Somalia - Mogadishu Second Water Supply Project

Document of The World Bank FOR OFFICIAL USE ONLY Report No. 9727 PROJECT PERFORMANCE AUDIT REPORT SOMALIA MOGADISHU SECOND WATER SUPPLY PROJECT (CREDIT 1236-SO) JUNE 28, 1991 Opfrations Evaluation Department This document has a restricted distribution and may be used by recipients only in the performance of iheir official duties. Its contents may not otherwise be disclosed without World Bank authorization. COUNTRY EXCHANGE RATES Somalian Shillings (SoSh)/US$ Appraisal Estimate (1980) Shl.00 6.295 US$ Actual: Average 1981 6.295 Average 1982 10.75 Average 1983 = 15.79 Average 1984 = 20.02 Average 1983 39.49 Average 1986 72.00 Average 1987 = 105.18 ACRONYMS AND ABBREVIATIONS EDF - Economic Development Fund (EEC) EEC - European Economic Community ENEE - National Power Company Government - Government of the Democratic Republic of Somalia (also GOS) MF - Ministry of Finance MMWR - Ministry of Mines and Water Resources MOA - Magistrate of Accounts MWA - Mogadishu Water Agency EORAD - Norway Development Agency PMPU - Project Management and Planning Unit SoSh - Somali shilling FOR 0MVCUAL USE ONLY THE WORLD BANK Washngton, D.C. 20433 U.S.A. 01ice of oiectot*Ceeai Opwaties avalation June 28, 1991 MEMORANDUM TO THE EXECUTIVE DIRECTORS AND THE PRESIDENT SUBJECT: Project Performance Audit Report on Somalia - Mogadishu Second Water Supply Project (Credit 1236-80) Attached, for information, is a copy of a report entitled "Project Performance Audit Report on Somalia - Mogadishu Second Water Supply Project (Credit 1236-SO)" prepared by the Operations Evaluation Department. Attachment 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. FOR OFFICIAL USE ONLY PROJECT PERFORMANCE AUDIT REPORT DEMOCRATIC REPUBLIC OF SOMALIA MOGADISHU SECOND WATER SUPPLY PROJECT (CREDIT 1236-30) TABLE OF CONTENTS Preface ............................................................... Basic Data Sheet . *........ . . . . .. . . . . 0....... .. . . ..... iii Evaluation Summary .............. . .......... .................. * ....... v I. BACKGROUND ....................................**.*.............. 1 II. THE PROJECT ....... ....................................... 2 A. Project Objectives. Description and Processing .............. 2 B. Project Scope .............. .......... ........ ...... 2 III. PROJECT IMPLEMENTATION ... ....................................... 3 A. Project Costs .......................................... 3 B. Water Audit ................-.............. .**** **** 4 C. Low-cost Sanitation ..... . ..........................* * * 4 D. Training ........... 0 ............. ......... 4 E. Financing Plan ................*................. ..... 000 5 F. Diabursements ...........0 0 .*....... **....... ... ... .. ......... 6 G. Cofinancing .......... ............... .. . .. .. . .. . ..... 6 H. Procurement ................ *****$*********************....... 7 I. Maturity. Grace Period and Exchange Risk .................... 7 J. Financial Covenants .................*......................... * 8 K. Basic Accounting and Auditing ............................... 10 L. Water Tariffs ........ 0 0.... 0....... ..... ................... 10 M. Institutional Development ...............*.*....... ....... 12 N. MWA Financial Situation ......................................* 13 IV. LESSONS LEARNED ................... .. ........................... ...... 14 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.I i PROJECT PERFORMANCE AUDIT REPORT DEMOCRATIC REPUBLIC OF SOMALIA MOGADISHU SECOND WATER SUPPLY PROJECT (CREDIT 1236-SO) PREFACE 1. This is a Project Performance Audit Report (PPAR) on the Second Water Supply Project in Mogadishu, involving a IDA credit in the initial amount of US$15.0 million to the Democratic Republic of Somalia to carry out a second phase of investments to increase water production and distribution in the capital. The credit was approved on April 27. 1982. The credit was entirely used after the closing date was postponed twice by one year. Half of the credit amount was refinanced by Norway in the form of a grant. 2. The PPAR is based on the Project Completion Report (PCR) prepared by the Africa regional department and issued on December 30, 1989 (Report No. 8454),; the President's Report. the loan documents, and on a study of the project files. An OED mission visited Somalia in June 1990 and discussed the effectiveness of the Bank's assistance with the Hogadishu Water Agency (MWA). The kind cooperation and valuable assistance provided by the Ministry of Mines and Water Resources (HHWR) in the preparation of this report is gratefully acknowledged. 3. The PCR provided a comprehensive account of the project experience. The PPAR elaborates on selected aspects of project implementation, in particular the financial and institutional building aspects of the project. 4. Due to the prevailing circumstances in the country and on the advice of the relevant country department no comments were requested. iii PROJECT PERFORMANCE AUDIT REPORT DEMOCRATIC REPUBLIC OF SOMALIA MOGADISHU SECOND WATER SUPPLY PROJECT (CREDIT 1236-S0) BASIC DATA SHEET KEY PROJECT DATA Appraisal Actual Actual as % Item Expectations Estimates of Appr.Est. Total Project Cost (US$ million) 42.3 40.2 -5 Loan Amount (") 15.0 7.5' same2 Date Physical Components Completed 06/85 12/85 Economic Rate of Return (Z) 5.6 4.0 ----------------------------------------------------------------------- CUMULATIVE ESTIMATED AND ACTUAL DISBURSEMENTS (US$ million) FY ended June 30, 1983 1984 1985 1986 1987 1988 (1) Appraisal Estimate 3.30 8.40 13.35 15.00 - - (2) Actual 6.53 9.52 11.69 12.63 13.65 14.40 (3) (2) as % of (1) 198.0 113.0 88.0 84.0 - - Date of Final Disbursement: 6/13/1988 ----------------------------------------------------------------------- PROJECT DATES Original Plan Revised Actual First Mention in Files 5/27/77 Negotiations 3/01/82 Board Approval 4/82 4/27/82 Credit Agreement Date 5/26/82 Effectiveness Date 8/15/82 10/12/82 Closing Date 12/31/85 12/31/87 1/ The PCR is incorrect in mentioning US$15 million since half was refinanced by Norway during disbursements. 2/ There was no change in IDA's gross contribution to the financing plan. iv STAFF INPUT (Staff-Weeks) not available in the PCR MISSION DATA No. of No. of Date of MonthfYear Days Persons Manweeks Report Identification 8/81 5 2 2.0 5/25/77 Prep/appraisal 12/81 6 3 3.6 5/16/78 SupLrvision 1 4/82 2 2 0.8 4/27/82 Supervision II 8/82 2 1 0.4 n.a. Supervision III 8/82 2 1 0.4 9/15/82 Supervision IV 9/82 15 1 3.0 10/01/82 Supervision V 11/82 5 1 1.0 12/22/82 Supervision VI 3/83 7 1 1.4 3/31/83 Supervision VII 7/83 2 2 0.8 7/26/83 Supervision VIII 9/83 3 2 1.2 10/18/83 Supervision IX 6/84 5 2 2.0 8/15/84 Supervision X 8/84 3 2 1.2 n.a. Supervision XI 9/84 6 1 1.2 10/05/84 Supervision XII 10/84 10 1 2.0 11/16/84 Supervision XIII 2/85 6 2 2.4 3/20/85 Supervision XIV 6/85 5 3 3.0 6/25/85 Supervision XV 11/85 7 2 2.8 12/23/85 Totals (averages) 5.4 1.7 29.2 ------------------------------------------------------- OTHER PROJECT DATA Borrower - Democratic Republic of Somalia Executing Agency - Ministry of Mines and Water Resources Fiscal Year of Borrower - January 1 - December 31 Follow-up Project - None V PROJECT PERFORMANCE AUDIT REPORT DEMOCRATIC REPUBLIC OF SOMALIA MOGADISHU SECOND WATER SUPPLY PROJECT (CREDIT 1236-SO) EVALUATION SUMMARY Introduction 1. In an arid country like water supply production capacity to Somalia. water is essential to the meet demand until 1987. The population's existence. It is physical component included power considered the most fundamental of generation facilities for the new basic needs by the Government whose pumping station, a transmission policy is to provide water as a main, extension to the distribution social service at a modest cost. system and equipment associated with Water resources are limited and ill these civil works. To strengthen distributed, and investigations to the water agency (MA). technical develop raw sources was a priority assistance focused on finance and in the late 1970s. This was an accounting was also included (PPAR. objective of the First Water Supply para. 2.02). The project cost was Project (Credit 822-SO) approved by estimated at appraisal at US$42 the Board in June 1978.1 Completed million of which 79% represented the in 1982 the First Project was an foreign exchange component; 18% of "interim measure" with a scope contingencies were also included to limited to catch up with the growing avoid the cost overrun financing demand in the capital. The problem of the First Project. The Mogadishu Second Water Supply project attra;ted co-financiers: the Project was the continuation of the Arab Fund for US$17.5 million, the First Project to develop the Afgoy EEC for US$5.3 million and Norway wellfield to full capacity. which refinanced half of the IDA Pri rity was again given to the credit; the last two sources were in capital because a quarter of the form of grants (PPAR, para. Somalia's population is in Mogadishu 3.06). and half of it is not connected to the water system (PPAR. para. 1.01). 