Uganda - Northern Uganda Reconstruction Project

Document of The World Bank FOR OFFICIAL USE ONLY Report No. 20664 PERFORMANCE AUDIT REPORT UGANDA NORTHERN UGANDA RECONSTRUCTION PROJECT (CREDIT 2362-UG) June 27, 2000 Sector and Thematic Evaluation Group Operations Evaluation Department This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization. Currency Equivalents Currenc Unit = Uganda Shilling (Ush) Average for US$1.00 month of: equals Ush: Jan 1992 942 Jan 1993 1,205 Jan 1994 1,107 Jan 1995 896 Jan 1996 994 Jan 1997 1,013 Jan 1998 1,148 Jan 1999 1,369 Jan 2000 1,523 Abbreviations and Acronyms AG - Auditor General (GOU) CDF - Comprehensive development framework CMEU - Coordination, Monitoring and Evaluation Unit (in OPM) GOU - Government of Uganda (=NRM) HIPC - Heavily Indebted Poor Countries (Initiative) IG - Inspector General (GOU) LRA - Lord's Resistance Army (rebel movement in Northern Uganda) MOF - Ministry of Finance MOI - Ministry of Interior NGO - Nongovernmental organization NRM - National Resistance Movement (Political party of the Museveni regime) NURP-I - Northern Uganda Reconstruction Project (the subject of this audit) NURP-II - Second Northern Uganda Reconstruction Project (proposed follow-on operation) OED - Operations Evaluation Department OPM - Office of the Prime Minister PIU - Project implementation unit SOE - Statement of expenditure UNDP - United Nations Development Programme UPDA - Uganda People's Defense Army USAID - United States Agency for International Development UVAP - Ugandan Veterans' Assistance Programme Fiscal Year July 1 - June 30 Director-General, Operations Evaluation : Mr. Robert Picciotto Director, Operations Evaluation Department : Mr. Gregory Ingram Manager, Sector and Thematic Evaluation : Mr. Ridley Nelson (Acting) Task Manager : Mr. Roy Gilbert FOR OFFICIAL USE ONLY The World Bank Washington, D.C. 20433 U.S.A. Office of the Director-General Operations Evaluation June 27, 2000 MEMORANDUM TO THE EXECUTIVE DIRECTORS AND THE PRESIDENT SUBJECT: Uganda: Performance Audit Report Northern Uganda Reconstruction Project (Credit 2362-UG) Attached is a Performance Audit Report (PAR) prepared by the Operations Evaluation Department on this project for which the World Bank approved a Credit in an amount of SDR 51.9 million (US$71.2 million equivalent) May 5, 1992. The Credit was closed on September 30, 1998 as planned. Co-financing was provided by the Government of the Netherlands (US$6.5 million) and DANIDA (US$2.1 million). Counterpart funding by government of Uganda was US$15.9 million. This was an ordinary development project that produced some extraordinary results in unusual circumstances. The experience of the Northern Uganda Reconstruction Project (NURP-I) provides a very rich source of lessons for future operations. Positive impacts upon local develop- ment were achieved, but undermined by inefficiency and corruption, although the project itself successfully took action against those accused and further investigation is under way at the time of writing. NURP-I was treated as a post-conflict operation by both the government of Uganda (GOU) and the Bank itself, even though the project was prepared, appraisal and implemented while conflict was on-going. Since 1988 to the present day, Northern Uganda, the last region of the country to be controlled by the National Resistance Movement led by President Museveni, has suffered conflict at the hands of the Lord's Resistance Army (LRA) rebels, based in the southern Sudanese city of Juba. GOU's military "Operation North", just prior to appraisal, failed to defeat the rebels. Conflict and the threat of it persisted throughout implementation and even prevented the audit mission's planned visit to the regional capital Gulu in January 2000. Project achievements were the result of a combination of good luck and unsullied dil- igence by many dedicated project staff in the region itself and in Kampala. A follow-on NURP-II project has to be more tightly designed for better control. It would be unwise to rely upon contin- ued good fortune or exceptional assiduity by project staff that is not only unrecognized, but even doubted. The project had significant intended and also unexpected results, some good, some bad. Among the intended results was the restoration of parts of the regional economy. While the North still remains the poorest region of a poor nation, there is evidence of economic improvement in those parts of the region visited by the audit mission, notably Lira and Arua, which have become bustling market towns. Even though North-South economic disparities may not have been reduced as the project intended, poverty in the Northern region fell by as much as in the rest of Uganda over the 1991-97 period. Security did not improve, nor did defense expenditures fall, as the NURP-I project designers hoped. A better understanding of the North as This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization. -2- a theater of on-going conflict might have led to reasoned expectations of continued violence and increased military spending, as in fact occurred. The project succeeded in involving women, as intended, although without the special attention that their status as war-widows, divorcees and single parent families deserved. NUPR-I impact upon building municipal capacity was insignificant, but not surprising for an operation that was highly centralized in Kampala. Nor did the project succeed in fostering much donor interest in the North. NURP-I alone is not responsible for this result. Donor trust and confidence of operating there has to be restored. For that, peace has to be brought to the area, and GOU must be the main peacemaker. Unexpected results included the empirical demonstration that management of on-going conflict situations is not where the Bank has comparative advantage in expertise. Nor did the Bank seek expert advice elsewhere. Thus, NURP-I overlooked key issues related to conflict, such as demobilization, security conditions required for private business to resume, or targeting poverty reduction interventions on potential combatants to increase the opportunity costs of participation in armed rebellion. The PAR endorses the finding of OED's 1998 review of Bank experience with post-conflict reconstruction that the Bank is not the agency that is best equipped to assess and decide about macrosecurity issues. The project's Kampala/Northern Region 'disconnect' that separated financial control in the capital from physical implementation in the remote region was an inducement to corruption. Cost overruns of 50% meant that NURP-I primary schools cost twice as much as similar structures in neighboring Rwanda. Whether this was the result of inefficiency or corruption remains to be clarified by on-going investigation. Another unexpected result was that the NURP-I project experience itself provided a case study in fighting corruption. Toward the end of implementation, use of the project's own control instruments-audits, spot checks by monitors, aide memoires, task manager oversight-identified a case of misuse of project funds that led to the indictment of a former project manager, through the Ugandan justice system. Further investigations are underway, triggered by allegations by a former NURP-I service provider. At the time of writing, results of the investigation are pending. In addition to ICR lessons which refer to the need for more decentralized implementation arrangements, more stakeholder participation in design and implementation, a clearer definition of target group, closer attention to post-conflict measures, and tighter monitoring and evaluation, the PAR highlights some important issues arising from the project that are relevant to Uganda and other countries. They include: (i) the Bank brings little comparative advantage to the management of situations of on-going conflict; (ii) project designers and managers have important responsibil- ities as partners in anti-corruption efforts; and (iii) the proposed follow-on project NURP-II, that should be supported once security conditions in the North improve, should incorporate all the anti-corruption lessons of NURP-I, particularly: (a) decentralization of project management and control to bring together financial management with physical implementation; (b) a decentralized design should include district administrations, local communities and final beneficiaries in project planning and implementation; (c) transparency of basic project information-especially costs- throughout the region, with hot-lines and help-desks to denounce abuses; (d) routine investiga- tions of allegations of abuse by project stakeholders; (e) encourage press and NGOs as watch- dogs; (f) publicly recognize and reward honest and diligent project staff and stakeholders., too, to help build trust. The PAR rates the outcome of the project as marginally satisfactory. The efficiency of achieving some important development results was undermined by overspending, some of it associated with alleged corruption. Also, NURP-I only partially achieved its objectives related to improving security, building local government capacity and stimulating other donors to assist the North. Hence the lower rating than the ICR/ES of satisfactory. The PAR also downgrades the Bank performance rating to unsatisfactory from satisfactory reported by the ICRIES. The Bank's -3- inadequate understanding of and response to the on-going conflict, its contribution to a project design with poor control and accountability and hands-off supervision during the first half of implementation contribute to that assessment. Other ICR/ES ratings remain unchanged by this PAR, namely uncertain sustainability, modest institutional development impact and satisfactory borrower performance. With regard to the last rating, the PAR considers that the outstanding merit of the government of Uganda's (GOU) anti-corruption pursuit of high level officials through the courts deserves recognition, even though the borrower itself contributed to poor project design and implementation and an inadequate assessment of the security situation facing the North. Attachment  1 Contents Principal Ratings ...........................................................................................................................iii Key Staff Responsible ...................................................................................................................m Preface ............................................................................................................................................. v 1. Introduction ............................................................................................................................. 