is essential and presented in the most effective format for making informed decisions on budget allocations. Except for 2006/07, there has been no feedback to the Sector Working Group (or MAAIF) from MoFPED on the content and quality of the SBFP, nor are there suggestions for improvement in future years. Such feedback should become a regular feature of the budget preparation process. 73 Detailed recommendations for improving the SBFP guidelines are summarized in Annex 1. 175. The credibility of MAAIF's Development Sector Investment Plan should be strengthened by giving greater attention to: (i) the criteria used for prioritization; (ii) the expected outcomes; (iii) detailed explanations of expenditure estimates; and (iv) linking investment plans more closely to anticipated MTEF ceilings by indicating how plans would change if MTEF ceilings increase or decrease. There is an urgent need for MAAIF to update its strategy and investment plan for 2008/09 onwards to provide a continued link between policy, planning, budget preparation, and negotiations. MAAIF should also consider whether a three-year or five-year investment plan is best suited to its purposes. A five-year plan would provide a longer horizon for investment decisions and reduce the perceived need for frequent revisions, but changing priorities and events might make the estimates for outer years unreliable. 176. Budget execution, the binding constraint for implementing projects, should take into account the strong seasonality of agriculture (and thus resource requirements) and the large costs of untimely, unpredictable counterpart financing in development projects. Within the overall cash flow, MoFPED should cater for the relatively small but particular cash flow requirements of agencies such as MAAIF and NAADS to improve their operational effectiveness. 177. Substantial benefits will arise from creating a results-based system to monitor and evaluate public expenditures. This system would contain provisions for regular program monitoring and for evaluating the impacts of major interventions. Currently MAAIF focuses on regularly monitoring project implementation. Such monitoring is necessary to follow the progress and achievement of targets; it should cover not only budget but also off-budget expenditure, which is likely to account for 10­20 percent of the sector expenditure. It is also critical to ensure that monitoring reports are used inside and outside of MAAIF to reward good performance (or impose sanctions on poor performance), address inefficiencies, or reallocate resources between different priority areas. 178. Rigorous impact evaluation should be done for most development projects. The budget for impact evaluation needs to be sufficient and must be reflected in the project budget. Results-based M&E is indispensable to good national management and policy planning. The regular monitoring system provides very little information about the real impact of public programs. The absence of a results-based M&E system makes it easy for interest groups to manipulate public debates and thus very difficult to change current approaches to public expenditures and agricultural policy in Uganda. Impact evaluations can range from the PETS analysis to more detailed assessments such as the one done for NAADS. Once these evaluations become the norm, policy makers and planners will be well equipped to guide budget allocations across sectors and address operational constraints in agriculture programs. 74 Annex 1: Recommendations to Improve the Guidelines of the Sector Budget Framework Paper To improve the analytical content and internal consistency of the SBFP guidelines, it would be desirable to make the following adjustments, based on the example of the 2007/08 SBFP:48 x Section 1.3.1 Priorities. Undertakings to be implemented as a result of the PMA Joint Annual Review are listed. It is stated that they will also guide MAAIF's program of work, but currently there is no indication that this is the case, as the undertakings are not discussed in the remainder of the Budget Framework Paper. x Section 2.2 Overview of FY 2005/06 Performance. Low outturns are simply explained as "Shortfalls...in releases." A more detailed explanation of the reasons and responsibility for these shortfalls and whether these were one-time or recurring problems would be more helpful. Five pages of text are devoted to listing key outputs of activities as outlined in the 2005/06 Ministry Policy Statement, but this information is unlikely to be sufficient for proper planning. Currently all that is required in the SBFP is a listing of highlights of selected physical and qualitative performance of activities. It is difficult, however, to assess performance from this listing in the absence of the targets set for the preceding