3. The project became The Ministry of Mines and Water effective in October 1982 (two Resources (MMWR) is responsible for months behind schedule to allow for implementing the government policy cross-effectiveness with co- country-wide. The agency for the financiers) and civil works were capital is the Mogadishu Water launched rapidly because procurement Agency (MWA) which was the had started prior to negotiations implementation agency for the with monitoring under the First project (PPAR, para. 1.02). Project. This contributed to greater than expected initial Project Objectives, Components and disbursements (PPAR. paras. 3.01 and mplementation 3.07). The implementation of the physical components remained on 2. The main objective of schedule until 1984 when it was the project was to add onto the realized that the National Power vi Company (ENEE) would not be able to appropriate monitoring of the provide the power necessary for the financial situation. In 1985 it was Afgoy wellfield (PPAR. para. 2.03). estimated that MWA had a deficit of Advantage was taken from the unused 1002 after depreciation and contingencies to finance the interest. The continuous increase additional cost of diesel pumps. in accounts receivable, mainly from Despite the extension in scope, the the Government, offset the benefits project had a cost underrun due both of large tariff adjustments (PPAR, to good planning and management by para. 3.22). MWA failed to meet the the engineering supervisor and to financial covenants: rate of return the appreciation of the US dollar on fixed assets, accounts receivable (PPAR, para. 3.02). There was ro and debt service ratio. The rate of major problem with procurement return formula imposed in practice (PPAR, para. 3.09), but co-financing to index tariffs on the exchange created one after the Arab Fund rate which depreciated rapidly from suspended disbursemen.s in 1983 1984. The Government had, (PPAR, para. 3.08). therefore, difficulties in following IDA recommendations (PPAR. paras. 4. The strengthening of 3.13 and 3.18). Efforts to improve MWA, the water agency. was an billing and collection were important objective of the Troject. postponed repeatedly mainly because The means used were mainly the the Government owed up to eight transfer of knowledge through months of water consumption. After foreign experts in finance, the water audit documented the accounting and training appointed as seriousness of the unaccounted-for- ad-visors (PPAR, para. 3.05). As water problem and the extent of illustrated by the progress during under-billing, temporary the financial advisors' stay and the improvements were observed (PPAR, decline after their departure, the paras. 3.03 and 3.15). The terms approach was necessary to help run for on lending IDA funds to MWA were the water agency, particularly its inadequate given its uncertain accounting, but inadequate to ensure financial viability (PPAR, para. its survival after loan closing. As 3.11). a result MWA was still on cash accounting by 1990 (PPAR, para. Conclusions and Le.isons 3.16). IDA endorsed the training consultant's recommendations but did 5. The objective of the not finance a training component and project was primarily to expand the therefore did not have the proper pumping capacity and the lever to achieve long-lasting distribution network to supply more results (PPAR, para. 3.05). Ten water to the growing population of years after project appraisal, MWA Mogadishu. This physical objective was plagued by the same staffing was attained, but its sustainability pay, morale and motivation problems is in doubt due mainly to MWA's (PPAR, para. 3.21). Despite seven weakness. Managed essentially as a tariff increases in eight years "turn-key" project under the close totalling 1,567%, MWA's financial supervision of a foreign engineering situation has deteriorated consultant, capital investments were continuously since project launch completed on schedule and with cost (PPAR, paras. 3.18 and 3.22). Poor savings which is a noteworthy management combined with the absence achievement. IDA's management of of useful accounting prevented the the project implementation schedule vii was efficient. The procurement of Three components amounting to US$10 all items of civil wovks was million were cofinanced between IDA completed before the planned credit and the Arab Fund. When the latter effectiveness date. Actually it is suspended disbursements in November the delay in declaring effectiveneso 1983, contramtors and advisors were which almost set back the good not paid. IDA's offer to substitute procurement planning since several for the Arab Fund to pay the contracts called for advance advisors' salaries was refused for payments on a set date. The lesson over two years. IDA did not get from this project is that good involved with the payment arrears planning combined with close due to the contracto.s, however. supervision suffice to ensure the The lesson from teis experience is success of the physical side of a that there should be a mechanism project (PPAR, paras 4.01-4.02). whereby alternative funding is arranged on a temporary basis when 6. One exception to the one of the cofinanciers suspends otherwise flawless planning was to disbursements if the other assume that electric power would be cofinanciers decide to pursue the provided by ENEE by 1984 to run the project. As there are cross- pumps at the Afgoy wellfield. A effectiveness clauses and cross- risk analysis would have shown the default clauses, there must be drawback of relying on the power cross-financing clauses, short of company to meet WA requirements and undermining reasonable causes for thus the need to plan for and suspension, when uninterrupted finance an alternative. The project implementation is paramount additional expenditures could be such as basic power, sewer and water absorbed in the project because of infrastructures (PPAR, para. 4.04). unexpected large cost savings. Without those, the Afgoy wellfield S. The IDA funds we.e would have been non-operational onlent at terms unfit for MWA given until additional financing could be its lack of financial independence. mobilized to acquire diesel The 20-year loan maturity was too generators. There was a contrast short for a project to be between planning generously for depreciated over 33 years. Tne 3- contingencies while making a heroic year grace period was too short assumption about a key physical because it ended before input being on stream when required. commissioning when the project was The lesson to be drawn is that a key not yet generating cash flows. MWA input in the project should not be was also asked to bear the full left to another agency outside the exchange risk on the sub-loan when enforceability of the Project it did not have any means to hedge Agreement. It should instead be against other than passing it onto incorporated in the project design, the water users. The result was to even if it implies that the project practically index the tariffs on the cost is greater and the credit exchange rate which all economic amount has to be increased (PPAR, evidence pointed toward a continuous para. 4.03). depreciation. The lesson from this ?roject is that onlending IDA funds 7, The financing of the must not penalize the implementation project foreign currency agency's financ.al viability. As a expenditures was shared between IDA, benchmark for tosting the Norway, the Arab Fund and EEC. affordability of the loan terms, it viii should have been considered unlikely maintain the system built under the that with a ERR projected at only Credit. This predicament reached 5.6% MWA could service a 102 nominal the point where maintenance at the interest plus an open-ended exchange Afgoy wellfield is currently done by risk (PPAR, para. 4.05). a foreign contractor with EEC financing. Since the result was a 9. The three financial non-operational project witho4t covenants were not met and moreover addi.ional and recurrent funding could not be properly monitored from foreign sources. greater risks because no commercial accounting was conld have been taken during project in place. The First Project aimed implementation to suspend its course already to shift HWA from single- to achieve crucial institutional entry accounting limited to a cash changes. The lesson from this income statement to double-entry Project is that IDA should not be accounting which iaterfaces income shy to use a broader menu of statement and balance sheet. All pressures to obtain institutional or the efforts during implementation of financial outcomes which are more the Second Project have yielded important for