1 2. Project Design and Implementation................................................................................. 2 3. Intended Outcomes ........................................................................................................... 4 Restoring the Economy .................4............ .................4 Reducing North-South Disparities ................................................5 Security and Defense Expenditures................6......... .............6 Gender Aspects ......................................................7 Municipal Capacity Building ................................. ........... 7 Fostering Other Donors in the North ....................... ..... ........... 8 4. Unexpected Results .......................................................................................................... 8 Partial Treatment of Ongoing Conflict.................................8 A Project Design to Induce Corruption ............................ 10 A Case Study of Fighting Corruption. ..................................... 13 5. Conclusions and Lessons .................................................................................................. 15 Ratings ........................................................... 15 Bank's Comparative Advantage Is Not in Dealing with Ongoing Conflict ..................15 The Responsibilities of Project Designers and Managers for Anti-Corruption............16 Some Elements for the Profile of a NURP-II Operation, Once Peace Is Achieved ..........17 Bibliography. .....................................................19 Annex: Basic Data Sheets.... .........................................20 Attachment: Final Project Costs by District................. ................22 Figures Fig. 1 Northern Uganda: share of households living in poverty .................5 Fig. 2 Uganda: military expenditures ......................... ......... 6 Fig. 3 Unit construction costs of primary school classrooms......................11 Boxes Box 1: Northern Uganda: anatomy of ongoing conflict................................1 Box 2: NUPR-I CAP gives Betty 'Best Girl' Award ... .. ........................7 Box 3: Evaluating primary schools in Arua..............................12 This report was prepared by Roy Gilbert (Task Manager), who audited the project in January 2000. William B. Hurlbut edited the report. Romayne Pereira provided administrative support.  111 Principal Ratings Northern Uganda Reconstruction Project (Credit 2362-UG) ICR ES PAR Outcome Satisfactory Satisfactory Marginally Satisfactory Sustainability Uncertain Uncertain Uncertain Institutional Development Modest Modest Modest Bank Performance Satisfactory Satisfactory Unsatisfactory Borrower Performance Satisfactory Satisfactory Satisfactory Key Staff Responsible Task Manager Division Chief Country Director Appraisal Gemhard Techanneri Stephen Welssman Francis Colaco Completion Iraj Talai Jeffrey Racki James Adams  V Preface This is a Performance Audit Report (PAR) on the Uganda: Northern Uganda Reconstruction Project (Credit 2362-UG) for which the World Bank approved a credit in an amount of SDR 51.9 million (US$71.2 million equivalent) on May 5, 1992. The credit was closed September 30, 1998, as scheduled. Cofinancing for the project was provided by the Government of the Netherlands (US$6.5 million) and DANIDA (US$2.1 million). Counterpart funding by the Government of Uganda was US$15.9 million. This report is based upon reviews of the Implementation Completion Report (ICR), the Staff Appraisal Report (SAR), legal documents, project files and discussions with Bank staff involved with the project. OED fielded a mission to Northern Uganda January 2000 to review the project results. The mission visited central government departments and project agencies in Kampala and Entebbe but spent most of its time in the Northern Region itself, visiting the towns of Arua, Soroti and Lira and their surroumding districts. The audit mission appreciates the courtesies and attention given by interlocutors in the region and in the capital as well as the logistical support provided by the Bank country office in Kampala. The mission is particularly grateful for the efforts made by the NURP-I team in the Prime Minister's Office to provide, for the first time, detailed estimates, by district and component, of expenditures incurred through the project. A summary table of data produced by that work is attached to this report. Following standard procedures, copies of the draft PAR were sent to the relevant government officials and agencies concerned for their review and comments, but none were received.  1 1. Introduction 1.1 On the surface, the Northern Uganda Reconstruction Project (Cr2362-UG), known as NURP-I, looked ordinary enough, but it had some extraordinary results which represent a very rich source of lessons for future operations. Positive impacts on local development were achieved, but undermined by corruption, although the project itself successfully took action against those accused and further investigation is underway at the time of writing. NURP-I was designed and implemented in unusual circumstances. Treated as a post-conflict reconstruction effort by both the Bank and the government of Uganda (GOU), the project was nevertheless conceived, implemented and completed while conflict was still ongoing, with tragic results for contractor crew members killed by rebels in the Kitgum district in 1996 and mixed results for peace and economic development for the Northern region as a whole. A combination of good luck and unsullied diligence by many dedicated project staff in the region and in Kampala brought Box 1: Northern Uganda - Anatomy of ongoing conflict some development effectiveness. The North remains the last area of Uganda still to be fully For follow-on work to be pacified and incorporated into the political and economic model successful, though, the Bank of the National Resistance Movement (NRM). Although conflict cannot continue to rely on good between the North and southern Uganda has historical roots, the fortune or assiduity by project failure of the Gulu peace accord of 1988, which led to the staff that is not only unrecognized creation of the dissident Lord's Resistance Army (LRA) guerilla butmovement-based in the southern Sudanese city of Juba, was a but vendoubed.NURPI ~determining moment for the conflict and, indirectly, the fortunes mistakes and attempts to resolve of NURP-I. The 1989-91 period-as NURP-I was under them provide valuable learning preparation-saw intensive rebel activity. An estimated one-third inputs for the proposed follow-on of the region's population was displaced from their home villages, project NURP-II to succeed in a while the NRM rounded up 75,000 civilians into six "protection" much less risky fashion. camps set up to deprive the guerillas of succor. Aiming to defeat the rebels militarily, the NRM embarked on a major counter- 1.2 NURP-I itself was born in insurgency "Operation North" in mid-1991. Despite its failure, unpropitious circumstances of the "stick" of Operation North was followed by the "carrot" of war. Uganda's Northern region- NURP-1. LRA rebel activity continued, as did counter- the project area-extended across insurgency, although peace talks were held in 1994, but once aagain broke down in the face of an ultimatum from the NRM for a lage sath of and350k by the insurgents to lay down their arms and surrender. Some of the 200km covering one third of the worst LRA massacres ensued during 1996/97, including savagery country, and bordering volatile in the mass abductions of children widely publicized in the Sudan and Democratic Congo. international media. After a comparatively peaceful interlude of With its towns 200-400km distant some 18 months during 1997-99, the LRA insurrection resumed from Kampala, the North is at Christmas 1999 with attacks in Gulu and Kitgum and continues remote from Uganda's capital. Its to the present day. Even experts today are hard pressed to transport and communication interpret the goals of the idiosyncratic LRA movement. The infrastructure is poor. Guerilla growing international dimension of the conflict-in which activities and military counter- Uganda supports a reciprocal rebel movement of the Sudan insurgencyPeople's Liberation Army (SPLA) in southern Sudan across the insugenc wee an reain ife border from the North-bodes ill for an early peace in the North, and constitute a constant threat. At given the overt support given by the international community to the time of NURP-I appraisal, six NRM efforts to contain Sudan. Meantime, the population districts-home to about two- continues to suffer from the insecurity and indiscriminate thirds of NURP-I investments- violence wrought by this ongoing conflict. were considered unsafe to visit. For more details, see: Ruddy Doom and Koen Vlassenroot%ony's message: a new koine? The Lord's Resistance An in Northern U anda in African Ars Vol. 98 No. 39 co Januarf 1999. Theeortareainathelas arasosUgnda,tilttobeiull 2 mostly in Kampala, was fielded almost immediately after a major government counter-insurgency campaign called "Operation North" aimed, without success, at the military defeat of the guerillas. Some eight and half years later the audit mission itself could not visit the regional capital Gulu, under attack by rebels of the Lord's Resistance Army (LRA) in early 2000 as it had been throughout NURP-I project implementation. Without the conflict, the North-as the poorest region of a poor country-would have enjoyed a high priority within the Bank's poverty reduction strategy. It is only to be hoped that this conflict will soon cease so that the needy poor, who have long suffered from the turmoil of this region, may benefit from stable economic and social development. 