Uganda - Assessment of public expenditure programs 1996/7-1997/8

OC7F. -H7 UGANDA ASSESSMENT OF PUBLIC EXPENDITURE PROGRAMS 1996/7-1997/8 Prepared for the Consultative Group Meeting of November 1997 Paris The World Bank Uganda Country Department Africa Region October, 1997  UGANDA: ASSESSMENT OF 1996/7-97/8 PUBLIC EXPENDITURE PROGRAMS' Prepared by the World Bank for the Consultative Group Meeting for Uganda Paris, November 12-13, 1997 Overview 1. The main features of the public expenditure reform program, and achievements to end- 1995/6, were reported to the last CG. This paper presents an assessment of budget execution during 1996/7 and analyzes the proposed 1997/8 budget --which at the time of this writing had not yet been approved by Parliament. The analysis reveals that 1996/7 was a second successive difficult year for the implementation of the budget. Factors contributing to these results include spending by certain agencies in excess of budgeted amounts, the cost of implementing the Government's important universal primary education (UPE) initiative and, for the first time in over 5 years, a significant shortfall in domestic revenue. In addition, there remained the need to improve budgetary outcomes and public service delivery. While the Government has taken tough decisions to reduce or contain lower priority and non-development expenditure, these difficulties are likely to continue into 1997/8. Key contributing factors include continued high, albeit stable, levels of security related expenditures, the cost of organizing local elections and the government's desire to launch and fund important new programs while remaining within overall macroeconomic objectives. 2. In 1996/7, in the face of the revenue shortfall and strong spending pressures on non-SAs which were not offset by savings from lower than envisioned defense expenditures and payments on external debt, the Government reduced spending on key programs, albeit by only 5 percent. Even with this, the deficit was slightly higher than originally programmed. However, the impact of the higher resulting deficit was met through non-bank financing (including the issuance of promissory notes), and the GOU avoided recourse to inflationary financing. Furthermore, recognizing that greater public sector efficiency and improved public service delivery were the key to bridging the gap between sustainable macroeconomic strategies, and meeting economic and social objectives, the Government embarked on an ambitious budget reform process, supported by a third Structural Adjustment Credit (SAC III). To ensure that public and donor resources are utilized effectively, SAC III calls for introduction of greater outcome-orientation and transparency in the budget process. To minimize problems during the transition to outcome-oriented budgeting, which is a fairly lengthy process, the government is deepening civil service reform, through the planned implementation in 1997/8 of results oriented management (ROM), and has agreed to improve PIP budgeting and protect expenditures budgeted for high priority sectors and the PIP from cuts. 3. The Government's expenditure choices will continue to be severely constrained over the medium term. While the recently approved HIPC initiative should reduce the debt service burden and free up resources for high priority activities, the Government has undertaken significant large and on-going expenditure commitments (notably UPE) and has potentially large additional charges (retrenchment costs associated with parastatal reform). Accordingly, the This review was prepared by Messrs. Iradj Alikhani and Robert Blake (World Bank). Contributors included Mmes. Ritva Reinikka, Harriet Nanyonjo and Mr. Fitz Ford. The mission benefited from, and is thankful for the effective support of the Government, especially the staff of the Ministries of Finance, Planning and Economic Development, Public Service, and Local Government.  difficulties noted in this paper, especially with regards to funding the PIP and high priority economic programs, can be expected to persist. This points to the need for the Government to continue its efforts to improve expenditure prioritization and effectiveness. General Assessment of the 1995/6 and 1996/7 Budgets 4. The 1996/7 budget and its outturn and the 1997/8 budget are presented in Table 1 below. The main finding of this analysis is that there was less progress than envisioned in the original 1996/7 budget in shifting budgetary resources towards priority programs, with the notable exceptions of roads and primary education. It is hoped that implementation of the Poverty Eradication Action Plan will lead to greater prioritization of expenditures and the development of coherent sector strategies. In parallel, the process of reforming budget systems, essential to increasing transparency and reducing overhead costs, needs to be sustained so that more can be achieved within the same resource envelope. 