the sustinability of temporary improvements which were the project than the on-schedule erased after the foreign expatriates completion of the investment portion left. By 1989, MWA had returned to (PPAR, para. 4.07). cash accounting which made practically impossible to monitor 11. The affordability of and to manage MWA's fragile Iroposed water tariff increases was viability (PPAR. para. 4.07). not an issue beyond the appraisal stage. Between 1983 and 1987 IDA 10. While efforts were staff recommended tariff increases steadily pursued by IDA to introduce totalling 9002 (on a compounded double-entry accounting, no leverage basis). During the same period the was aver applied to force the needed official consumer price index outcome. The Action Plan of 1984 increased by 623% and the US dollar was a constructive attempt at inmroased by 878Z against the Somali resolving the problem, but the shilling. The lesson from this incentive for MWA was merely the project is that, in a rapid earmarking of the salaries of inflationary environment, a good foreign experts on the Credit. This balance needs to be maintained raises the issue of appropriateness between maximizing the cost recovery of means used by IDA to effect for the implementing agency and institutional changes in a follow-up ensuring that users are not taxed project. Given the prerequisite beyond their obvious means. The nature of adequate accounting to main reason for the tariffs to be assess and to correct MWA's increased so drastically was its financial situation, stronger pegging onto a depreciating exchange pressure should have been applied. rate to enable the water agency to Disbursement suspension would have produce a constant rate of return on been appropriate although it would its assets, themselves revalued have meant delayed implementation of every year on a replacement cost the physical component. The result basis. Given that the additional is that the latter was successfully population in Mogadishu was arriving implemented from an engineering mostly from the poor rural area, the viewpoint, but that MWA is unable to full cost recovery aimed at was generate sufficient cash flows to imposing a disproportionate burden ix on the urban poor targeted by the appropriate tool for either managing project. To this extent, it or monitoring. A simpler covenant contradicted Somalia's stated policy both accepted by the borrower and to provide water at a modest cost. considered by IDA aw a valid proxy On the other hand, cost recovery is for the desirable covenant is necessary to ensure the preferable (PPAR, para. 4.09). sustainability of project benefits. Yet, implementing this principle 13. Sustainability in the .ould be spread over time when the water supply sector is too . affordability of tariffs proposal is frequently equated to the continued in doubt (PPAR, para. 4.08). ability to pump water and to distribute it to the end users. 12. When dealing with an This is a necessary condition. Yet, agency whose financial autonomy if the water agency is unable to depends on the Government's generate enough cash flows to remain willingness to pay its bills, solvent, the project benefits will covenants should be aiming at not be sustainable. The combination protecting the agercy against of unhedgeable exchange risk, lack insolvency rather than to maximize of adequate accounting, and nominal returns on assets. The two excessive accounts receivable, has financial covettants did not protect worsened MWA's financial situation IA against financial difficulties. without the appropriate remedies The ra*, of return covenant produced being offered. The paradox is that greats triff increase MWA was considerably less recommendations than was required to creditworthy by the project closing keep MWA creditworthy. The accounts date than before its start. The receivable covenant was useful only lesson from this project is that the if enforced which was not the case. financial autonomy of a water agency Letting arrears go over 8 months of is an important objective and a test sales meant insolvency for MWA. As of whether the project is a result, MWA resorted to the sustainable. Covenants and their expedient of defaulting on its debt monitoring should be geared toward service. The lesson from this achieving short-term goals of experience is that, if .the financial viability which are key to prerequisites for a covenant are not project sustainability (PPAR. para. *in place, the covenant is not an 4.10). 1/ PCR of June 27, 1985 (report No. 5761). 1 PROJECT PERFORMANCE AUDIT REPORT DEMOCRATIC REPUBLIC OF SOMALIA MOGADISHU SECOND WATER SUPPLY PROJECT (CREDIT 1236-SO) I. BACKGROUND 1.01 In an arid country like Somalia, water is essential to the population's existence. It is considered the most fundamental of basic needs by the Government whose policy is to provide water as a social service at a modest price. Water resources are limited and ill distributed; 1iere are only two perennial rivers which flow across the South. Ground water although limited in quantity and quality is the principal source of water supply. Since the First Development Plan (1970-73) urban areas, especially Mogadishu, have received priority for investment. Still, the share of planned water infrastructure investment declined from 122 to 4% in the Second Plan (1974-78) and only 50% of the projected expenditures were realized due to various difficulties (mainly drought and poor project preparation). When the First Project was launched in 1978, access to safe water was limited to 33% of the country population, of which 58% for the urban population and 20% for the rural population. Only 6 out of 38 urban centers offered piped water systems and 15% of the urban population had house connections,' the majority relying on public standpipes. In Mogadishu the situation was marginally better. While about 87% of the population was served from public supply (although 10% through individual connections), more than half of the water was used by ministries and official agencies. Public sewerage existed only in Mogadishu and yet it was limited to offices, hotels and about 3,000 residents in the central district. The First IDA Project added 8,000 m3/day (or 50%) to the declining production at Balad Road but by the time of its commission two years behind schedule (1982), demand in Mogadishu was again ahead of supply. The Second IDA Project was designed to avoid the repetition of water shortages and to satisfy the demand projected until 1987. 1.02 The Ministry of Mines and Water Resources (MMWR) through its Department of Water is responsible for providing water for both human and livestock country-wide. Many of the functions are carried out by four agencies of which the Mogadishu Water Agency (MWA) covers the capital. This Audit addresses solely the institutional aspects of the implementation agency. 1/ Initial Project Brief (December 20, 1978). The PCR estimated at only 10% the population of Mogadishu with direct connections in 1981 (para. 2.05). / PCR, para. 2.05. 2 II. THE PROJECT A. Proiect Objectives, Description and Processing 2.01 The project's objectives were to provide water to meet demand in Mogadishu until 1987 and to strengthen the water agency up to a financial self- sufficiency level. The project financed the development phase following the successful investigations of additional groundwater resources of the Afgoi-Balad- Mogadishu triangle undertaken under the First Project. It was originally considered to aim for 1990. but decided that the additional investment should be left for a follow-up project given the funding constraints on the limited project (PPAR, para. 3.06). During preparation, IDA tried to scale down further the scope of physical investments for the same reason. Savings of 20% were obtained by removing several expenditure items, including the back-up power generation at Afgoy. The Government insisted for a time to stay with the original project but eventually accepted the US$32 million cost alternative. The Arab Fund commitment to the larger project with a US$17.5 million loan, however, removed IDA's objections which were tied to the financing plan. 2.02 The physical component of the project comprised (1) the extension to the Afgoy Road wellfield to its full capacity of 34,560 m3/day, (2) the construction of a 14000 m3 service reservoir, of an elevated 450 m3 storage and . pumping station in Mogadishu, (3) the extension by 68 km of the distribution . system, and (4) the installation of 200 kw of additional power generation at Afgoy. The technical assistance component comprised (1) the appointment of two financial advisors and of a training consultant to MWA management, and (2) the supervision of the above physical component and the design of the follow-up phase. The project was processed in a short time since it was prepared under the First Project. Credit effectiveness was delayed by two months to October 12. 