1.3 The 1990s were a promising time for Uganda's economy as a whole, but not for the North. The government of Uganda, called the National Resistance Movement (NRM) under President Museveni, earned a solid reputation among donors for good macroeconomic management and governance. As a sign of support of national policies, the Bank agreed in 1998, that Uganda would become one of the first pilot countries to adopt the Comprehensive Develop- ment Framework (CDF) proposed by Mr. Wolfensohn. In 2000, Uganda became the first country to qualify under the second stage of the Bank/IMF Heavily Indebted Poor Countries (HIPC) init- iative, under whose first stage debt relief of US$2.0 billion was granted. Throughout NURP-I implementation during 1992-98, Uganda's GDP grew at an annual rate of 7.1%. Reliable region- al GDP data for the North is not available. Before the project, in 1991, Uganda's net capital flight was US$17.3 million and net foreign investment was only US$1.0 million. By 1997, toward the completion of NURP-I, Uganda had a net capital inflow of US$311.3 million, while net foreign investment soared to US$180.0 million. Despite these national achievements, the North continued poor and excluded from the remarkable economic progress reported by the NRM. 2. Project Design and Implementation 2.1 NURP-I was a multisectoral operation with nine components, each prepared as a separate mini-project by a GOU line ministry and Bank sector specialists.' This audit focused upon five largest components that, by completion, accounted for 82% of all expenditures: namely, community action program (US$6.5 million), education (US$19.2 million), highways (US$32.1 million), urban (US$13.5 million), and water supply and sanitation (US$9.6 million).2 The audit mission planned to visit three project towns-Gulu, Soroti and Lira-whose districts accounted for 60% of all NURP-I investments. Rebel attacks on the regional capital Gulu while the mission was in Uganda, forced the audit to substitute Arua, bringing the share of project costs incurred in districts visited by the audit mission down to 42% of the total. 2.2 Lack of security like this was a problem throughout project preparation and implementation. While project files rarely provide many details, Bank staff and consultant forays 1. The Region further clarified that: "Not only the project was appraised by a team of different sector specialists from the Bank, it was also supervised in the same manner, under the overall coordination of one task manager. It is noteworthy that, since sector specialists from different units did have their own work program and mission plans, it was difficult, if not impossible, to include all specialists for each supervision mission. In fact, in many cases, each specialist pianaging her /his relevant component fielded missions and supervised the project component at different times and issued an aide-memoire delivered to the sector ministry (with copy to the Prime Minister's office)." 2. The remaining four components were: agriculture (US$5.0 million), feeder roads (US$2.7 million), telecommunications (US$0.3 million), and sub-projects (US$.7 million). For further details and project costs by district, see cost tables attached to the annex of this report. 3 into the North itself were rarely more than a small part of mission time.' None of the many district officials contacted by the audit mission in the North could recall previously meeting any Bank officials. The appraisal mission of 17 Bank staff and consultants visited Kampala for 28 days in September 199 1,4 when it was only safe to visit four district in the North. Of 11 supervision missions, three did not venture out of Kampala at all, while the remainder could field only brief excursions to the North. Things were no easier for the Ugandan staff. On one visit to Soroti in 1995, for instance, NURP-I's manager from Kampala reported how he had to be rescued by district officials from the wrath of angry protestors complaining about inadequate project allocations for their area. 2.3 In such conditions, Bank project preparation was inevitably top-down. Neither Bank staff nor GOU officials were in listening mode in determining priority investments for the region. A large, expensive and useless hole in the ground in a residential neighborhood of Arua is one result. It marks the place where NURP-I experts from Kampala wanted to locate a sewerage treatment plant against the wishes of the district authorities and local residents. NURP-I's modus operandum for planning of this and other investments was for line ministries in Kampala to propose subproject types, while the Ministry of the North-located variously in Gulu or the Office of the Prime Minister (OPM) in Kampala-would determine where to locate them within the districts. With little field presence, Bank supervision itself was remote from on-the-ground activities. Worse still, the project Coordination, Monitoring and Evaluation Unit (CMEU) in Kampala was reluctant at times to send OPM project officials into the field, prompting official correspondence from Bank management to CMEU requesting more diligence in project oversight. 2.4 NURP-I's design was, therefore, highly centralized. Of course, it was prepared before the 1997 Local Government Act that today places prime responsibility for service provision to the poor on the shoulders of local authorities in Uganda.s Even in 1991, however, NURP-I had a declared aim of building local government capacity in reconstruction and rehabilitation. At the time of appraisal, the Bank's country office in Kampala argued that districts should play a key role in NURP-I, since the weaknesses of local administration had been exaggerated, and the strength of central government exaggerated.' GOU, however, preferred the centralized arrange- ment, to which Bank management eventually deferred.' The resulting design may be faulted, among other things, for separating financial management and control (in Kampala) from physical implementation (in the North), thereby opening up a potentially rich arena for corrupt gains. 2.5 Notwithstanding the extraordinary circumstances of war and exclusion described earlier, NURP-I had-with the exception of those concerned with security-strikingly ordinary 3. The Region considers that the PAR does not sufficiently take into account the efforts of missions of other than those by the project core team in visiting the North, citing the following examples: "(i) For the first three years of the project an architect was actively involved in the design of schools and supervision of first schools to ensure the adequate understanding and execution of the designs; (ii) the Agriculture component was mostly supervised by a specialist from the Resident Mission who traveled frequently to the North; and (iii) the highways component was mostly supervised by staff together with other roads project under implementation at that time." 4. It is unlikely that all mission members stayed the full 28 days. That would imply 68 staff-weeks just for the appraisal mission, while the ICR reports a total of only 45 staff-weeks for the entire appraisal/effectiveness effort. 5. See: Roberta Mutiso and Deborah Kasente Improving the Performance of the Country Programme in Reaching the Urban Poor Government of Uganda/UNICEF Country Programme 1995-2000. 6. Country office memo to HQ in June 1991. 7. In September 1995, shortly after midterm, the Bank tried unsuccessfully to decentralize NURP-I by moving the project coordination from Kampala to Gulu, but such a move was resisted by the Minister in Gulu and by CMEU/OPM in Kampala, whose point of view prevailed. 4 development objectives: namely, restore economic productivity, reduce historical disparities between the North and south of the country, improve security and reduce defense expenditures, reduce poverty and address the needs of women, build capacity for local government, stimulate other donors to finance projects in the North. Individual components did not have individual goals designed to meet specific objectives or address specifically conflict-related matters. Rather, project designers hoped that, collectively, the multi-sectoral components would contribute to the enhanced development mapped out by the operation's explicit objectives. By good fortune and the diligence of many key players, some of the intended outcomes sought by these objectives were achieved. In addition, this report looks at some unexpected project results-good and bad- that the project design had not foreseen at the outset. 3. Intended Outcomess Restoring the Economy 3.1 Despite economic progress in recent years, Uganda is still one of the poorest in the world with a GNP per capita of US$330 in 1997, compared with US$260 in 1991, prior to NURP-I. The North still remains the poorest region of a poor nation, but there is evidence of improvements in the regional economy, at least in those parts visited by the audit mission. 3.2 In the towns of Lira and Arua, for instance, local commerce was very active. In Lira, now an important market town, the urban market provided under NURP-I could not accommodate all traders and was overflowing with goods. Stallholders reported brisk business in response to audit mission inquiries. Some modest new construction and repair of existing building are signs of a reviving urban economy, stimulated by the substantial improvements NURP-I brought to the downtown through paved streets, drainage and better highway access. Cross-border trade with Democratic Congo, whose frontier is just 8km from Arua, was responsible for a busy local economy in the town, helped by having the region's only daily air connections with Kampala/Entebbe. NURP-I's own contribution to Arua, where it spent five times less than in Lira, was much less, though. The principal component in Arua-a water treatment plant costing US$459,000-was not operational for lack of gasoline to fuel the water pumps, leaving the entire town dry and still dependent upon boreholes even in the downtown area. 3.3 NURP-I significantly improved Soroti through major drainage works, urban street paving and a large market. But a bustling urban economy was not the impression gleaned by a brief visit to the town, which had previously been a major center of Ugandan-Asian business, yet to be resumed. The audit cannot report on the economy of Gulu, the regional capital and principal economic hub of the region, since on-going conflict prevented the mission's visit there.' 3.4 A US$13.7 million investment to pave the 93 km Kafu-Karuma main highway outside the region but linking it to the south, was reported to have allowed the near quadrupling of traffic, 8. It is worth repeating that the assessment here is based upon observations during audit visit missions to three districts of the Northern Region where 42% of project expenditures were made, and where some significant development impacts were achieved through the useful exploitation of infrastructure provided by the project. The two most insecure districts, where field visits were not possible, namely Gulu and Kitgum, accounted 27% of total project expenditures. 