5. The implementation of the 1996/7 budget faced a number of challenges. On the positive side, the Government should be commended for keeping supplementaries to about 5 percent of the government budget However, there was a significant shortfall in revenues of about USh 40 billions. This shortfall put considerable pressure on all programs very early in the budget year, and led to implementation difficulties throughout the year. Table 1: Uganda Budget and Outturn, 1994/5-1997/8 (USh Billion) 1994/95 1995/96 1996/97 1997/98 Category Budget Outturn Budget Outturn Budget a/ Outturn Budget Wage Bill 125 135 160 174 229 227 252 of which Defense 25 23 23 25 38 31 30 Non-Wage Recurrent 247 252 282 277 302 329 310 of which Non-Defense SAs 85 85 96 87 103 96 110 Defense 49 70 72 88 89 92 89 Non-SAs b/ 113 115 104 102 110 141 131 Statutory, URA etc. 25 37 30 54 41 49 87 Interest (Inc. External) 42 40 70 47 61 63 79 PIP (excluding taxes) 74 66 65 74 85 105 92 - Highway Maintenance 14 13 13 16 17 14 18 - Projects 60 53 52 58 68 91 74 of which Defense 4 4 5 5 9 8 10 External Debt Payments 79 82 98 56 81 68 71 Notes: a/ October 1996 version agreed with the IMF. b/ The SA and non-SA series are not consistent over time. Prior to 1996/7 a portion of the block grants to districts was assumed to be spent on SAs. For the last two FYs, recognizing that shadow budgets were often not realized, only central government resources spend on SAs have been tabulated and district flows are included in non-wage recurrent. 6. A portion of this shortfall was offset by lower than envisioned payments on external debt. Overall, the fiscal deficit was somewhat higher than originally envisioned. While inflation for 1996/7 was, for the first time in 3 years, above 10 percent, this was caused by the impact of the weather on agriculture production and not monetization of the deficit. The underlying inflation rate remained low, on the order of 2 percent.  7. Recognizing that a repeat of this experience would eventually undermine macroeconomic stability, during June-July 1997 the Government took a number of remedial actions. These include: (a) overhauling URA management and institutional set up; (b) limiting issuance of promissory notes solely to clear arrears; (c) formulating the 1997/8 budget on the basis of conservative revenue estimates; and (d) keeping supplementaries to less than 3 percent of the budget. 8. The revenue shortfalls last year contributed to the suboptimal funding for priority programs, a problem which was exacerbated further by overspending in certain major areas. Unlike in past years, over-spending by defense was not a problem in aggregate, with lower than envisioned wage expenditures more than offsetting higher than envisioned non-recurrent wage spending. However, this was more than offset by large overspending (by over 30 percent) on lower priority programs (non strategic areas; i.e., non SAs). This over-spending took the form of supplementary spending, additional to the originally budgeted levels. Spending in excess of budgeted levels was particularly striking in the cases of the State House and the National Assembly (see Table 2 below). In addition, some agencies notably Defense, issued promissory notes in 1996/7. Table 2: Supplementaries and Promissory Notes by Key Ministries: 1996/7 Recurrent Budget (USh Billion) Approved Net Promissory Budget Supplementaries2 Notes State House 10.500 12.700 1.200 Defense 127.600 3.200 12.000J Health 11.200 6.300 0 Mulago 10.400 2.300 0 Hospital National 6.700 7.200 0 Assembly 9. For the second consecutive year, spending in non-defense Strategic Areas (SAs) fell below originally budgeted levels, by about 5 percent (para. 5), although the actual impact of the shortfall was mitigated to some extent by absorptive capacity problems. In addition, the revenue shortfall contributed to the problems of expenditure management by giving rise to strong monthly variability in the level of releases. This variability added considerable uncertainty and made planning and budgeting difficult in spending ministries. However, this is a necessary 2Excluding supplementaries used to clear previous fiscal year promissory notes and net of reductions. 3Payment of these is to be covered as part of the Ministry's 1997/8 budget allocation.  consequence of the Government's reliance on the cash release system to attain macroeconomic objectives. 