1982 due to cross-effectiveness clauses with the Arab Fund and the European Development Fund. B. Proiect Scope 2.03 The project scope was changed marginally during its implementation, but the additions required two amendments of the Credit Agreement. The first credit reallocation was to earmark funding for the diesel power generation facility at Afgoy (May 22, 1984). The second amendment was to allow IDA financing of the Arab Fund's share of the foreign advisors' salaries (March 18. 1986). A crucial assumption for the operation of the project after commissioning was the availability of electric power to run the pumps at the Afgoy wellfield. Although the Project Brief acknowledged that the power company (ENEE) would not meec MWA requirements by 1983-84, only a compromise solution consisting of diesel generators as back-up was recommended. The concern was then to save on total project cost to fit financing constraints. Given that a risk analysis would have shown a substantial probability for the completed project to be non-operational for lack of electric power, these were misplaced savings. Indeed, it became clear about a year before commissioning that ENEE would not provide any electricity to the Afgoy wellfield although there was sufficient warning time / Not 1986 as incorrectly stated in the PCR, para. 3.06. 3 since ENEE was supposed to have provide 500 I.W by 1981 and 2,000 KW by 1982.' The project scope was then extended to include the investment for diesel pumps generating 1920 kw and their housing. III. PROJECT IMPLEMENTATION 3.01 Implementation of the project had started before credit effectiveness because design and feasibility were a component of the First Project. Procurement started after negotiations and contracts for the physical components were awarded four months before effectiveness. This rapid project launching contributed to greater than expected initial disbursements (PPAR, para. 3.07) and to costs savings which could be used later to finance the diesel pumps, the water audit and the full salaries of the foreign advisors. All supervision missions fielded during 1983 and 1984 reported all civil works on schedule. Competent contractors and good engineering supervision provided the right combination for a successful physical implementation. Lessons from the previous one had been learned as the engineering firm was involved in both projects. The PCR seems to overstate the role played by the Project Management and Planning Unit (PMPU; PCR, paras. 3.08 and 3.37) set up within MWA in 1978 under the First Project was nominally in charge of construction supervision, but actually the tasks were performed by foreign consultants. A. Proiect Costs 3.02 Until the power generation component was reintroduced into the project, there was a large cost underrun. Eighteen months into implementation, IDA expected 25% of savings (38% in Somali shillings) "even with the just- approved cost of power supply at Afgoy".' This was due to the head-start in launching procurement, to better than expected project management which avoided tapping the contingencies and to limited exchange rate fluctuations (152 depreciation vis-A-vis the US$ between effectiveness and the first credit amendment). The PCR reported final cost in local currency to be only 2.1Z above appraisal estimates while there was a cost underrun in US$ (5.22) due to the accelerated depreciation of the Somali shilling. More relevant for the respective financiers, the local currency component registered a saving of 23% while the foreign currency component was less than 1% over budget. Given, however, that the power generation component replaced about SoSh30 million of unused contingencies at mid implementation, it can be argued that the foreign currency component was also under budget. Contingencies were relatively high with 18% of total costs and 70% for the local component alone. The rationale was to avoid the repetition of the cost overruns under the First Project which could not be financed for about two years. 4/ Telex from engineering consultant of February 3, 1981. 5/ Letter to MWA of May 22, 1984. 4 B. Water Audit 3.03 The Action Program of 1984 called for a water audit which was financed by a grant from EEC. Unexpected delays in executing the contract for supplying and installing the metering equipment justified a second extension of the closing date to end 1987. The last supervision mission (April 1987) declared the audit ready to start after many valves installed about twenty years before were repaired or replaced. The delayed completion of the audit led to a cost overrun of US$58,700 or 12% which was covered by savings on the technical assistance category. The files do not document what benefits MWA drew from this assessment of its distribution system although the push for better billing and collection toward the end of 1987 seems to be related (PCR, para. 5.08). The detailed mapping of Mogadishu (1:2,000 scale) done at this occasion was, however, an overdue tool for the management of a growing capital without updated maps since 1962. C. Low-cost Sanitation 3.04 The project was designed as a water supply stand-alone project, but it was felt during preparation that a linkage between water and sewerage had to exist. Two reasons were given: (1) technical: sewers and latetals had to follow closely the geographical layout of water piping and connections, and (2) financial: the cost of sewers would have to be recovered through the water bills.6 The lack of Government priority did not support a component in the project beyond earmarking US$300,000 for sanitation studies. A low-cost sanitation component (consisting mainly of demonstration latrines and related technical assistance) was prepared nonetheless and IDA assisted in looking for financing. Although the cost was modest (US$600,000), no foreign grant funding could be mobilized. The Government indicated before effectiveness of this Credit that it would finance it, but the commitment became less clear after.' Two years later, IDA was still prospecting for foreign assistance on behalf of the GOS without result.8 Neither the remainder of the correspondence file nor the PCR mentioned the low-cost sanitation component. D. Trainina 3.05 Training was an important priority since the First Project whose one among several priorities was to carry out a thorough assessment of MWA's training needs. Despite the obvious need, delay in appointing a consultant postponed the diagnostic while the staffing situation at MWA deteriorated. The Training Consultant Report (based on terms of reference drafted in 1977) was available only before the scheduled effectiveness of the Second Credit. The report concluded that training was no longer the prime issue. MWA was then "in urgent need of rehabilitation assistance to redress an overall state of plant 6/ Memorandum of December 29, 1980. / Telex from MWA of August 12, 1991 and letter to the Minister of Finance of December 22, 1982. a/ Supervision report of May 4, 1984. 5 malfunction or failure. A training program alone will not solve MWA's operational problems." Two alternative training programs were proposeds a comprehensive one costing US$1.2 million and a "absolute minimum" at US$737,000. IDA endorsed the second proposal but did not include any funding.10 WA favored the first recommendation for setting up an independent training facility and searched for grant financing. A request to EEC was approved after a year of waiting, but no information could be found on the outcome of this crucial component of institution building. The PCR mentions only the consideration of additional training in a follow-up project which did not materialize. Given that actual results were expected ten years earlier under the First Project, it underscores the lack of commitment toward training as the long-term solution to a sustainable water agency. The opportunity was missed to make a meaningful progress on this score. The lack of an IDA-financed component expia ins only partly this failure, however. Other issues, mainly poor salary levels. contributed to the lack of personnel motivation, thus absenteeism, high turnover and short life of physical assets. E. Financing Plan 3.06 The financing plan involved several financiers due to the project size. During preparation the project cost was reduced by US$12.8 million by IDA" to take into account the anticipated difficulty to close the financing gap. After the proposed US$15 million IDA credit, US$6 million were needed to cover the foreign cost component alone. Moreover the US$9 million of local costs were a burden on Somalia after its difficulties coming ur with the funds necessary to finance the cost overrun on the First Project' . IDA agreed to process the original US$43 million project only when the Arab Fund committed US$17.5 million in October 1981 and the EEC Economic Development Fund committed US$5 million equivalent one year later. Shortly after Norway offered to refinance half of the IDA credit as a grant. This fact was not mentioned in the PCR although it was the only IDA loan to Somalia which benefitted from such a favorable refinancing. MWA was supposed to self-finance a share of the project investment and the tariff increases recommended by IDA initially emphasized a rate of return sufficient to cover 10% of expenditures." The weighted-average cost of funds for MWA was originally 6.2%.