year. A tabulation of achieved outputs against targets would be much more informative and would help to establish a results-oriented management culture in MAAIF. This could be complemented with a list of, say, the 10 major achievements if required by the authors of the National Budget Framework Paper. x Section 3. Overview of FY 2006/07 Budget Allocations and Objectives. The observations made for Section 2.2 about budget and physical performance apply to Section 3, which deals with budget performance in the first half of the fiscal year. Although MAAIF needs to prepare quarterly Budget Performance Reports, which list activities undertaken in each department, they contain no systematic tabulation of planned and actual activities and outputs that could be used in the SBFP. x Section 4. Sector Budget Priorities for the Medium Term. The MAAIF Development Strategy and Investment Plan 2005/06­2007/08 had been used last year for the first time as the basis for sector planning and prioritization of spending plans. At first glance, there is a clear link between policy, planning, and budgeting decisions. However, as discussed in Section 1.4 of this report, there are wide variations between planned expenditures as outlined in the DSIP and the approved budgets proposed in the SBFP. 48 The content of the SBFP is complicated, because some agricultural activities (such as agro-processing, marketing, and collecting agricultural statistics) are shared with other ministries, agencies, and local government and not completely under the control or budget of MAAIF and its agencies. This situation leads to ambiguities over the precise content of the agriculture SBFP. 75 x Section 5. Expected Outputs, Performance Indicators and Planned Activities for the Medium Term. Eight pages of text list outputs, indicators, and activities for the 12 priority areas. A tabular presentation would have been more concise and effective, as it could have more clearly linked indicators to specific outputs and activities. A table in this format could then be used in conjunction with the table required in Section 9, showing outputs, indicators, targets, and allocation of recurrent and development budgets to specific activities. x Section 6. Proposed Budget Allocations for FY 07/08. Having listed and explained the 12 DSIP priority areas in Section 5, the table used in Section 6 to summarize recurrent, development, and total budget allocations is confusing, as it shows only 9 priority areas listed in a different order and consequently numbered differently from those in the DSIP and Section 5. In Section 6, two DSIP priority areas (livestock and fish disease control and livestock regulation) have been amalgamated as one priority. "Processing and marketing of crops, livestock, and fish" and "Promotion of increased agricultural production and productivity" are omitted; the activities and limited expenditure under these two priority areas are incorporated in other priority areas. Annex 2, indicating the sources of funding for each area, uses the same classification as the table in Section 6. x Section 7.1 Medium-Term Challenges with Implications for Additional Funding for Sector MTEF. This section is the one place in the SBFP where a claim can be made for additional resources. A table presents the gaps between the sums required to implement the DSIP fully in six priority areas,49 amounting to UShs 72.5 billion in total, but no attempt is made to identify priorities within this sum. Developed properly, the SBFP should provide an evidence-based justification for additional resources that might be compared by MoFPED with the claims made by other sectors. At present, the agriculture SBFP cannot be used in this way. Until MoFPED requires SBFPs to indicate and justify their expenditure priorities if their budget ceilings are raised or lowered (by, say, 5, 10, or 20 percent), MAAIF is likely to avoid making these hard prioritization decisions and much, or indeed most, of the rationale of the SBFP process will continue to be lost. x Section 7.2 Challenges with Implications for Additional Funding Outside the Sector MTEF Ceiling. This section allows the agricultural sector to highlight constraints in other sectors that should be addressed, because they impinge on agriculture's performance. Six areas are indicated, including finalizing the restructuring of local government production departments. Funding for the Sembeguya goat multiplication project is also included in the list, but it is difficult to understand why this is considered to be outside the Sector MTEF ceiling. x Section 8. Non Tax Revenue. Tables of nontax revenue (NTR) are shown for various MAAIF cost centers and the semiautonomous agencies. The table for MAAIF shows 2005/06 Budget NTR and outturn, 2006/07 Budget NTR and estimated outturn, and Budget NTR estimate 2007/08. The latter is identical to the 49 Livestock disease control and regulation are again amalgamated, and a heading appears to be missing for "Capacity building in irrigation, etc.," so 8 of the 12 DSIP areas may be represented. 