10. The Government has sought to address some of these shortfalls in the 1997/8 budget. The overall budgetary envelope is projected to grow by 5 percent, remaining roughly constant in real terms. Within this envelope, the Government is increasing non-wage recurrent spending in the non-defense SAs by 6.8 percent from the 1996/7 budgeted levels (equivalent to an increase of 14.6 percent relative to the 1996/7 outturn). In addition, it is to be hoped that the combination of more realistic revenue estimates and the expenditure control measures discussed earlier will lead to sectoral budget outcomes more consistent with budgeted levels than in the past. However, given efforts to protect non-Defense SAs and envisioned expenditures on local elections and UPE, the PIP appears to be significantly under-budgeted. More fundamentally, the inability of government to fully fund priority economic expenditures reflects a growing inconsistency between the overall macroeconomic program and government's desire to devote additional resources to new initiatives. It is hoped that government will give high priority to addressing this issue during 1997/8 and arrive at a budget in which all key economic programs are properly funded and executed. Some Additional Features of the 1996/7 and 1997/8 Budgets l. To ensure macroeconomic consistency over the medium term, improve expenditure programming and increase predictability, the Government developed beginning in 1995/96 a 3- 4 year Budget Framework Paper (BFP) . This paper has been the basis of the annual budget and has provided a useful context for developing the medium term implications of certain programs, i.e., the roads program. However, the BFP has not been as effective as hoped. The most serious shortcoming is the weak link between the BFP and the annual budget; decisions made in the context of the BFP are not systematically reflected in the subsequent budget much less in actual outturns. This problem was exacerbated in the case of the 1996/7 budget which was not preceded by any BFP. The BFP can play a useful role in framing expenditure choices in a medium term context but to this end the linkages between the BFP and the annual budget need to be tightened. It is expected that the Government will take steps this year to enhance the effectiveness of the BFP in the preparation of the 1998/9 budget. 12. The Cabinet put in place in July 1996 led to the creation of new ministerial positions and to the splitting the Ministry of Finance and Economic Planning into two separate ministries. While the Government believes that the enlargement of the number of ministries is justified, it has led to some additional expenditure, however modest 13. The 1997/98 budget calls for a slight decrease of the military budget in nominal terms. As some of the 1997/98 budget allocation is to be used to pay promissory notes, this implies a sharply lower level of actual defense expenditures. Given other budget priorities, it will be critical that actual defense spending be maintained within the envisioned envelope. Defense expenditure in 1997/8 will account for about 9 percent of all spending (including donor financed development expenditures). This represents a fall from the 1996/7 level, a downward trend which is encouraging and will hopefully continue to be sustained, as security concerns diminish. 4The Budget Framework Paper itself dates from 1992/3.  14. A key government objective is to improve expenditure prioritization. To achieve this it was decided in 1994 that selected high priority recurrent programs known as strategic areas (SAs) would receive substantial real increases in allocations. During 1996/7 aggregate expenditure on SAs did increase somewhat relative to 1995/6 outturns, with road maintenance and primary education being the largest beneficiaries. The Government sought in the 1997/8 budget to give priority to spending in SAs but its success will depend upon ensuring that Defense and non-SA spending is held strictly within budgeted limits. Table 3: Budget for SAs (USh Billion) SA Description 1996/7 1997/8 Budget Increase (%) Justice and Judiciary 3.67 7.52 105 Agriculture 4.24 4.24 0 Education 39.76 48.15 21.1 Health 22.92 23.00 0.4 Transport a/ 32.90 37.80 14.9 of which Recurrent 10.44 13.20 26.4 PIP 22.46 24.60 9.5 Natural Resources 0.40 0.49 22.5 Audit 1.35 1.33 -1.5 Inspectorate of Government 0.80 1.00 25.0 Police 11.89 12.31 3.7 Prison 4.78 4.78 0 Land and Housing 0.11 0.11 0 Other agencies 2.15 1.74 -18.9 Note: Funding of road maintenance program from PIP included in SA. Excludes expenditure on SAs by districts. 15. A significant medium term risk to the budget is related to the cost of funding parastatal reform. In some important individual cases, sales proceeds will either not cover divestiture costs (e.g. UCB) or revenues are deferred while certain divestiture costs, such as compensation of retrenched workers, have to be paid up front (e.g. UCB and UPTC). Furthermore, it is unclear whether the overall privatization program will generate enough funds to cover total divestiture costs, much of which are labor related. Given that the burden of the parastatal sector on the budget has been estimated at about US$150-200 million per year (20-25 percent of the budget), the need to push forward with the reforms is self evident. However, this will only be possible to the extent that associated costs can be covered through additional external support. To this end, 5Includes amounts provided for under statutorv expenditures and some wage costs.  