* When the Norwegian government refinanced half the IDA credit with a grant, the cost was lowered to 4.52. The 9/ Training Consultant Report of August 18, 1982. 10/ Telex to MWA of October 19, 1982. "/ Project brief, June 28, 1980 and Issues paper, August 11, 1980, respectively. n/ Its estimate fluctuated from US$5.2 million in January 1981, down to US$3.8 million in June 1981, back up to US$7 million in February 1982. 3/ Letter to MWA of September 24, 1980. 14/ It is assumed that Government and MWA funds had to earn the 10Z long- term lending rate. 6 exchange risk has added 37% on the ovtstanding principal due IDA as of the end of 1990. The rate of return covenant was, therefore, set too low to allow MWA to generate enough cash flows to remunerate its lenders. Amount Share Interest Maturity Source US_ ml. 2 % V.a. years Arab Fund 17.5 41.7 4 20 IDA 15.0 35.7 10 20 EEC 5.3 12.6 0 grant Government/MWA 4.2 10.0 10 equity F. Disbursements 3.07 As civil works contracts were awarded before credit effectiveness, this allowed initial disbursements to be greater than projected. The first supervision report noted that they were almost double (US$ 2.9 million vs. 1.5 million) at the end of 1982 mainly due to the advances paid to contractors. The table below shows the advance taken on the anticipated disbursement profile, which contributed to the overall low project cost. During 1985 the disbursement performance seemed to falter but this was due to the savings on contingencies. (US$ million) 12/82 6/83 3184 12/85 Actual 2.9 6.0 8.5 9.5 Projected 1.5 2.9 6.2 9.9 While the PCR compares the actual disbursement profile (5.5 years) to projection (4 years), it should be corrected for the fact that a subcomponent was added in 1984. Efforts to use cost savings also contributed to stretch the disbursements profile. Yet the two closing date extensions were not granted mainly for the purpose of promoting more disbursements, but to make progress in the institution building area. It is the extension of the project scope which allowed the credit to be almost fully disbursed except for SDR 12,186 cancelled in July 1988 (PPAR. para. 2.03). G. Cofinancing 3.08 While disbursements under the IDA credit were better than 4xpected. it was the opposite for the share cofinanced by the Arab Fund (between 431 and 801 of three components). In Novemb.- 1983 it suspended disbursements because of the late payment of a service charge." This created difficulties especially for the financial advisor who received only 40% of his salary (the IDA share) from May 1984 to April 1986. IDA staff recommended that IDA finances 1001 of his salary for one year, but the Arab Fund claimed to have paid him'6 which proved to be partly incorrect since only eight months of back pay were released and the "/ IDA did not despite a 59-day delay in receiving payment on the first service charge for Credit 1236-SO. "/ Telex from the Arab Fund of M.,. 14, 1985. 7 second financial advisor was also affected." Another attempt by IDA at covering the salaries of the two advisors was refused by the Arab Fund, but again no payments were made.8 Eventually the Credit Agreement was amended to allow IDA financing the foreign cost of the advisors (March 20. 1986). Although the episode was not disruptive (the advisors continued their assignments and the contractors their civil works), it is an embarrassment to have a major cofinancier default on its funding commitment while allowing project implementation to continue without appropriate resolution. IDA did not put pressure on the Arab Fund to resume its reimbursements nor stopped the project. The PCR did not deal appropriately with the issue (PCR, para. 3.07). H. Procurement 3.09 The procurement of physical goods was problem-free mainly due to the able supervision done by the engineering consultants. For all practical purpose, they were responsible for designing the project, for justifying its feasibility, for selecting the bids and for monitoring work progress. This approach saved time as the perfect adherence to the original schedule confirms. Consequently it saved on the physical and price contingencies which were earmarked at 182 of project cost (PPAR, para. 3.02). In turn, these savings allowed IDA to expand the project scope to meet emergency needs (PPAR. para. 2.03). One contractor tried to be declared winner on two contracts by offering conditional discounts although it was the lowest bidder on only one. The episode which took place when the Credit was being approved by the Board was handled professionally by IDA staff and the engineering consultant who followed Bank guidelines. When it was decided in 1984 to finance the power generation facility at Afgoy. IDA was careful to follow accepted procedures despite the urgency. The engineering consultant did extensive international shopping to ensure the optimal choice of diesel generators given that the maintenance benefits for WA of keeping the same supplier as for the Balad Road wellfield. I. Maturity, Grace Period and Exchanze Risk 3.10 The Credit was granted to the Government of Somalia on the atandard IDA terms of 50 years of maturity including 10 years of grace and no interest but a service fee equal to .75% p.a. What is more relevant for the project, however, are the terms of the subloan granted by the Government to MWA. The subloan has a maturity of 20 years including 3 years of grace and carries an interest rate of 10% p.a.; MWA bears also the exchange risk. At the time of appraisal IDA insisted that the subloan be charged the current Bank rate then 9.62. The Government refused to apply a rate different than its usual long-term rate, then 7.5%. Without any change in the project's anticipated results, the rate was eventually increased to 10.0% in 1982.19 17/ Telex from Resident Representative of November 9. 1985. 16/ Telex to Resident Representative of January 3, 1986, from the Arab Fund of January 8, 1986, and from the Bank's Resident Representative. "/ Memorandum of February 12, 1982. 8 3.11 These subloan terms defeated the goal of financial viability which was a key component of MWA's institutional strengthening. All three lending features, maturity, grace period, and exchange risk, were unfavorable to MWA given its specific sectoral situation. The loan maturity was too short for the lumpy and long-gestation investment. Twenty years were a clear case of mismatch of maturities to finance assets to be depreciated at 3% p.a. ie. over 33 years. The grace period of only three years was also too short because it implied MWA started to repay the principal before the plant was commissioned, thus when no project-related cash flows were generated yet. MWA was also asked to bear the full exchange risk on the subloan when it did not have any means to hedge against other than passing it onto the water users. The result was to force MWA to index tariffs on the US dollar, thus raising income distribution and social equity issues. Five years after the end the grace period, the actual cost of funds of the IDA subloan for MWA increased from the nominal 10% to .% assuming no further worsening of the exchange rate. Given, however, that MWA is not repaying its debt to the Government, this additional cost is only theoretical and the exchange risk has been passed on back to the Government. By the end of 1990 the accumulated exchange risk reached US$2.7 million or 37% of the nominal debt. When repayments to IDA will start in July 1992, the servicing of the IDA credit will have to be included in the Government's general budget if MWA has not been able to start principal repayments. J. Financial Covenants 3.12 MWA's financial viability was protected and monitored by three covenants. The rate of return covenant required a positive return on fixed assets. The target were nominally low: 2% in 1983 and 1984, 3% in 1985 and 1986, and 4% thereafter. The definition was the ratio of operating income (water revenues less production costs) over average revalued net fixed assets. The accounts receivable covenant stipulated that the Government would not owe MWA more than 60 days of sales at any time. MWA was also required to meet a debt service covenant by generating cash flows to cover at least 140% of its scheduled principal repayments. 