76 Budget 2006/07 figure, even though outturns for some expenditure items in 2006/07 differ from the budget estimates. NARO shows 2005/06 actual, 2006/07 estimated actual, 2007/08 budget, and budget forecasts for 2008/09 and 2009/10. This approach seems more sensible and informative and might be adopted by all agencies. x Section 9. Summary of Proposed Budget Allocations for FY 2007/08. This table uses the 12 DSIP priority area classifications with the outputs, indicators, and activities that were listed in bulleted text form in Section 5. It includes targets for each indicator and a breakdown of recurrent and development budget allocations for each activity. Because different classifications are used in the tables in Section 6 (and Annex 2) and Section 9, there are apparent inconsistencies in the amounts allocated to each priority area, leading to confusion. x Annex 1. Details of Expected Outputs, Performance Indicators, and Planned Activities, 2007/08 to 2007/10. This is the only opportunity in the BFP to indicate how funds available under the MTEF ceilings for future years for each Vote would be allocated. An effective approach would be to have a summary table indicating whether any reallocation of funds between programs or subsectors within a Vote was being contemplated or whether any changes in budget ceilings would be distributed equally between programs. Justification for these adjustments could then be presented, showing the predicted effects on outputs and activities. This would appear to be more informative than the current approach in the SBFP, in which some Votes and programs give detailed accounts of intended outputs, targets, activities, and allocations of recurrent and domestic development funds over the next three years, whereas others: (i) do not indicate whether the targets and activities are annually recurring events or totals for the whole period; (ii) contain typing errors; and (iii) lack some data and totals for the programs. 77 Annex 2: Distribution of Goats for Restocking as at 30th September 2007 in NLPIP Human Household Population Poverty Allocation by indices Goats District Balance population consumption index index supplied Population Poverty Total Apac 40,5524 27,677 0.0357 0.0174 1,476 481 1957 0 1,957 Bushenyi 731,392 37,625 0.0643 0.0237 2,663 654 3317 400 2,917 Kaberamaido 131,650 34,491 0.0116 0.0217 479 599 1079 0 1,079 Kamuli 558,536 34,014 0.0491 0.0214 2,034 591 2625 198 2,427 Kamwenge 263,730 39,870 0.0232 0.0251 960 693 1653 550 1,103 Katakwi 118,928 30,560 0.0105 0.0192 433 531 964 208 756 Kayunga 294,613 48,576 0.0259 0.0306 1,073 844 1917 259 1,658 Kibaale 177,000 38,729 0.0156 0.0244 644 673 1317 400 917 Kiboga 229,472 44,547 0.0202 0.0280 835 774 1609 622 987 Kitgum 282,375 24,696 0.0248 0.0155 1,028 429 1457 0 1,457 Lira 530,527 30,094 0.0467 0.0189 1,932 523 2454 0 2,454 Luwero 341,317 43,085 0.0300 0.0271 1,243 749 1991 302 1,689 Masindi 459,490 33,085 0.0404 0.0208 1,673 575 2248 400 1,848 Mbarara 453,172 51,612 0.0399 0.0325 1,650 897 2547 500 2,047 Mpigi 407,790 59,549 0.0359 0.0375 1,485 1035 2519 1023 1,496 Mubende 423,422 39,135 0.0372 0.0246 1,542 680 2222 719 1,503 Nakapiripirit 154,494 19,691 0.0136 0.0124 562 342 905 0 905 Nakasongola 127,064 39,467 0.0112 0.0248 463 686 1148 296 852 Ntungamo 379,987 53,391 0.0334 0.0336 1,383 928 2311 690 1,621 Pader 326,,338 24,696 0.0287 0.0155 1,188 429 1617 0 1,617 Pallisa 3840,89 35,554 0.0338 0.0224 1,398 618 2016 0 2,016 Rakai 404,326 43,264 0.0356 0.0272 1,472 752 2224 873 1,351 Sembabule 180,045 45,458 0.0158 0.0286 656 790 1445 1026 419 Sironko 283,092 43,320 0.0249 0.0273 1,031 753 1783 0 1,783 Soroti 369,789 34,491 0.0325 0.0217 1,346 599 1946 0 1,946 Ibanda 198,635 51,612 0.0175 0.0325 723 897 1620 550 1,070 Kaliro 153,543 34,014 0.0135 0.0214 559 591 1150 0 1,150 Kirihura 121,219 51,612 0.0107 0.0325 441 897 1338 700 638 Amolatar 96,189 30,094 0.0085 0.0189 350 523 873 0 873 Amuria 180,022 30,560 0.0158 0.0192 655 531 1186 0 1,186 Mityana 266,108 39,135 0.0234 0.0246 969 680 1649 720 929 Nakaseke 137,278 43,085 0.0121 0.0271 500 749 1248 635 613 Budaka 136,489 35,554 0.0120 0.0224 497 618 1115 0 1,115 Oyam 268,415 27,677 0.0236 0.0174 977 481 1458 0 1,458 Buliisa 633,63 33,085 0.0056 0.0208 231 575 805 0 805 Bukedea 122,433 34,036 0.0108 0.0214 446 591 1037 0 1,037 Isingiro 316,025 51,612 0.0278 0.0325 1,151 897 2047 400 1,647 Lyantonde 66,039 43,264 0.0058 0.0272 240 752 992 616 376 Total 11,370,782 1,588,684 1 1 41,400 27,600 69,000 10,616 58,384 78 MAP SECTION