consideration should be given to the creation of a donor-supported retrenchment fund, which could parallel the successful experience of civil service reform. Civil Service Reform 16. The pace of civil service and payroll management reform, which had slowed during the previous two years, picked up during 1996/7 and sustainability was improved. The most notable achievements include: (a) an effective payroll review and cleaning exercise; (b) the implementation of improved payroll monitoring and control system; and (c) a further review of ministerial structures to more realistic levels. 17. A payroll review exercise was spearheaded by MPS between October 1996 and February 1997. This review was motivated by the realization that, as reported to the November 1996 CG, effective payroll monitoring and control had not yet been achieved. For example, the civil service roster still included people who were no longer civil servants. Using a simple but effective methodology, MPS with the collaboration of other ministries was able to identify 2,800 possible ghost workers, about 2,000 of which were later confirmed and struck from the civil service register. MPS believes that most ghosts have now been removed and the introduction of monitoring measures discussed below should help ensure that such cases do not reoccur. 18. Since January 1997, the following monitoring system has been implemented to improve payroll control: (a) MPS regularly monitors the trend in payroll on a monthly basis, to identify any upward trend indicative of a potential problem; (b) access to Uganda Computing Services (UCS) for the purpose of adding or deleting a civil servant must be authorized by the PS of MPS; (c) department heads in each ministry are required to sign the payroll certifying that staff appearing on it are employed and physically present; and (d) with the notable exception of UPE teachers, a tight control over the payroll is being maintained by MPS and the freeze on new employment essentially remains in place --the only significant exception, which was approved by the PS of MPS, concerns the recruitment of medical workers. Despite these improvements, it is recognized that the payroll software maintained by UCS is outmoded and lacks basic security/safety features to prevent fraud or mistakes. This system is due to be replaced at the latest by 1999/2000 and donor financing for this is currently being secured. In the meantime, system shortcomings will have to be mitigated through enhanced monitoring and random checks. 19. There was a general consensus that the initial round of ministerial reviews completed in 1994/5 were unsatisfactory because it had (a) been conducted in the absence of a hard budget constraint; and (b) had been prepared in advance of the new Constitution and provisions of the Decentralization Act (March 1997). The result was to delay the process. During 1996/7 MPS revisited the reviews for 10 ministries. In the case of 6 of the ministries, service delivery objectives, as specified through the pilot results oriented management (ROM) study, were the driving force behind the new revised structures. MPS expects that the review of other ministries will be concluded and approved by government in early 1998, and that the 1998/9 central government wage budget will be based on the results of this work. The initial results and recommendations, which tend to call for strengthening of planning and monitoring functions, are encouraging. They have allowed redundant functions to be identified and have led to the gap between theoretical and actual staffing being reduced or eliminated. A significant shift in intra- departmental staffing of Ministries will be an important indicator of a successful restructuring.  20. Nevertheless, the government recognizes that further cross ministerial rationalization is needed --e.g. the Ministries of Finance, Economic Planning and Local Government each have aid coordination units-- and that fine tuning of structures will be needed during 1998/9, when the issue of service delivery priorities in the light of affordability will need to be revisited. To the extent that this has delayed certain decisions, there is a small risk that the necessary follow up will not take place, and the evolution of this process will need to be closely monitored. In addition, the issue of the rationalization of the number of ministries merits further review in order to free up additional resources to finance the PIP and strategic areas. 