3.13 The rate of return covenant was basically a flawed approach as applied to the case of MWA. It would have been more correct to require enough free cash flows to cover the exchange risk on foreign currency borrowings which inflate the liabilities of MWA. This mandatory revaluation of liabilities should not have been offset by an equal revaluation of fixed assets allegedly at their replacement cost. Since the producing facilities contain a majority of foreign components, the depreciation of the Somali shilling make them more costly to replace. Yet, it is a theoretical situation barred a catastrophe. Annual provisions for depreciation take care of the need to replace assets at the end of their useful life, but they are not mandatory, merely a measure of financial prudence. In contrast, provisioning for the greater borrowing liabilities is both a financial reality, a Lugal obligation and an accounting requirement under IDA-financed project accounting. The liabilities increase must have a funding interface on the assets side of MWA's balance sheet, invested in either cash or 9 some forms of liquid financial instrument,20 not a portion of fixed assets totally illiquid. 3.14 There were other drawbacks in the application of the rate of return covenant. Except for the lower initial targets, no correction was made for the assets under construction which were not producing any revenues until plant commissioning (end 1985). No correction was made either for the accounts receivable if they went over the target set by the covenant. The large accounts receivable of the Government and parastatals resulted in foregone income since they increased continuously between 1982 and 1987. In the absence of financial projections no average assets was ever computed between beginning and end of year as stipulated by the legal definition. This tended to understate the required tariff increases since the exchange rate fluctuations averaged % p.a. between 1983 and 1988. On the other hand, computations available in files omitted the annual depreciation, thus inflating the recommendations. Furthermore, revaluations were applied on all assets regardless of the local costs share which may have increased more slowly than the foreign component due to the manpower content. The operational definition used by IDA staff changed also in 1987 compared to 1985: a 4% depreciation rate was used instead of 3% and the income was taken before interest payments instead of after interest in 1985.22 Overall the application of the covenant did not follow the legal definition and resulted in greater recommendations which the Government did not follow. It would have been more effective to propose lower targets, but to enforce them. Although no recommendation was fully implemented, no suspension of disbursements was ever considered. 3.15 It was the same with the accounts receivable covenant which was met only as a condition of credit effectiveness in late 1982. The rise in accounts receivable was not limited to the public sector. Lack of meters, of meter readers, of staff in the billing and collection department resulted in a 45% billing-production ratio by mid 1987.2 This figure resulted from the water audit, but a year earlier it had been estimated at 70%24 only because of MWA practice of charging 18 or 30m3 per month whenever a meter was not working. It is in mid 1986 when the physical project was completed that IDA concluded that billing and collection was "now the most critical area requiring management action"." It was estimated that about 25% of private sector sales were not billed and 50% of the individual customers were not read or rarely; there was also a large amount of payments was not posted to the correct accounts due to 20/ Similar to the interest-bearing account at the central bank proposed for the depreciation of fixed assets (Supervision report of April 10, 1987). 2'/ For example Supervision report of 1985, Attachment A, p. 2. 22/ Supervision report of April 10, 1987. 23/ PCR, para. 5.07. 24/ Supervision report of August 14, 1986. 25/ Supervision report of August 14, 1986. 10 poor accounting practi-ces. The PCR noted marked improvements toward the and of 1987 (PCR, para. 5.08), but it does not seem that they have been sustained beyond the closing date. Cimilarly the debt service covenant was declared met in all supervision reports" when it had not been the cace since February 1983. There was a contradiction in setting aside provisions for depreciation of fixed assets at 32 p.a., thus implying a life of 33 years for MWA's fixed assets and requiring MWA to repay its loan over 20 years. This discrepancy was not mentioned in the PCR although it offers a partial explanation for MWA's structural financial difficultics. K. Basic AccountirA and Auditinx 3.16 It is a clear failure of this project to have been unable to fully introduce and maintain proper accounting in MWA. The bookkeeping is done on a cash basis or single-entry accounting. For an entity managing sales to a large client base, having debts to finance capital investments, it is not adequate recording. It is not possible to monitor the true cash flows situation or whether the balance sheet structure is viable. To justify tariffs increases in such an accounting vacuum is not practical because the basic data supporting the proposed levels is missing. IDA staff was obliged to make multiple assumptions to estimate the data necessary to justify the tariff recommendations (PPAR, para. 3.18). To convert MWA's bookkeeping to "commercial" accounting was an objective of the First Project appraised in 1977. As it was an objective of the on-going First Project, the accounting conversion was neither made a condition of the Second Credit nor a covenant. As a result no lasting change took place because no legal lever could be applied. The field mission for this Audit confirmed that accounting was still in disarray thirteen years later despite the support of financial advisors, of accounting clerks and of repeated requests for improving the billing and collection department at MWA. This failure costed NWA its financial viability and is a major reason to conclude that the project is not sustainable. 3.17 No external audit report of MWA's accounts was submitted to IDA during the implementation period. The Magistrate of Accounts (MOA) which was supposed to carry them out alleged not being aware of this obligation (PCR, para. 5.12). One for the period 1982-88 was done in 1989. MOA concluded that some expenses were not justified and the balance sheets were incorrect. It recommended that the last five years be corrected by MWA itself; when it was not done one year later, MOA sent a team to correct the accounts despite its own limited staff. This step confirmed that, 18 months after the closing date, MWA was unable to keep its own accounts up to date. L. Water Tariffs 3.18 Raising water tariffs to enable MWA to self-finance 102 of the project capital expenditures was a clear priority.27 It was estimated in 1980 thet tariffs would need to be increased by 80% when the project would be 26/ For example Supervision report of September 30, 1985. 27/ Letter to MWA of September 24, 1980. 11 commissioned then scheduled for 1984. Such an increase seemed insufficient given the inflation rate observed in 1980 (about 752), but appeared large given the targeted population of urban poor. IDA staff recommended the completion of the tariff study (scheduled for September 1979 under the First Project) as well as a proposed timetable for the tariff increases be made a condition for neotiations." The conditionality was hardened later by requiring an "interim" tariff increase prior to negotiations and making the tariff study a condition of effectiveness. The first tariff increase since 1973 was enacted in May 1981. Although large at 50% to SoSh4.5/m3, it did not offset the 457% inflation during the previous eight years. As a result, it was estimated before negotiations that increases of 84% would be needed in 1983, 312 in 1984, and 26% in 1985 to meet HWA's commitment to the project financing plan.' No specific increase was, however, required as condition of effectiveness or dated covenants. A tariff proposal for 1983 was requested following the tariff study,30 but the Government waited until after effectiveness to raise tariffs to SoSh7 rate (+55.5%) when SoSh9 were needed.31 More significantly the Government refused the study recommendation for progressive rates tied to volumes consumed which would have implied steep increases for the Government since it consumed then 70% of the water sold in Mogadishu." The failure of IDA to insist on the full implementation of the long-awaited tariff study meant financial difficulties for MWA and explains its inability to meet the rate of return covenant from then on. 3.19 The following table compares the tariff increases enacted with those recommended by IDA staff and with inflation and exchange rate fluctuations. It shows that increases followed more closely the domestic inflation rate than either IDA recommendations or the Somali shilling depreciation. Tariff Increase IDA Recommenda- Inflation Exchange Rate Year SoSh % tion % % Channe 2 May 1981 4.5 50 n.a. 457 n.a. Aug.1983 5.0 11 56 84 148 Aug.1984 10.0 100 30 84 22 May 1985 14.0 40 160 35 106 Oct.1985 18.0 29 n.a. -1 4 Apr.1986 25.0 39 28 41 74 Dec.1988 25/75" 200 80 200 293 21/ Issues Paper of August 11, 1980. / Memorandum of February 12, 1982. 