9 At the request of the audit mission, the project team prepared detailed cost tables by district which showed total NURP-I investments of US$18.2 million in Gulu, or 23% of total project costs. Completed components reported there were: agriculture (US$1.8 million), education (US$2.3 million), highways (US$10.3 million), urban (US$3.2 million), water supply (US$0.6 million), 5 despite frequent interruptions due to security checks. The ERR estimated at completion was, nevertheless, only 13%, half the appraisal forecast, owing to a cost overrun of nearly 50%. The economic impact of NURP-I investments of US$18.4 million in highway improvements within the region was less. ERRs reported by the ICR averaged only 12%, comparing unfavorably with 40% for road transport projects elsewhere in Uganda.o The case of the 113km Soroti-Lira road was particularly disappointing, as it constitutes a potentially important link in Uganda's national highway network. Today, the gravel surface provided under NURP-I and convex cross-section for drainage are practically non-existent. Despite that, the ICR reported an investment with a 7-year useful life and benefits still flowing today even though the improved operational conditions no longer existed at the time of the visit by the audit mission. A breakdown of maintenance equipment was cited by MOPW as the cause of the failure of the Soroti-Lira road. The shortened life of this investment, together with much lower daily traffic-162 vehicle per day according to MOPW, instead of 240 per day reported earlier-implies a much lower ERR than the 9.2% reported in the ICR. The audit mission's own ground observations at both ends of this highway and a noontime over-flight of its entire length found poor road surface conditions and minimal traffic flows. The Kumi-Soroti highway, on the other hand, was in much better condition and carried more traffic, even though it is of lesser importance in the national network. 3.5 The repatriation of capital, restoration of seized assets and inducements to resumed private investment must remain at the heart of the economic recovery of Uganda's Northern Region. NURP-I neither addressed these issues nor prepared the groundwork for when they might be addressed in the future. Evidently, the complete restoration of peace and security-yet to be achieved in this region-is a precursor for private investment to resume. Reducing North-South Disparities Fig. 1 Northern Uganda: share of households living in poverty 3.6 In promising to reduce these disparities, NURP-I set itself 80 a tough and challenging task of 70_ many dimensions. Here, we look at the poverty and social aspects, where the audit found some 50 0199 satisfactory results. The good news 40 is that poverty reportedly fell in the -S30 North at almost the same rate as in 20 Uganda as a whole over the 1992- 10 97 period" (Figure 1). While poverty still remains higher in the 0 North than in the country as a (Uganda- NORTH - NORTH - NORTH - whole, urban poverty has fallen Total) total urban rural significantly in the region, confirming positive impressions 10. Azarn et al. Economic Conseuences of the Transition from Civil War to Peace World Bank Policy Research Department, December 1994 p.12 11. Data taken from Simon Appleton 'Change of Poverty and Inequality in Uganda' in An African Phoenix: Assessing the Uganda Recovery (forthcoming in 2000) 6 gleaned from audit mission visits to the three towns. Two factors contributed to the different performance: at least half of NURP-I's investments were directed at urban areas2; and war has had a bigger impact upon rural areas than the towns. Thus, NURP-I probably helped reduce urban poverty through allowing trade to resume in some urban areas, while evidence of project benefits in rural areas is less convincing. Security and Defense Expenditures 3.8 Contrary to the hopes of the Fig. 2 Uganda Military Expenditures designers of NURP-I, lack of security remains a severe obstacle to 180 45 development in the North and 1! AS 140- Ugandan defense expenditures 1 35 climbed throughout the years of project implementation (Figure 2)." 100 -2.5_Q_ How NCRP-I would bring peace and " 0t 1 60 defense savings in a region of . 1.5 ongoing conflict was not explained g in detail in the project's appraisal 0 0 documentation. The cycle of war and0 I 0 docmenatin. he ycl ofwarand1991 1992 1993 1994 1995 1998 1997 1998 violence in the North has not been milions of 1998 dollars (left axis) - 4" as % of GDP (right axis) broken to this day. NURP-I itself had Information from the Stockholm International Peace Research Instittte (SIPAI), htW.1/WWW.siori.s8. its own tragic victims; workers on a Bank-financed water borehole project slain at the hands of guerillas in Kitgum. district in 1996. In considering Bank support for developmeat operations in such circumstances, Bank staff need to reflect upon the moral responsibilities involved if workers lives are lost when governments implement projects that the Bank finances. A careful up-front assessment of the risks to lives of stakeholders should be part of project preparation in conflict prone areas, and a decision not to proceed would be an appropriate response if those risks are considered to be high. 3.9 From a macrosecurity perspective, common sense would suggest that the introduction of equipment, infrastructure and people, into a theater of ongoing conflict would likely increase insecurity and defense expenditure, rather than reduce them as NURP-I designers hoped." New assets such as these increase the number of targets" and require a greater security effort to protect 12. The audit estimates that 50% of project investments were urban, through the share of benefits accruing to urban areas from each component, graded as follows: 100% urban - Telecom and Urban; 50% urban - Community Action Program, Highways, Water Supply, Sub-Projects, Project Administration and PPF; 33% urban - Education and Training; 0% urban - Agriculture and Feeder Roads. 13. Data from the Stockholm International Peace Research Institute (SIPRI). See: http://www.sipri.se:8020/ 14. There is a striking parallel here with the experience of Tamil Tiger guerilla attacks on infrastructure in Sri Lanka, following the disavowal of the 1987 Indo-Sri Lanka Peace Accord. OED's recommendation from its own case study of this is that: "When major infrastructure is being attacked, but the overall security situation has not deteriorated to the point that no assistance is feasible, assistance could be provided directly to households for survival and rehabilitation requirements, and for small-scale infrastructure, possibly of a temporary design so as to minimize the cost of damage from continued or renewed conflict." (See: "The World Bank's Experience with Post Conflict Reconstruction: Vol. 5 Desk Reviews of Eritrea, Haiti, Lebanon, Rwanda and Sri Lanka7 Operations Evaluation Department, May 4, 1998 Report Number 17769 - Sri Lanka case written by John Eriksson and Margaret Arnold) 15. In Northern Uganda's war, schools and their pupils have been considered prime targets by the LRA. 7 them. Ongoing conflict to this day and increased military expenditures point to a different outcome." Gender Aspects Box 2: NURP-1 CAP gives Betty "Best Girl" 3.10 Improvements to the North are Award likely to benefit women in this devastated region. Some 25% of adult women in the Under the project's Community Action Program reio. om 2%fduwlen n futhe (CAP) component a Dutch NGO, SNY, helped build North are war widows, whileself-esteem among women and girls. SNV had 50% are divorced after returning to their expatriate and Uganda staffers living in the field, who parents' home through wartime separation worked directly with beneficiaries. In a well-known from their husbands.' Such tragic social case in the Nebbi district, the CAP awarded a "best- statistics point to single parent households girl" prize to a student named Betty. With this as the norm. In such circumstances, support, Betty was able to proceed to secondary NURP-I benefited women-both as school, overcoming the resistance to her continuing mothers of pupils and as teachers- schooling, given that her family had accepted a cow as through increased primary education. Girl a bride-price, knowing that they would not be able to enrolment-a traditionally low priority for afford the secondary school fees. Education benefited many families in Uganda-could increase. more generally from CAP. In two NURP-I primary Byschools, community ownership and involvement led By poviingagriultraltool an seds, teams of parents and teachers to provide their own NURP-I helped some-how many we labor and work directly on building additional don't know-of the predominantly female classrooms. farmers of the region. In Soroti and Lira, the audit mission saw that women are the principal traders in the local markets built under NURP-I's urban development component. 3.11 The project included a social fund in the form of a Community Action Program (CAP) co-financed by the Netherlands government in an amount of US$6.5 million. Implemented by a Dutch NGO, SNV, in the A,a, Moyo and Nebbi districts of the North, the CAP aimed to build women's and girl's self-esteem through support of local women's organizations, especially those connected with schools. Adult literacy classes and informal programs in family hygiene to helping women gain greater control of their own and children's lives were enacted through a Women's Empowerment Program (WEP), which culminated in a national workshop in Arua in 1998. While good work was done through the CAP (Box 2) the scope and sustainability of its impact is difficult to judge for lack of data on the scale of potential demand throughout the region for this kind of service provided. Municipal Capacity Building 3.12 NLJRP-I's centralized management structure relegated local governments to a minor role, so that capacity building at the district level was limited primarily to managing the operation and maintenance of physical infrastructure provided under the project. From the audit mission's contacts with district administrations in three towns in the North, however, it was clear that the project experience served as an important "wake-up call" to alert local governments about the 16. Despite the growth reported, Uganda's defense expenditures remain among the lowest in a politically volatile region, according to figures of annual expenditures for 1998/99 published by the CIA: Thus Uganda total US$95 million (1.9% of GDP - Note: compared with 2.4% in 98 according to SIPRI); Kenya total US$197 million (1 .9% of GDP); Rwanda total US$95 million (3.4% of GDP); Sudan total US$550 million (GDP unknown); Democratic Congo total US$10 million (3.8% of GDP); and Ethiopia total US$138 million (2.5% of GDP). See: http://www.odci.gov/ciapublications/factbookilug.html#gov. 17. Nat Coletta et al. op. cit. p.14 8 need for their contributions to the development of their own areas in the future. Local officials informed the mission of their desire to play an increasing role. On the other hand, district governments lacked the resources to fulfill broader responsibilities adequately. Decentralization since the 1998 Local Government Act now casts the districts in a much more important role for the follow-on NURP-[I, once peace is established. Fostering Other Donors in the North 3.13 NURP-I alone was unable to overcome many partners' reluctance to becoming involved in the North. NURP-I's own co-financing commitments involving US$6.4 million from the Netherlands and US$10.5 million from DANIDA, were only half used by completion. Lack of security led DANIDA to largely withdraw, although the partnership with the Netherlands continued in full until project completion through the CAP component already mentioned. NURP-I alone cannot be faulted for the mixed performance of donor partnerships in the North. Donor trust and confidence of operating there has to be restored. For that, peace has to be brought to the area. GOU itself is obviously the key player as it addresses the causes and consequences of the war. More important still is how GOU engages efforts by all parties to establish peace within the broad context of relations with the government of Sudan and the international community. In its peacemaking efforts, GOU should avail itself of the best conflict resolution expertise available, knowing that it is likely to be found outside the Bank (see: para. 4.2 below). 