21. Finally, there are a number of other still unresolved/pending priorities which need to be addressed as soon as possible: (a) As reported last year, mid to high-level civil servants in the Ul(unentitled)- U5(b) pay categories received a lower overall pay increase than expected. As a result salaries for this group are on average about 20 percent less than their minimum living wage levels. However, the Government has indicated its intention to review this situation this fiscal year. (b) Rationalization of benefits and allowances and monetization of benefits was a cornerstone of the 1995 reforms. A review of the latest circular on benefits shows that the monetization process is not yet complete and that emoluments still continue being paid from the non-wage recurrent budget --albeit less than before. Completing the monetization of benefits and the elimination of those that are no longer justified should be pursued as a matter of urgency. (c) In response to the UPE initiative, districts have recruited teachers who are not yet on the government payroll. Wage arrears are thus being incurred and it is important that these additional teachers be put on the payroll as soon as possible and paid. (d) Available information shows that personnel actions are subject to delays. For example, it might take 2-6 months for someone who has resigned from the civil service to be deleted from the payroll. There is a need for MPS to ensure that the payroll system can be updated quickly and that line ministries provide the necessary information for this expeditiously. (e) In the event that the improved monitoring system discussed above reveals persistent abuses, the Government will need to formulate and implement appropriate sanctions. The Public Investment Program 22. The previous PER reported that funding of the 1996/7 PIP for projects was inadequate. As shown in Table 1, budget allocation for projects was increased by USh23 billion --the resulting supplementaries were by and large fully released. However, much of this increased allocation went to projects which were 100 percent GOU-funded, and the amount of additional funds provided to mobilize donor funding was relatively  limited.6 As a result, lack of counterpart funding remained a problem throughout the FY for some donor-financed projects. This points to the need for the Government to consider particularly rigorously projects requiring 100 percent GOU funding. 23. In addition, there was some evidence of a deterioration in investment programming processes. In a few instances public institutions signed contracts for fairly large investments -- ferry landings, Mityana-Fort Portal road -- without obtaining prior authorization for the expenditures. This is a clear violation of current procedures. To avoid paying heavy penalties, the Development Committee (DC) and the Ministry of Finance were forced to provide ex-post approval to these projects. Given that much of these expenditures were financed through domestic resources, this exacerbated the PIP resource constraints. 24. The above problems affected the content and quality of the 1997/8 PIP. Further problems stemmed from PIP financing for projects being about 30 percent below outturn for 1996/7. The prioritization of PIP projects into priority areas and others did imply that some projects were less affected by low counterpart funding than others. Nevertheless, data provided by Ministry of Planning shows that to fully fund priority project an additional USh32.1 billion would be required and that to fund other areas, which receive no funding at all, an additional UShl 5.3 billion would be needed. This performance clearly falls short of government commitment to fully fund the PIP. This problem of underfunding, evident since 1994/5, appears, if anything, to have worsened and needs to be addressed as a matter of urgency. Implementing UPE 25. In early 1996/7 the Government announced a major new initiative aimed at ensuring sustainable long term development and tackling poverty through universal primary education. This initiative was well received by all stakeholders and implementation began in January 1997. This resulted in an increase in enrollment by almost 80 percent. The program currently has and will continue to have significant public expenditure implications. Two recent studies have been completed on issues facing UPE, the key findings of which are set out below. One of the main practical issues is that faced with sharply higher enrollments the Ministry of Education needs to give priority to rapid improvement in school infrastructure, even if this means making do with temporary, less expensive facilities. 26. The main findings of the studies of implementation of UPE are as follows: (a) Educational infrastructure is inadequate. In 7 out of 8 schools there were over 100 students/class in P.1. Only 38 percent of the pupils sit on benches, and water sources, libraries and stores were lacking in most of the schools. The mean number of students per latrine varies between 48:1 to no facility at all and half of the latrines used by girls have no doors. 6Significant examples include: USh 4.23 billion for the State House; USh 10.73 billion for the Mityana-Fort Portal Road; and USh 1.95 billion in assistance of the Uganda Police.  (b) In the face of large enrollment increases, delivery of educational services has suffered. Major problems identified include: excessive numbers per classroom, lack of classroom facilities, lack of furniture (desks and tables), lack of qualified teachers, inadequate scholastic materials, delay in payment of teachers' salaries, delays in remitting UPE funds to schools, increased number of fees charged under various guises. Some groups proposed charging PTA fees of some sort, and strengthening management committees to improve accountability. (c) The partnership between government and parents continues to be perceived to be very poor, with some eligible children still remaining outside the system. The most common reason for keeping children out of school was their use for domestic work or productive labor e.g. cattle keeping and fishing. Proposals to address these problems include: the development of more innovative approaches, increased priority to disadvantaged school groups, and greater encouragement to school management committees, local authorities and head teachers to monitor drop out levels and their reasons. (d) Parents value education, but worry about the impact of UPE and free education on the quality of education. 27. The key to implementing UPE is the preparation of an education sector plan (ESP). To be effective, ESP will need to tackle the above issues. The Ministry of Education has held several consultation meetings. The process now needs to become more participatory so that the Plan to be presented to Cabinet in October and subsequently to donors for financing, reflects all stakeholders' views. The need for continued effective coordination has also been identified as another issue. The potential risk of inconsistency in donor approaches is being mitigated through a regular dialogue amongst local donor representatives. Progress on Decentralization 28. Although the progress on decentralization was broadly positive, there were some set backs during the year which had to be rectified in the 1997/8 budget. Among the positive elements was the passage of the new Local Government Act in March 1997. This Act incorporated the promised change in the local government financial year to coincide with the central government's. This eventually will contribute significantly to planning, programming and monitoring of expenditures on priority programs. Districts have had difficulties in the past preparing final accounts partly as a result of the anomaly in the financial years. For the FY ending June 30, 1997, this problem will be aggravated by having to close their books in the ninth month of their approved budgets; however, measures to assist them are being taken by the Decentralization Secretariat. The Auditor General's Department has undertaken to complete audits of these accounts within the four months prescribed under the Law. 29. Another important achievement was the completion of the study of the decentralization of the development budget and the preparation of recommendations to be presented to Cabinet early in 1997/8, for implementation in 1998/9. This coincides with commitments made under the SAC III, and will mark a major step forward in the decentralization process.  30. The major difficulties which arose during 1996/7 resulted from back tracking on commitments to de-link block grants from the wage bills of districts, and from limiting releases of these grants to about 82 percent of authorized budgetary provision. The combined effect of these actions was to direct local governments expenditure choices increasingly towards wages and away from programs. Primary health care services appear to be among those which suffered as a consequence. However, in the 1997/8 budget a formula has been used for the distribution of block grants which is de-linked from wages. Greater use of conditional grants and integrated budgets by local governments should also contribute to more rationalized expenditure programs in this year. 3 1. It is critical for the credibility of the decentralization program that budgeted funds be released promptly. To ensure that there is greater transparency in this process, releases to districts are being published monthly in the press. However, in 1996/7 the Ministry of Finance had to reduce unconditional grant releases to transfer the equivalent amount to conditional grants to fund UPE, reducing budgetary flexibility at local levels. Over the near term, top priority needs to be given to improving accounting and budgeting systems at the local levels to permit increased accountability in the use of unconditional grants. 32. At the level of the districts, a minority of districts (15) have approved budgets which are in deficit, in clear violation of the Local Government Law. In addition, most districts appear to have overestimated local revenues in violation of budget guidelines, probably in attempts to circumvent the law as well as to permit greater spending on overheads (which are limited to 10% of local revenues). The Decentralization Secretariat is proposing that joint teams from the Secretariat, MLG and the Ministry of Finance review these budgets and where deficits exist and overestimates would result in deficits, districts would be required to revise and reapprove legal budgets. It is critical that appropriate sanctions be applied to districts that fail to do so promptly. The Poverty Eradication Action Plan 33. The Government's Poverty Eradication Action Plan (PEAP) focuses on four main areas: macroeconomic policy, the institutional framework, policies to increase the incomes of the poor, and measures to improve their quality of life. The document represents the outcome of a process of consultation which lasted almost two years and involved government ministries and agencies, donors, NGOs, and local academicians. While the PEAP will only start being implemented in 1998/9, a number of public expenditure issues have already emerged from the process. 