30/ Telex to MWA (through the Resident representative) of July 15, 1982. "/ Letter to MWA of February 9, 1983. 32/ Supervision Report of February 16, 1983. 3'/ SoSh25 for public fountain kiosk users and SoSh75 for other users. 12 The main reason supporting the frequent IDA requests for water tariffs increases was the need to meet the rate of return covenant (?PAA. para. 3.13). This emphasis while legitimate was carried too far with a clear disregard for the affordability of the proposed tariffs. Not once was a recommendation justified on the basis of the water users' incomes while the PCR claims a substantial beneficial impact on the urban poverty ,roup.3. IDA even challenged the assumptions underlying the tariff study, calling unrealistic the income growth projected as low as 1/6 of the projected inflation. This scenario was however realized. It did not imply that the ability to pay for water services did not exist as water vendors were able to charge SoSh3,O0O/m3 when NWA tariff was only SoSh25. The revaluation of fixed assets contained in the rate of retturn covenant formula forced IDA to favor an indexing of tariffs on the exchange rate. The discrepancy with the tariff increases called for by an affordability analysis widened. This explains why the Government lagged considerably behind IDA recommendation in enacting the actual increases which rarely exceeded the local inflation rate. 3.20 It seems also that double counting produced excessively high recommendations. The rate of return covenant stipulated the operating income as numerator of the ratio. The recommendation of 1985 required, however, the 3% return after generating another 3% for depreciation of fixed assets whereas the covenant required only 3% including depreciation. This approach resulted in a figure about double what was actually needed.35 As a result, it was not followed. M. Institutional Development 3.21 The strengthening of MWA was an important objective, but the improvements were not sustained. The institutional weakness of MWA was known before the First Project whose effectiveness was conditional upon appointing foreign advisors to MWA management. MWA's poor performance was then "due mainly to unqualified staff, low morale, and lack of motivation"." This diagnostic was still correct at the time of the field mission for this Audit. The means used to improve MWA's effectiveness were mainly the transfer of knowledge through foreign experts appointed as advisors. Before the start of this project, IDA staff recommended more than doubling foreign advisory assistan-ce, but only appointing a tariff consultant and a training consultant was made a condition of negotiation.38 As illustrated by the progress during their / PCR, para. 2.05. 5/ The figure recommended was greater due to a mistake in computing the increase in percentage which should have been reported as 61% not 159.7% (Supervision Report, Appendix A, March 12, 1985). 3/ PCR for the First Mogadishu Water Supply Project, para. 3.01 (June 27, 1985). 37/ Project Brief, June 28, 1980. 38/ Decision Memorandum of August 28, 1980. 13 stay and the decline after their departure, the approach wae necessary to help run the water agency, but inappropriate to ensure its survival. IDA endorsed the training consultant's recommendations, but it did not finance a training component and therefore did not have the proper lever to achieve long-lasting results. Ten years after project appraisal, MWA was plagued by the same staffing pay, morale and motivation problems recognized then. Government officials were convinced Lhat pay increase is the solution. While catching up with inflation is neces&ary, staff productivity and efficiency depends on changing absenteeism and other habits which havL led to failures to take simple responsibilities. N. MWA Financial Situation 3.22 MWA' s financial situation has deteriorated continuously since psroject launch whereas it was considered "satisfactory until the end of 1981". ' IDA monitoring relied on indicators generally associated with institutions on the verge of bankruptcy such as cash in bank, cash available to meet debt service, and accounts receivable. The tone of IDA's assessment changed after effectiveness. The first supervision report noted early 1983 MWA's "inability to produce any surplus to finance capital spending".0 It was actually worse when IDA reported that MWA had an estimated operating deficit in 1982: SoSh8 million or 35%." The same supervision report anticipatd a break-even in 1983 and the following report a SoShl5 millior. surplus in 984." By early 1985, however, the operating deficit reached 60% of revenues. 11oreover, the Government arrears had increased from about 65 days to 200 in 18 moAths, thus worsening the actual deficit on a cash basis. Final figures for 1985 showed a deficit of SoSh17 million, double the previous estimate. On a commercial accounting basis, it was estimated to be equivalent to SoSh69 million after depreciation and interest on revenues of SoSh69 million. This meant that 1WA was covering either its operating expenses or its debt service in principal and interest, not both. In 1986 SoSh20 million of surplus were made and IDA recommended that it be set aside in a capital asset replacement fund instead of avoiding a tariff increase." The latter solution was, however, retained and the next increase took place in December 1988. After the departure of the financial and accounting advisors, MWA bookkeeping has relapsed and it was not possible to obtain updated financial data for this Audit. Officials confirmed, however, that MIA's finances were in disarray. 39/ Back-to-office report of January 5, 1983. "'/ Back-to-office report of January 5, 1983. 41/ Back-to-office report of April 19, 1983. 42/ Supervision reports of April 19, 1983 and July 31, 1984, respectively. 43/ Supervision report of August 14, 1986. 4/ Supervision report of April 10, 1987. 14 IV. LESSONS LEARNED 4.01 The objective of the project was primarily to expand the pumping capacity and the distribution network tn supply more water to the growing population of Mogadishu. In physical terms, this objective was attained. Its sustainability is in doubt, however, due to a combination of adverse factors and institutional weakness. This case typifies the difficulty to qualify a project which has been successful in its physical implementation but a failure in terms of institution building and financial sustainability. Managed essentially as a "turn-key" project under the close supervision of an engineering consultant, capital investments were completed on schedule and with cost savings which is a noteworthy achievement. 4.02 IDA's management of the project implementation schedule was efficient. The procurement of all items of :ivil works was completed before the planned credit effectiveness date. Actually it is the delay in declaring effectiveness which almost set back the good procurement planning since several contracts called for advance payments before a specified date. The lesson from this project is that good planning combined with close supervision suffice to ensure the success of the physical side of a project. 4.03 One exception to the otherwise flawless planning was the crucial assumption that after commissioning electric power would be available to run the pumps at the Afgoy wellfield. A risk analysis would have shown the drawback of relying on the power company to meet MWA requirements and thus the need to plan for and finance an alternative. The project scope could be extended when the worst scenario was realized because of unexpected large cost savings. If this had not been the case as the cost overrun under the First Project led to anticipate, the Afgoy wellfield would have been non-operational until additional financing could be mobilized. It is a disturbing anomaly to plan generously for contingencies while making a heroic assumption about a key physical input being on stream when required. The lesson to be drawn is that a key input in the project, whether energy to run the plant or spare parts to maintain it, should not be left to another agency outside the enforceability of the Project Agreement. I )hould instead be incorporated in the project design, even if it implies that the project cost is greater and the credit amount has to be increased. It is less costly for the country to cancel unused funds than to be faced with non-operational assets due to optimistic planning. 