4. Unexpected Results Partial Treatment of Ongoing Conflict 4.1 Why was the NURP-I project designed as if the war had ended? Part of the answer may lie in the wishful thinking of some Ugandan authorities and military at the time of appraisal, who had just completed their anti-rebel 'Operation North'. Another part of the explanation may come from the Bank's own lack of expertise in appraising security issues in theaters of military conflict. 4.2 Rather than recruiting expertise that was abundant outside the Bank," the project team hoped that implementing NURP-I, by itself, would help peace break out. Both theory and practical experience elsewhere would suggest otherwise. Introducing new assets into a war zone through a project like this is likely to exacerbate conflict, but this was not an issue addressed by NURP-I's disparate design team. Each sector expert and team-member competently prepared his or her part of what was a very ordinary development project without taking a close look at the quite extraordinary conflict situation in which all the components would have to be implemented. 4.3 Thus, key conflict-related issues were not addressed by the project. Demobilization of ex- combatants, for instance-now widely accepted as a central focus of conflict management-was specifically excluded from NURP-I. Bank management considered it too "political" and at risk from abuse, even though the Bank's own Africa Region had set up a working group on demobilization in 1992. This group found the Uganda Veterans Assistance Program (UVAP) to be a useful instrument-especially when land would available to settle ex-combatants as it was in 18. Especially given the limited professional experience of conflict within the Bank itself, the NURP-I appraisal could have tapped the considerable knowledge and understanding of conflict and conflict-related issues that existed in agencies such as the International Committee of the Red Cross, International Red Cross and Red Crescent Movement, International Alert, Human Rights Watch, M6dicins sans Fronti6res, Africa Fund, and the Carter Center, among others. 9 Northern Uganda-for jump-starting a war-stricken economy, albeit only in a truly post-conflict situation." 4.4 Another post-conflict issue ignored by NURP-I, and that will have to be addressed in due course when peace is restored, is the security conditions required for the private sector to resume investment and business in the area. The NURP-I experience sheds no light on this important aspect. 4.5 On the other hand, NURP-I did address two other important conflict-related issues, albeit perhaps without full cognizance of their implications for war itself. The first was poverty reduction, which can raise the opportunity cost for local people in the North to participate in a rebellion, and thereby reduce the risk of continuing conflict. 2 NURP-I did not however, adopt a poverty reduction strategy specifically related to the ongoing conflict, which might have targeted upon those most at risk from involvement in the violence either as perpetrators or victims. Second, NURP-I addressed the needs of women in the region, but did not recognize their peculiar suffering from the war, particularly through high rates of widowhood, divorce, single-parent families and disruption of family and community structures. NURP-I incorrectly treated women's development in the North in the same way an ordinary project would treat women's development issues in a peaceful and stable society. 4.6 OED's review of the Bank's experience with post-conflict reconstruction studied the case of Uganda, cautioning against Bank involvement in ongoing conflict, while security is uncertain: "Clearly the World Bank is not well-positioned to make the political and military calculations which underlay decisions about macrosecurity. Programs that may have an impact on macrosecurity are not among the Bank's comparative advantages." 4.7 This audit endorses this interpretation by the OED study. The Bank cannot do and know everything related to development. It could well leave to security experts the assessment of these and other issues. The NURP-I experience shows that, by itself, a development project-a core specialty of the Bank-is unlikely to capture the peace.' 19. This was in spite of the successful demobilization of 36,358 Ugandan soldiers through the Ugandan Veterans' Assistance Program (UVAP). See: Nat Colletta, Markus Kostner and Ingo Wiederhofer Case Studies in War-to-Peace Transition: The Demobilization and Reintegration of Ex-Combatants in Ethiopia, Namibia, and Uganda World Bank Discussion Paper No. 331 Africa Technical Department Series, June 1996. OED's own study of post-conflict reconstruction found the Ugandan demobilization experience to be a model for other post-conflict societies. See: Alcira Kreimer, John Eriksson, Robert Muscat, Margaret Arnold, Colin Scott, The World Bank's Experience with Post- conflict Reconstruction, World Bank, 1998. 20. See: Paul Collier The Challenge of Ugandan Reconstruction 1986-1998, World Bank (November 2, 1999 draft). 21. Alcira Kreimer, John Eriksson, Robert Muscat, Margaret Arnold and Colin Scott The World Bank's Experience with Post-Conflict Reconstruction - Volume IV: Uganda Case Study Operations Evaluation Department, The World Bank 1998 p. 34. 22. Operational staff involved in project preparation confirmed that conflict mitigation through infrastructure improvements was very much part of the project culture at that time, when the Bank side itself did not know any better. 10 A Project Design to Induce Corruption 4.8 NURP-I's design lacked effective control systems, something that undermined accountability, thereby providing opportunities for waste23 and corruption, some of which were seized." In other words, NURP-I management in Kampala knew how the money was being allocated, but not what was being done on the ground. By contrast in the region itself, those involved in the project knew what was being built, but were unaware how much it cost or what the programming schedule looked like. In this way, NURP-I became fair game for corrupt officials. Somewhat half-heartedly, the Bank attempted to correct this anomaly by proposing to shift the NURP-I project coordination from Kampala to the North, a move strongly opposed by GOU officials in both Kampala and in the North." 4.9 This NURP-I Kampala/Northern Region 'disconnect' provided opportunities and unsolicited temptations for graft, but which were nobly resisted by many honest officials in Kampala and the North. The audit mission was able to verify on the ground the completion of every single sub-project among the 25 that the mission itself chose to inspect during field visits. The very existence of these physical artifacts bears testimony to the probity of the large numbers of people involved in their planning, financing, implementation and operation. Transgression by others unfortunately can cast a pall of suspicion over the entire endeavor and undermines trust among stakeholders that is a necessary condition for a successful project outcome. On the other hand, significant GOU efforts to rein in miscreants (details paras. 4.17-4.21) can help rebuild that trust by demonstrating publicly that controls are taken seriously 4.10 Among other things, such controls should avoid cost overruns without signs of correlated increases in physical quantities or improvement in the quality of outputs. NURP-I components that suffered the largest overruns were primary schools, urban streets and drains and the Kafu- Karuma main road. The audit paid most attention to primary schools because of the relative ease of locating them and estimating their costs. The approach adopted here, would in principle, be applicable to any other NURP-I component. 23. Wastage of economic resources through unproductive rent-seeking activities has also been classified as corruption. See: Jacqueline Coolidge and Susan Rose-Ackerman High-Level Rent Seeking and Corruption in African Regimes World Bank Policy Research Paper, June 1997. 24. Breakdown of accountability was the main theme of a review of corruption in the education and health sectors in rural Uganda. See; J.C. Munene "Organizational pathology and accountability in health and education in rural Uganda" in: P. Langseth, J. Katorobo, E, Bret and J. Munene Uganda: Landmarks in Rebuilding a Nation Fountain Publishers, Kampala, Uganda 1995 pp. 133-154. The study nevertheless found that public confidence in education high, citing the 1998 National Integrity Survey which found that only 4% of households rated education as the most corrupt service (compared with 27% for health and 60% for the police). 25. The Region argues that Bank efforts to change these arrangements were insistent over the course of 18 months, but which could not overcome GOU rejection of decentralization. The PAR agrees with the lessons that the Region draws from this, namely that: "Control mechanisms and the right set-up should exist from the outset otherwise the mid-course corrections face resistance from borrowers' management and staff', and that "the Bank should be cautious with its initial institutional approach", lest an undesirable arrangement becomes solidified. The Region further confirms that the Bank preparation team got locked into an implementation arrangement with the PMO that the team "did not like". Working initially with the PMO was intended to ease access to the area, not to make the PMO itself into the executing agency. 11 4.11 Actual NURP-I primary Fig. 3 Unit construction costs of primary school school classroom costs- classrooms measured by unit costs per 5.0 classroom-overran significantly 4.5 and were considerably higher 4.0 than those of a Bank-financed 35 project supporting the construction of similar schools in - 2.5 neighboring Rwanda26 (Figure 3). 