34. First, the share of priority basic services in total recurrent (including donor) spending has been estimated at 23 percent in 1996/7. The Plan envisions raising this share to 50 percent over the medium term. This will require freezing and reducing of the rest of government spending. Most of the increased spending should be at the district level. Although most donor spending is included in the PIP, the role of sector strategies needs to be strengthened. This is a necessary condition for effective expenditure prioritization. Once adequate sector strategies exist, donor funding of recurrent expenditure may in many cases be more effective than capital spending. 35. Second, the proposed mechanism for achieving national priorities at the district level is a defined package of basic services available to all Ugandans, to be supported by dissemination of the package to create bottom-up pressure (as has happened in primary education), and conditional grants to districts for specific services. Initially, the basic package includes agricultural extension, primary health, feeder road maintenance, and primary education. In the  long run, through appropriate use of conditional and equalization grants and with the effective decentralization of the development budget, districts should be free to prioritize within basic services but not to divert away from them. Currently, the Government is committed to free primary education for most children. In primary health care there is no such policy. In practice neither private nor public health services of decent quality are typically available or affordable to the poor. Feeder roads and the extension services are mainly publicly funded, although some good commercial private extension services are available. 36. Costing the package is key. So far, costings are most clearly developed in roads and agricultural extension; the existing envelope does not fall too much short of the requirements of the minimum package. But the actual service delivery needs major improvements. Costing of an appropriate health package is badly required. Preliminary costings exist for education but there are some elements which may be flexible, such as school structures. 37. Third, the Plan stresses mass service delivery rather than targeting subgroups of the poor, but there is some scope for regional and gender targeting. Incentives for staff to go to remote areas do not exist at the moment, and prioritizing roads in poorer areas needs to be re- examined. 38. Fourth, public funding need not imply public provision. For example, the Ministry of Health is hoping to administer funds through community sick funds which procure care for their members. In addition, the Government has been extending grants to private health units for provision of services. Flexibility and cooperation at district levels between govemment and NGOs will be important for delivering the services efficiently. 39. Finally, fine-tuning the prioritization of public spending to protect basic services from within-year fluctuations and spending cuts is important. Improved monitoring of outputs and outcomes is badly needed, including better publication of the monitoring results. Immediate priority is monitoring quality as well as quantity in primary education. Central Government Accounting and Transparency Issues 40. The findings of past public expenditure reviews and work undertaken in the preparation of the future Economic and Financial Management II (EFMP II) project have identified many shortcoming in the government accounting, expenditure tracking and reporting systems. It is clear that some of the needed capacity will take time to be built. But, existing systems can be improved and the AG's report for the year ending on June 30, 1995, which became available recently, identifies some the issues and suggests possible solutions. This report is the most comprehensive and up-to-date in recent memory and is the fruit of a commendable effort by Government to prepare and make available to the AG for his appreciation government accounts, in some cases for the first time in several years. Even though the report covers the period preceding the review herein, some of its key findings and recommendations remain topical and merit to be summarized here, in the hope that they will be followed up by government and other stakeholders: (a) Transparency of Classified Expenditure - About 