4.04 The financing of the project foreign currency expenditures was shared between IDA, Norway, the Arab Fund and EEC. Three major components were cofinanced between IDA and the Arab Fund. When the latter suspended disbursements in November 1983, contractors and consultants were not paid. IDA's effort to substitute for the Arab Fund to pay the advisors' salaries failed twice and it is only over two years later that the Credit Agreement was amended to allow this. IDA did not get involved with the payment arrears to the contractors, however. The lesson from this experience is that there should be a mechanism whereby alternative funding is arranged on a temporary basis when one of the cofinanciers suspends disbursements if the other cofinanciers decide to pursue the project. As there are cross-effectiveness clauses and cross-default clauses, there must be cross-financing clauses, short of undermining reasonable 15 causes for suspension, when uninterrupted project implementation is paramount such as basic power, sever and water infrastructures. 4.05 Although nominally an autonomous agency, MWA was not endowed with the necessary and sufficient means to be self-sufficient. The sub-loan terms were unfit for HWA given the reality of its financial independence. The 20-year loan maturity was too short for a project to be depreciated over 33 years. The 3-year grace period was too short because it ended before commissioning when the project was not yet generating cash flows. MWA was also asked to bear the full exchange risk on the sub-loan when it did not have any means to hedge against other than passing it onto the water users. The result was to practically index the tariffs on the exchange rate which all economic evidence pointed toward a continuous depreciation. The lesson from this project is that onlending IDA funds must not penalize the implementation agency's financial viability. As a benchmark for testing the affordability of the loan terms, it could have been considered unlikely that with a ERR projected at only 5.6% MWA could service a 101 nominal interest plus an open-ended exchange risk. 4.06 The three financial covenants were not met and moreover could not be properly monitored because no commercial accounting was available. It was one of the objectives of the First Project to shift MWA from single-entry accounting limited to a cash income statement to double-entry accounting which interfaces income statement and balance sheet. It was acknowledged prior to the First project that without this tool it was meaningless to try and monitor MWA's financial situation. All the efforts during implementation of the Second Project have yielded temporary improvements which were erased after the foreign expatriates left. At the time of the field mission for this Audit, MWA had returned to unsystematic t:ash accounting. The main consequences of the lack of useful accounting have been to render (i) meaningless the provisioning for depreciation of assets in service, (ii) baseless the formulation of appropriate tariff levels, a-ad (iii' moot the commitment to service of IDA credits onlent by the Government to MA since no balance sheet was produced. 4.07 While efforts vere steadily pursued by IDA to introduce double-entry accounting, no leverage was ever applied to force the needed outcome. The Action Plan of 1984 was a constr Ective attempt at resolving the problem, but the incentive for MWA was merely the earmarking of the saaries of foreign experts on the Credit. This raises the issue of appropriateness of means used by IDA to effect institutional changes in a follow-up project. Given the prerequisite nature of adequate accounting to assess and to correct MWA's financial situation, stronger pressure should have been applied. Disbursement suspension would have been appropriate although it would have meant delayed implementation of the physical component. The result is that the latter was successfully implemented from an engineering viewpoint, but that MWA is unable to generate sufficient cash flows to maintain the system built under the Credit. This predicament reached the point where maintenance at the Afgoy wellfield is currently done by a foreign contractor with EEC financing. Since the result was a non-operational project without additional and recurrent fundings from foreign sources, greater risks could have been taken during project implementation to suspend its course to achieve crucial institutional changes. The lesson from this Project is that IDA should not be shy to use a broader menu of pressures to obtain institutional or 16 financial outcomes which are more important for the sustainability of the project than the on-schedule completion of the investment portion. 4.08 The affordability of proposed water tariff increases was not an issue beyond the appraisal stag ?. Between 1983 and 1987 IDA staff recommended tariff increases totalling 900% (on a compounded basis). During the same period the official consumer price index increased by 623% and the US dollar increased by 878% against the Somali shilling. The lesson from this project is that, in a rapid inflationary environment, a good balance needs to be maintained between maximizing the cost recovery for the implementing agency and ensuring that users are not taxed beyond their obvious means. The main reason for the tariff. to be increased so drastically was its pegging onto a depreciating exchange rate to enable the water agency to produce a constant rate of return on its assets, themselves revalued every year on a replacement cost basis. Given that the additional population in Mogadishu was arriving mostly from the poor rural area, the full cost recovery aimed at was imposing a disproportionate burden on the urban poor targeted by the project. To this extent, it contradicted Somalia's stated policy to provide water at a modest cost. On the other hand, full cost recovery is a key feature of IDA financed projects to achieve positive financial rates of return and there should no exception to the principle. Yet its implementation could be spread over time when the affordability of tariffs proposal is in doubt. 4.09 When dealing with an agency whose financial autonomy depends on the Government's willingness to pay its bills, covenants should be aiming at protecting the agency against insolvency rather than to maximize nominal returns on assets. The two financial covenants contained in the Credit Agreement did not protect MWA against financial difficulties. The rate of return covenant produced greater tariff increase recommendations than was required to keep MWA creditworthy. The accounts receivable covenant was useful only if enforced which was not the case. Letting receivables go over 8 months of sales meant insolvency for MWA. As a result, MWA resorted to the expedient of defaulting on its debt service. The lesson from this experience is that, if the prerequisites for implementing a covenant are not in place, the covenant is not an appropriate tool for either managing or monitoring. A simpler covenant both accepted by the borrower and considered by IDA as a valid proxy for the desirable covenant is preferable. 4.10 Sustainability in the water supply sector is too frequently equated to the continued ability to pump water and to 4atribute it to the end users. This is a necessary condition. Yet, if the water agency is unable to generate enough cash flows to remain solvent, the project benefits will not be sustainable. The combination of unhedgeable exchange risk, lack of adequate accounting, and excessive accounts receivable, has worsened MWA's financial situation without the appropriate remedies being offered. The paradox is that MWA was considerably less creditworthy by the project closing date than before its start. The lesson from this project is that the financial autonomy of a water agency is an important objective and a test of whether the project is sustainable. Covenants and their monitoring should be geared toward achieving short-term goals of financial viability which are key to project sustainability. 17 4.11 This water supply project illustrates the dilemma of development lending. Additional investments in the water sector were urgently needed to avoid the recurrence of water shortages. On the other hand, the magnitude of tariff increases needed tm meet the initial self-financing comitment and the debt service later was hardly aff,rdable to the 37Z of urban pQor targeted by the project. Compromises on the pace and amount of tariff increases led the implementation agency to face a growing insolvency problem. The lack of parallel progress between strengthening the agency and developing its water production and distribution facilities resulted in a project whose physical achievement are unsustainable without further outside assistance.