2.0 In Uganda, the Bank financed 8 15 100% of foreign and 50% of 1.0 local costs. By completion, 0.5 NURP-I classrooms cost, on 0.0 - average, 57% more than planned. NURP-1 NURP-1 Rwanda Rwanda NURP-I's planned unit costs (planned) (actual) (planned) (actual) were already 50% higher than those of the similar Rwanda operation, where final costs were slightly below those planned.17 The NUJRP-1 overruns and relatively high costs are clear. What is not clear is the cause of those overruns. Fromn the simple structures visited by the audit mission, high standards or specification or expensive materials were obviously not responsible. For the successful implementation of NURP-Hl it is important for GOU and the Bank to satisfy themselves that cost overruns were either justified for technical reasons, or that appropriate remedies, including the reimbursement of unjustified payments are sought where technical justifications are absent. The arithmetic of inefficiency and corruption is simple. Incurring twice the necessary cost means reaching only half the population who could have benefited. 4.12 During its detailed field review in Arua, the audit mission was pleased to be able to glean evidence of the existence of all the primary schools reported as completed by the NURP-1 project, and to visit and directly observe 8 of them (Box 3). Before, during and immediately after the field visits, it was still not clear, however, how much each of these schools cost. The Arua District Education Office, which was not responsible for construction under NUJRP-I' s centralized proc- urement arrangements, was not familiar with the cost data. School officials, teachers and parents were even more in the dark. Those who spoke to the audit mission had no idea about the intended specifications of the school-size, materials, length of construction and, of course, cost-to this day. Lack of transparency like this is an incentive for overspending and misspending. Simple billboards at the construction site posting basic sub-project data-cost, design drawing, timing, contractor's name and where to lodge suggestions/complaints-would help. 4.13 During supervision, NURP-I project audits highlighted a number of problems with project accounts. Auditors' qualifications on project accounts from 1993 onwards were consistent with instances of wasteful appropriation and use of project resources. Successive auditors' reports revealed that many things were not quite right. Staff from the Region report that each supervision 26. Rwanda: Education Sector Credit (Cr. 2227) approved shortly before and executed over a similar period to NURP-1 27. Comparisons of planned/actual and country unit costs of primary school classrooms were based upon the following assumptions: (i) accurate reporting of appraisal and completion costs in both cases; (ii) no exchange rate distortions between the two countries, nor over time within countries when estimating US dollar equivalents; (iii) construction costs account for 69% of the total (remalnder consists of furniture, technical assistance, books and taxes); and (iv) in Rwanda case unit costs of classroom rehabilitation-which includes painting, furniture repair and replacement of broken windows and doors-amounts to 15% of unit cost of construction of a new classroom (task team estimate). 12 mission would carry out ex-post Statement of Expenditure (SOE) Box 3: Evaluating Primary Schools in Arua reviews, based upon audit During its visit to the district of Arua, the audit mission spent findings. The Bank's most of the first day with the District Education Officer and his disbursement officer-from the team, who kindly responded to all the detailed inquiries made resident mission in Kampala- about primary schools financed by NURP-1. Arua district was participated regularly in these cited in allegations being tried in Kampala about payment for reviews. From 1995 onwards school furniture that was not delivered. Other allegations made project files include Bank requests were that schools paid for did not exist and those that did were fornot completed, although reported as such. In the midst of these dubious expenditures o cl r allegations, the audit mission arrived with project dubius ependture andfordocumentation that detailed the number of schools, their names explanations of corrective action and specifications. As a first step at a meeting in the District by local NURP-I management, Education Office, the local education team was able to identify even though it was not evident the exact location, one-by-one, of each school as its name was when reimbursements of mis- called. From their quick identification and agreement on procured amounts were requested locations, it was clear that all schools existed and that local of the borrower. Examples of officials were quite familiar with all 24 reported built by auditors' comments included: NURP-1. Then, without any interference by the local team or officials from Kampala on the field trip, the mission selected 0 Prjectaccuntswer delyed six schools to visit across the district of Arua. Two more were * Project accounts added to these at the request of the District Education Officer. and not prepared inThus, the mission immediately made an unimpeded and accordance with Bank unannounced visit to eight primary schools financed by standards (FY93). NURP-1. There was very little time to war the schools of the * 9% of project expenditures impending visit. Contrary to many allegations, all the schools (US$40,100) were without visited existed as completed units and were operational. All supporting documentation, so were properly roofed, most used shutters on windows. Building thatmaterials were mostly local. Construction quality varied. The thatreleant OEs ouldnot poorest two schools had ill-fitting doors and inadequate finish be traced to underlying being attributed to one particular contractor. Most schools were accounting records (FY94). fully furnished, although some classrooms in three of them had * Not all expenses-such as no desks at all, the focus of the alleged fraud under trial in US$37,600 for roofing Kampala. Overfall the schools represented for the staff and materials-were supported by pupils a substantial improvement over previous conditions payment vouchers (FY95). where classes had to be held under the shade of mango trees. * NUR.P-I's Agricultural credit fund (final total cost US$ 1.1 million) was not separately audited (FY95) (NURP-I considered such auditing superfluous duplication). D Delayed appointment of auditors of up to six months,e prevented auditors from conducting physical year-end inventory counts as required under Bank regulations (several FYs). Inaccurate allocation of expenditures to expense categories. * Quality and timeliness of accountabilities was inadequate, incorrectly allowing new advances for districts' components before earlier advances had been fully accounted for (FY96). 4.14 While these audits were not identifying hajor elements of fraud through their random surveys, they were indicating that some things were out of line. NteRP-I management response in OPM in these early years appears to have led to some improvements. The FY97 audit had no major qualifications on the project accounts. Nevertheless, in their letter to NURP-I management in Kampala, the auditors reminded the team, "while preparing the final accounts, a full and 28. Regional staff report that the Bank insisted on immediate and unannounced audits whenever a district officer had not submitted accounts. One such audit resulted in the dismissal of a district officer in Gulu. 13 accurate trail should be prepared to link up the figures and the final accounts tot the underlying records for accounting purposes." Audits for FY98 and FY99 together with spot checks on the ground by NURP-I monitors pointed to lack of clarity of spending on and missing items of school furniture, an observation picked up by the Bank's task manager, who requested a specific investigation into the situation, which led to the case currently at trial in Kampala.29 4.15 As well as providing opportunities for corrupt practices to go undetected, neglect of control can undermine stakeholder confidence and beneficiary trust, without which any operation is doomed to fail. Through tightening controls, trust can be shored up, as it urgently needs to be in Uganda. The country's 1998 National Integrity Survey found that 70% of households across all regions believe corruption to be rampant in public service, and 57% think the problem is getting worse. The audit mission itself heard many unprompted and unsolicited allegations of corruption from many different sources. Altogether, and leaving aside double counting, the amounts allegedly embezzled from NURP-I would have exceeded the total funding of the project, according to these sources. 4.16 From the way in which NURP-I has responded thus far to problems arising from its own lack of control, a fight against corruption should begin within a project itself. Thus, NURP-II's design needs to have tight ex ante and ex post controls built into it. Once improved security conditions permit such an operation-according to informed expert opinion-NURP-II's implementation needs to be decentralized into parts of the North where financial and implementation management can be closely observed from both above and below. Implementution should be transparent enough for local communities and beneficiaries to monitor progress according to plans and targets laid out very publicly. There would need to be some form of hot line or complaint desk in the North, as well as in Kampala, where local beneficiaries can fearlessly and anonymously report suspected abuse in project implementation. "Flying squads" of inspectors from outside, who can make surprise unannounced checks can provide effective control as they were able to do to some extent under NURP-I. Cleaning up project acts cannot rely solely on national programs, such as Uganda's National Integrity Action Plan. Analysts have criticized the latter's "shopping-list" of proposed actions-awareness seminars, a "living wage" for civil servants, special courts for corruption cases, strengthening existing agencies including Auditor General and Inspector General.' Whether the criticism is well placed or not, the NURP-I experience makes clear that the clean-up has to start 'at-home' within a project itself. That places a heavy responsibility on the shoulders of designers and managers of individual projects for preventing and monitoring corrupt practices within their domain. A Case Study of Fighting Corruption 4.17 To date, the most striking NURP-I anti-corruption effort has been an investigation into over-payment31 for school furniture, leading to the filing of charges against the former NURP-I project manager (and permanent secretary to the prime minister), accountant and business 29. See para. 9 of Aide M6moire of completion mission dated Sept. 28, 1998. 