10 percent of the recurrent spending in 1995/96 consisted of unaudited classified expenditure. The AG  points out that such expenditures need to be audited, but does not propose a system for doing so. (b) Domestic arrears - Based on the analysis of a sample, the AG estimates that about USh 62 billion of non-salary arrears were accumulated during thie preceding 3-4 years. Some of these were with public utilities. The problem is now being partially addressed through debt settlement. However, given that in the medium term some utilities will be private, expenditures on utilities need to be budgeted properly to avoid accumulation of new arrears. (c) Accounting for Cash and Stock - The AG notes that only 12 Ministries have submitted accountability reports for the former and 10 Ministries/departments out of 40 have accounted for the latter. Part of the problem is inadequate staffing at the Treasury Office of Accounts which has hindered the work. (d) Doubtful and Due Revenue - The report provides many instances where revenue collected by public agencies/Ministries is not paid to the consolidated account. The problem seems to have peaked in 1993/4. (e) Excess Expenditure - Unauthorized expenditures amounted to over USh 2 billion. This is symptomatic of problems at UCS, which controlled expenditures globally and not by line items. (f) Financial Statement by Treasury - The AG points out inconsistencies in the accounting report, which was presented for the first time by the Treasury. While the situation has improved and Treasury officials are looking into the discrepancies, they believe that the problems encountered are fundamentally caused by lack of integrated government accounting and budgeting, especially the absence of double entry book keeping. 41. The above issues are expected to be addressed as part of the review of the audit report by the Parliamentary Accounts Committees (PACs). The Future Agenda 42. In spite of difficulties, which have slowed down the rate at which the budget is being reoriented towards high priority programs, a number of key initiatives have been announced and are in the process of implementation. They include: (i) UPE; (ii) the 10 year highway program, presented to the donors in November 1996; (iii) PEAP; (iv) decentralization of the development budget envisioned for 1998/9; (v) the introduction in 1998/9 of output/outcome oriented budgeting for three pilot ministries, Agriculture, Health and Education, which would put greater emphasis on optimizing public service delivery; and (vi) strengthening of public accounting and auditing and improved dissemination of information about public prograins and their results, a process that has already started. 43. These new programs if successively implemented should substantially improve the quality of public expenditures. However, implementation of these initiatives will not  be costless. In light of this, the Government will need to strengthen substantially existing mechanisms for expenditure prioritization and reinforce procedures for budgetary execution to ensure a better initial allocation of resources and budgetary outtums more consistent with initial budgetary provisions. The Government's Budget Framework Paper (BFP) can play a critical role in this regard, as a vehicle for identifying and resolving budgetary choices. To be effective, though, the BFP needs to be linked closely with the preparation of the FY 1998/9 budget, so that the budgetary choices are clearly made. This implies in turn more careful preparation of the BFP so that policy options are identified and costed. Finally, there is a need to reinforce the cash release system to ensure that expenditures are closely linked to budgetary allocations. In this regard, the Government's decision to limit supplementals and restrict the issuance of promissory notes to paying arrears is to be welcomed. These decisions highlight, though, the necessity of ensuring acceptance of budgeted levels. 44. In conjunction with this effort, continued donor support will be critical. The speed with which the Government will be able to tackle its development agenda will depend on resources that are available to it. On the donor side, the reform agenda is expected to be supported through budget support and capacity building initiatives. This should help ensure the financing of priority programs. 45. A last point meriting to be highlighted is that as government budgeting shifts towards defining expenditures in terms of programs, the way in which donor support is provided would also need to evolve. The principle modification would involve government defining a medium term expenditure program for each key sector, which defines investment as well as incremental recurrent cost requirements. Each donor, as is already the case for roads, would then finance a subprogram within the overall program. This approach which should get underway during 1998/9, would have implications not only for future projects, but also for some ongoing ones.