30. Rachel Flanery and David Watt 'The state of corruption: a case study of Uganda' in: Third World Quarterly Vol. 20 No. 3 1999 p. 534. In a case of NURP-l's alleged corruption, the chief defendant was former Permanent Secretary (PS) of OPM. As part of Uganda's anti-corruption measures, PS salaries had been increased 42,464% over the 1993-97 period to a level of US$1,550 per month. 31. According to anti-corruption experts, this is one of the most common modalities of fraud. See: Augustine Ruzindana 'Combating Corruption in Uganda' in: P. Langseth, J. Katorobo, E, Bret and J. Munene Uganda: Landmarks in Rebuilding a Nation Fountain Publishers, Kampala, Uganda 1995 pp. 191-208 14 associate. This involved a total of US$233,000 paid between October 1997 and January 1998 for furniture that was not fully delivered. The case first came to court in Kampala in April 1999, when the suspects, who had been remanded in custody, were released on bail pending the trial. At the time of writing, Ugandan government prosecutors are appealing the judge's decision in April 2000 to throw out the charges for lack of evidence. The case continues. 4.18 Checks available to the NURP-I project itself were responsible for bringing this case to light. During a surprise field inspection to Arua inspired by doubts raised by project audits, NURP-I monitors found that the furniture paid for had not been delivered. They reported back to NURP-I management in Kampala, who did not act upon this tip-off, precisely because management itself was allegedly involved in the matter. Meantime, a Bank supervision mission independently uncovered more evidence of this case through an audit of SOEs that traced and reconciled them back to progress reports issued by the project monitoring unit. Upon learning of this, the Bank's task manager requested, through the Aide Memoire of the completion mission, a full investigation by the Ugandan authorities. The ensuing investigation was conducted through the regular channels of the office of the government Inspector General (IG) which brought charges now being tried through the Ugandan justice system. While the outcome of this case is still unknown at the time of writing, Ugandan authorities are to be applauded for their determination thus far in bringing to justice a case that involved a high-ranking official of the regime. 4.19 Two important lessons from this case are relevant to future operations. First, the NURP-I experience shows that, when properly applied with the active collaboration of the Bank and GOU project staff, normal project procedures-particularly financial audits, routine monitoring and spot checks-can identify and inhibit corrupt practices. Second, control mechanisms that help fight corruption need to be built into each lending operation and be given active attention by project staff throughout implementation. In order to help reap full benefit from lessons such as these, and also to help the Bank fulfil its fiduciary responsibilities to fight corruption, it is important that all cases such as this one are reported in future to the Bank's central Anti- Corruption and Fraud Secretariat. 4.20 After the furniture case came to light, further fraud allegations by a NURP-I service provider were made to the Bank's Anti-Corruption and Fraud Secretariat, who decided to investigate further. This investigation-which counted on the full collaboration of this audit-is currently underway. At the time of writing, it's findings, which can lead to recommendations for criminal prosecution of wrongdoers, are pending. 4.21 From the anti-corruption results achieved thus far, success in pinpointing problems has come from the rigorous application of checks and control instruments available to the NURP-I project itself. Nation-wide anti-corruption drives, such as the one in Uganda currently supported by the Bank, 32 may help broaden general awareness of the issue, but may not provide the instruments needed on the front line in the battle against corruption. 32. Report ofa World Bank Mission to Support the Program ofthe Government of Uganda to Improved Economic Governance and Combat Corruption 1998 15 5. Conclusions and Lessons Ratings 5.1 Despite the extraordinary difficulties it faced, the NURP-I project did achieve some important outcomes, but efficiency was undermined by higher project expenditures than expected. Some NURP-I objectives were relevant to situation in the North, and some were partially achieved, notably restoring economic productivity, reducing poverty and addressing the needs of women, but overspending-through waste or corruption-undermined the efficiency of these achievements. For these reasons, the audit rates project outcome as marginally satisfactory, differing from the ICR/ES's (Evaluation Summary) satisfactory outcome rating. The field audit, gave insight into lower levels of efficiency in achieving project objectives than were apparent from the ICR and, hence, reported by the ES. Observation by the audit mission in the Northern Region itself, nevertheless, confirmed important development gains on the ground, in the form of useful infrastructure, albeit exploited to different degrees. On the other hand, the audit agrees with the ICR/ES ratings of uncertain sustainability and modest institutional development impact. Bank performance is rated here as unsatisfactory, since the Bank brought an inadequate understanding of and inappropriate responses to the ongoing conflict in the North, project design had poor control and accountability, and hands-off supervision during the first half of implementation helped leave accountability in abeyance. There were, however, satisfactory Bank efforts to begin to recover control under a new task manager in the closing months of the project-that led to a successful anti-corruption investigation and prosecution by the Uganda authorities. This worthwhile push can too late, however, and the audit does not agree with the ICR/ES rating of satisfactory for Bank performance overall. The audit nevertheless agrees with the ICR/ES rating of satisfactory for borrower performance. Although the borrower had similar weaknesses as the Bank at the outset and shared much of the responsibility for the poor project design, inadequate assessments of the security situation, and made a less than satisfactory contribution to implementation management through CMEU/OPM, GOU's highly satisfactory efforts thus far in fighting corruption at a high a level, deserve recognition in a satisfactory rating overall. Bank's Comparative Advantage Is Not in Dealing with Ongoing Conflict 5.2 The experience of NURP-I showed that the project never really came to terms with the fact that Northern Uganda was (and remains until the time of writing) a theater for ongoing conflict. With hindsight, it is clear to this audit and now to some Bank staff involved in NURP-I, that the Bank should not have tried to put together a more or less conventional investment project in a conflict situation. Although it was unsafe to travel to most of the region during appraisal, the Bank-probably through lack of experience-erroneously treated the North as a case of post- conflict reconstruction. From reviewing NURP-I, this audit endorses the findings of OED's own evaluation of Bank experience with post-conflict reconstruction, that it is not the Bank's comparative advantage to analyze and manage security matters in war situations. Issues arising from these are best left for security and military experts to assess. Such specialists can apprise GOU the Bank itself when conditions may be propitious for reconstruction operations to begin after the conflict is settled. The following evidence from NURP-I suggests the need for caution and reticence by the Bank when presented with proposals to intervene in war zones:" 33. Bank operational staff involved with NURP-I mentioned Sierra Leone and Liberia as two countries to which these lessons may apply. 16 * The ongoing conflict in the North itself and the loss of lives of construction workers engaged in implementing project components at the hands of insurgents; * The Bank's inability or unwillingness to design and supervise the project on the ground for much of the time for justified fear of insecurity; * The overriding importance of military issues of counter-insurgence and demobilization that were not addressed by those preparing the NURP-I project; * Failure to see the importance of education as a measure of conflict mitigation that could raise the opportunity costs of students joining the rebel movement; * Lack of clear understanding that bringing new project equipment, infrastructure and people into a zone of on-going conflict introduces new targets for rebels. With more assets to defend, military expenditures rose instead of falling, as the NURP-I project hoped. The Responsibilities of Project Designers and Managers for Anti-Corruption 5.3 Individual operations such as NURP-I should be at the heart of a government's anti- corruption drive. NURP-I's initial centralized design, which separated financial management from physical implementation, undermined control and accountability and served as an inducement to corruption. At the same time, NURP-I's standard project elements of control, notably financial audits and monitoring with active involvement of Bank and borrower project staff, were successful in identifying and investigating some alleged corruption. In this tale of failure and success, the following lessons can be applied to NURP-II, once the war in the North is over: * Project planning, design and implementation should be decentralized not least of all to unify financial management with physical implementation. * As part of decentralization, project design should involve district administrations, local communities and final beneficiaries in project planning and implementation. * Basic information-notably costs---on project components should be shared with local people who should have access to some kind of hot line or help desk where they can denounce abuse of project resources. * Investigations of allegations of abuse should be routine. * The local press and NGOs should be regularly briefed about project progress to encourage their watchdog roles. * Projects should publicly recognize and reward honest and diligent project staff in order to build up trust that has been eroded. With very poor controls, NURP-I had some significant achievements, but the Bank should not have to rely upon people's outstanding altruism to guarantee the diligent implementation of a follow-on project. 17 Some Elements for the Profile of a NURP-II Operation, Once Peace Is Achieved' 5.4 As soon as the conflict is settled, the Bank should be in a good position to help the reconstruction of Northern Uganda. From what has been learned from the completed operation, a follow-on NJRP-II project could address the following issues: * Incorporation of all the anti-corruption lessons above (para. 5.3). * Seek the advice of security and military experts regarding the security risks of the area to determine the minimum satisfactory conditions for proceeding and communicate these conditions to GOU. * Proceed with NJRP-II only when project management and control structures can be safely put in place in the North, in the expert opinion of security and conflict specialists. * Phasing of NURP-II operations to focus on the most secure districts first. * Extensive surveys of the needs and priorities of the current residents. Community participation through CAP social fund has proved effective on a small scale, but it is unlikely to be enough given the scale of the devastation wrought and the large components of infrastructure that need to be introduced. * Directly addressing key conflict-related issues such as demobilization, rebuilding family and social lives (especially for women) and inducements to restarting private businesses especially through repatriation and restoration of assets. 34. In comments on this PAR, staff from the Bank's Post Conflict Unit confirmed that the design of a follow-on NURP-II "will/should be a very different animal in both process and composition" from NURP-I.  19 Bibliography Appleton, Simon 2000 "Change of Poverty and Inequality in Uganda" in: African Phoenix: Assessing the Uganda Recovery (forthcoming). Azam, et. al. 1994 Economic Consequences of the Transition from Civil War to Peace Washington D.C. World Bank Policy Research Department. Colletta, Nat J., Markus Kostner and Ingo Wiederhofer 1996 - June Case Studies in War-to- Peace Transition: The Demobilization and Reintegration of Ex-Combatants in Ethiopia, Namibia and Uganda Washington D.C. World Bank Discussion Paper No. 331 (Africa Technical Series). Collier, Paul 1999 - November How to Reduce Corruption Washington D.C. Paper prepared for the Conference on Good Governance and Sustainable Development, held in Abidjan November 22-24, 1999. World Bank. Collier, Paul 1999 - November The Challenge of Ugandan Reconstruction 1986-1998 (draft paper) Washington D.C., World Bank. Coolidge, Jacqueline and Susan Rose-Ackerman 1997 - June High-Level Rent-Seeking and Corruption in African Regimes: Theories and Cases Washington D.C. World Bank Policy Research Paper 1780. Doom, Ruddy and Koen Vlassenroot 1999 - January "Kony's message: a new koine? The Lord's Resistance Army in Northern Uganda" in African Affairs Volume 98 No. 390. Flanary, Rachel and David Watt 1999 "The state of corruption: a case study of Uganda" in Third World Quarterly Vol. 20 No. 3. Human Rights Watch 1998 Crises in Sudan and Northern Uganda Washington D.C. Testimony of Jemera rone, Human Rights Watch before the House Subcommittee on International Operations and Human Rights and the Subcommittee on Africa. www.igc.apc.org/hrw Kreimer, Alcira 1998 The World Bank's Experience with Post-Conflict Reconstruction Washington D.C. Operations Evaluation Department, World Bank. Munene, J.C. 1995 "Organizational pathology and accountability in health and education in rural Uganda" in Uganda: Landmarks in Rebuilding a Nation edited by: Langseth, P., J. Katorobo, E. Brett and J. Munene: Kampala, Fountain Publishers. Mutiso, Roberta and Deborah Kasente 1997 - September Improving the Performance of the Country Programme in Reaching the Poor Kampala, Government of Uganda/UNICEF Country Programme 1995-2000 Pouliquen, Louis Rural Primary School Construction Costs in Africa (undated draft report for the World Bank). Ruzindana, Augustine 1995 "Combating corruption in Uganda" in: Uganda: Landmarks in Rebuilding a Nation edited by: Langseth, P., J. Katorobo, E. Brett and J. Munene: Kampala, Fountain Publishers. Shah, Anwar Anti-Corruption Policies and Programs: A Framework for Evaluation (undated draft) Washington D.C. Operations Evaluation Department. World Bank 1998 The World Bank's Experience with Post-conflict Reconstruction: Volume IV Uganda Case Study. Washington D.C. Operations Evaluation Department of the World Bank. World Bank 1998 Post-Conflcit Reconstruction: Uganda Case Study Summary Washington D.C., OED Pr6cis. 20 Annex A Basic Data Sheet NORTHERN UGANDA RECONSTRUCTION PROJECT (LOAN 2362-UG) Key Project Data (amounts in US$ million) Appraisal Actual or current estimate estimate Total project costs 98.18 98.91 Loan amount 71.2 71.2 Cofinancing 16.96 8.61 Cancellation - . Date physical components completed 09/30/98 09/30/98 Economic rate of return 24% 12% Cumulative Estimated and Actual Disbursements FY92/93 FY93/94 FY94/95 FY95/96 FY96/97 FY19798 FY/9899 Appraisal estimate (US$ thousands) 2.3 9.1 18.2 39.4 57.5 69.6 71.2 Actual (US$ thousands) 3.9 6.6 21.9 39.1 59.03 72.1 72.9 Date of final disbursement 09/30/98 Project Dates Steps in project cycle Date planned Date actual Ildentification (IEPS) 12/06/90 12/06/90 Preparation 05/15/91 05/15/91 Preappraisal 06/10/91 06/10/91 Appraisal 09/16/91 09/16/91 Negotiations 03/16/92 03/16/92 Board Presentation 05/05/92 05/05/92 Signing 06/12/92 06/12/92 Effectiveness 12/09/92 12/09/92 Project Completion 09/30/97 09/30/97 Loan Closing 09/30/98 09/30/98 Staff Inputs (staff weeks) Stage of project cycle Staff week actual Amount US$('000) Through pre-appraisal 71.3 138.1 Appraisal-effectiveness 45.3 89.2 Supervision 238.8 478.4 Completion 3.3 7.3 TOTAL 358.7 713.0 Annex A 21 Mission Data Stage of project No. of Staff Specialized staff Implem. Develop. Types of problems cycle Month/year persons days in skills represented status Object field Through Appraisal Appraisal through Board approval Supervision 1 Aug.92 7 FA, HE, IS, ME, SE, S S UP Supervision 2 Apr.93 7 OA, EE, HE, UP, Ed S S Supervision 3 Jan. 94 1 0 U U Supervision 4 Oct. 94 5 Ec, OA, 0 PS U U Supervision 5 June 95 2 0 (2) S S Supervision 6 Nov. 95 3 Ec, 0, S S Supervision 7 Mar. 96 4 Ec, HE (2), 0 S S Supervision 8 Oct. 96 4 Ec, FA, HE (2) S S Supervision 9 Mar.97 5 13 HE(2), Ec, IT S HS Supervision 10 Oct. 97 4 5 Ec, HE(2), 0 S HS Supervision 11 Apr. 98 3 10 EE, FA, 0 S HS Completion Sept. 98 5 5 Ec, FA, 0, HE (2) E = Environmental Specialist Ed = Education Specialist IS = Institutional Specialist Ec = Economist HE = Highway Engineer IT = Information Technology EE = Environmental Engineer FA = Financial Specialist ME = Municipal Engineer PS = Procurement Specialist 0 = Operation Officer/ OA = Operation Analyst Proc.Specialist SE = Sanitation Engineer TS = Training Specialist UP = Urban Planner 22 Annex B Attachment: Northern Uganda Reconstruction Project (Credit 2632-UG) Final Project Cost by District (in millions of US dollars) Total - Total - final planned Apac Arua Gulu Kitgum Kum[ Lira Mayo Nebbi Pallisa Soroti (CR) (SAR) A. Agriculture 0.61 0.08 1.80 0.76 0.17 0.45 0.08 - 0.36 0.68: 5.0 5.1 i. Rural Credit 0.23 - 0.18 0.29 - 0.25 - - 0.15 1.1 1.4 ii.Agriculture inputs 0.02 0.08 0.09 0.12 0.02 0.08 0.08 - 0.01 0.12 0.6 1.8 iii. Equipment& Vehicles - - 0.31 - - - - - - 0.3 0.7 iv. Training - - - 0.05 - 0.02 - - - 0.1 0.1 v. Administration 0.17 - 0.03 0.13 0.07 0.02 - - 0.10 0.21 0.7 1.0 vi.Extension,Seed 0.19 - 0.03 0.17 0.08 0.08 - - 0.11 0.34 1.0 - vii.Civil Work/Gulu DFI - - 1.16 - - - - 1.2 - B. Community Action P. na na na na na na na na na na 6.5 6.3 c. Education & Training 1.97 1.64 2.33 2.34 1.80 2.12 1.20 1.60 1.09 3.01 191 10.2 i. NITEP 0.62 0.73 0.85 0.66 0.71 0.63 0.54 0.49 0.52 1.20 7.0 2.8 li Technical Training 0.10 0.10 0.10 0.10 0.10 0.10 0.10 - 0.10 0.10 0.9 1.0 lii.Primary Schools 1.25 0.81 1.38 1.58 1.00 1.40 0.57 1.11 0.47 1.70 11.3 6.6 D. Feeder Roads 0.15 - -- 0.15 0.19 - 0.19 28 8.7 i. Maintenance Equipment 0.12 - - - 0.12 0.12 - - - 0.12 0.5 na ii.Repair of Equipment 0.03 - - - 0.03 0.07 - - - 0.07 0.2 na iii. Cap building (DANIDA) - - - 2.1 na E. Highways - - 10.27 2.14 0.98 9.75 1.86 - - 7.07 32 1 25.6 i.Atiak-Moyo Road - - 0.73 - - - 1.71 - - 2.5 3.6 ii.Gulu-Achotibur Road - - 1.32 1.98 - - - - - 3.3 3.3 iii.Soroti-Lira Road - - - - - 2.92 - - - 2.92 5.8 6.1 iv.Kumi-Serere-Soroti - - - - 0.90 - - - - 3.59 4.5 3.4 v.Kafu-Karuma Road - - 7.54 - - 6.17 - - - - 13.7 8.1 vi.(a)Const.Supervision - - 0.51 - - 0.41 - - - 1.0 1.2 vi.Const.Supervion Contr. - - 0.17 0.16 0.08 0.24 0.14 - - 0.56 1.4 - F.iTelecom na na na na na na na na na na 0.3 11.8 G. Urban Development 1.79 - 3.16 1.54 1.62 2.27 - - 1.30 1.79 13.5 6.9 i.Rehab. Of streets 0.79 - 2.04 0.98 0.87 1.22 - - 0.77 1.03 7.7 4.4 ii.Reb. Of markets 0.22 - 0.22 - 0.10 0.22 - - - 0.10 0.9 - iii.Upgrading of Town Mkt. 0.22 - 0.22 - 0.10 0.22 - - - 0.10 0.9 1.4 iv.Maint. Of Urban 0.10 - 0.10 0.10 0.10 0.10 - - 0.10 0.10 0.7 0.3 v.Town bundry survey 0.28 - 0.28 0.28 0.28 0.28 - - 0.28 0.28 2.0 0.5 vi.SRPs-Aerial Photo. 0.07 - 0.07 0.07 0.07 0.07 - - 0.07 0.07 0.5 - vii.Training of local coun. - - 0.01 0.01 0.01 0.01 - - 0.01 0.01; 0.1 0.3 viii.Design and supervision 0.09 - 0.20 0.09 0.08 0.13 - - 0.06 0.09: 0.7 - ix.Supervision of urban 0.02 - 0.02 0.02 0.02 0.02 - - 0.02 0.02: 0.1 - H.Water Supply & Sant. 0.68 1.44 0.74 1.85 1.62 1.23 0.24 - 0.19 1.82; 9.8 5.6 i. Urban Water - 0.46 - 0.19 1.04 0.18 0.24 - 0.19 0.99: 3.3 na ii.Rural water 0.23 0.33 0.22 0.54 0.20 0.33 - - - 0.28: 2.1 na iii.Consultancy 0.04 0.05 0.04 0.09 0.03 0.05 - - - 0.04, 0.3 na iv.TechnicalAssistance 0.02 0.02 0.10 0.10 0.01 0.10 - - - 0.02 0.4 na v. Training 0.02 0.02 0.02 0.04 0.01 0.02 - - - 0.02 0.2 na vi.Equip.,vehicles & Admn. 0.38 0.55 0.37 0.89 0.33 0.55 - - - 0.46 3.5 na 1.Sub-Projects 0.14 0.05 - - 0.06 0.08 0.03 0.12 0.14 0.05 0.7 - i.Const. Of schools 0.12 0.04 - - 0.05 0.06 0.02 0.09 0.11 0.04 0.5 - ii.Consultants 0.01 0.01 - - 0.01 0.01 0.02 0.01 - 0.1 - iii.Admn. Costs, 0.02 0.01 - - 0.01 0.01 - 0.01 0.02 0.01 0.1 - J. Project Admin (CMEU) na na na na na na na na na na 8.5 2.3 K. Project Prep. Facility na na na na na na na na na na: 0.9 1.5 NURP-1 TOTAL 5.34 3.21 18.30 8.63 6.40 16.03 3.42 1.72 3.09 14.61; 99.2 (*)84.0 Sources/Notes: Actual cost tables by district prepared specially for this audit by CMEU. Planned costs refer to base costs reported by SAR and ICR, and (*)exclude contingencies. Project appraisal estimates of contingencies were US$ 14.3 million, bringing total planned costs to US$98 million. In this table, totals are given to one decimal place only, resulting in rounding errors in some cases.