Tanzania - First and Second Songo-Songo Petroleum Exploration Projects and Petroleum Sector Technical Assistance Project

Document of The World Bank FOR OFFICIAL USE ONLY Report No. 15802 PERFORMANCE AUDIT REPORT UNITED REPUBLIC OF TANZANIA SONGO-SONGO PETROLEUM EXPLORATION PROJECT (Credit S-027 TA) SECOND SONGO-SONGO PETROLEUM EXPLORATION PROJECT (Credit 1199-TA) PETROLEUM SECTOR TECHNICAL ASSISTANCE PROJECT (Credit 1604-TA) June 25, 1996 Operations Evaluations 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 Currencv Unit = Tanzanian Shilling (Tsh) 1984 US$1.00 Tsh 17.0 1992 US$1.00 Tsh 300.0 Abbreviations and Acronyms bcf billion cubic foot CNG Compressed Natural Gas ElB European Investment Bank GOT Government of Tanzania IDA International Development Association IMF International Monetary Fund IOC International Oil Company kWh kilowatt-hour LPG Liquid Petroleum Gas mscf thousand standard cubic feet MW megawatt MWEM Ministry of Water, Energy and Minerals OED Operations Evaluation Department OPEC Organization of Petroleum Exporting Countries PAR Performance Audit Report PCR Project Completion Report PIl Project Implementation Unit PR President's Report PSC Production Sharing Contract SDR Special Drawing Rights TA Technical Assistance TANESCO Tanzania Electric Supply Corporation TPDC Tanzania Petroleum Development Corporation UNDP United Nations Development Programme Measurement kin Kilometer Im meter TOE ton of oil equivalent ton metric toil Fiscal Year Government: July I - June 30 TPDC: January 1 - December 31 FOR OFFICIAL USE ONLY The World Bank Washington, D.C. 20433 U.S.A. Office of the Director-General Operations Evaluation June 25, 1996 MEMORANDUM TO THE EXECUTIVE DIRECTORS AND THE PRESIDENT SUBJECT: Performance Audit Report on Tanzania Songo-Songo Petroleum Exploration Project (Credit S-27-TA), Second Songo-Songo Petroleum Exploration Project (Credit 1199-TA) and Petroleum Sector Technical Assistance Project (Credit 1604-TA) Attached is the Performance Audit Report (PAR) prepared by the Operations Evaluation Department (OED) on the above projects approved, respectively, in FYs 80, 82 and 85. Total approved credits were SDR 55.5 million, of which all but SDR 1.8 million was disbursed. The main objective of the two Petroleum Exploration projects was to assess the oil and gas potential of the Songo-Songo field and thus help Tanzania develop a domestic supply of hydrocarbons. These two back-to-back projects comprised a US$75 million, five-well drilling program. They failed to find oil, but established proven reserves of over 700 billion cubic feet of natural gas by the end of the second project in 1984. The Technical Assistance (TA) project was implemented primarily to finance urgently needed repairs on some of the Songo-Songo wells (drilled prior to Bank involvement), which were leaking gas. There were also several unrelated project components, including support for a further exploration promotion effort. The promotion activity was reasonably successful in inducing exploration investment by international oil companies (lOCs). There has been no success so far, but lOCs continue to look for oil on a modest scale. Unfortunately, Tanzania's internal conflicts about the best use of the gas (fertilizers, industry, or private or public run power generation) have stalemated efforts to utilize these reserves during the past twelve years, even though the cost of gas would have been competitive and the country has, in recent years, been suffering from severe electricity shortages. The Region is currently working on a private sector program that would utilize the gas for power generation, but in the meanwhile, the well infrastructure is continuing to deteriorate. The Audit concluded that the outcome of the petroleum projects was marginally satisfactory, but that their sustainability was uncertain. Bank and Borrower performance and institutional development were all assessed as unsatisfactory for the exploration projects, but satisfactory for the TA project. In the past, IDA's energy sector policy dialogue with the Tanzanian Government has been fragmented by subsectors. IDA's nine energy projects during the past 15 years were insufficiently harmonized to ensure an optimum use of available domestic resources. The major lessons from these three projects is that when economically recoverable gas reserves are found, every effort should be made to ensure that they are exploited as soon as possible to generate revenues, create economic benefits for the country, and mobilize domestic resources. 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.  1 FOR OFFICIAL USE ONLY Contents Preface ... ....................................................... 3 Basic Data Sheet.................................................... 5 Evaluation Summary ......................................... ....... 13 1. Introduction .................................................... 19 Macroeconomic Context..................... ............... 19 Sectoral Context................................................. 19 2. Projects' Objectives and Descriptions .............................. .....21 Songo-Songo Petroleum Exploration Projects ...............................21 Petroleum Sector Technical Assistance Project..................l........21 3. Implementation .......................................................23 Songo-Songo Exploration Projects .................................. .....23 Petroleum Sector Technical Assistance Project .......................... 24 4. Issues ................................................................ 27 Project Management ........................................... ..... 27 Choice of Lending Instrument ..................................... .....27 Neglect of Scope for Gas Utilization..................................... 27 Gas Utilization and Power Planning .......................................... 28 5. Conclusions .......................................................... 33 Project Outcome and Impact ........................................... 33 Bank and Borrower Performance ........................................ 33 Project Ratings .................................................... 34 6. Lessons ............................................................. 37 Attachment Comments from the Borrower ............................................... 39 This report was prepared by Mr. Sunil Mathrani (Consultant), who audited the projects in August 1995, under the supervision of Richard Berney (Task Manager), with administrative assistance provided by Ms. Lorna Sibblies and Ms. Helen Watkins. The report was issued by the Infrastructure and Energy Division (Yves Albouy, Chief) of the Operations Evaluation Department (Francisco Aguirre-Sacasa, Director). 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 wiLhout World Bank authorization.  3 Preface This is a Performance Audit Report (PAR) on the above Bank Credits, respectively for US$30 million, SDR 17.4 million and SDR 8.1 million to the United Republic of Tanzania, approved between June 1980 and June 1985. The first two loans were fully disbursed. The third disbursed SDR 6.3 million. SDR 1.8 million was canceled. The PAR was prepared by the Operations Evaluation Department (OED) and was based on the Project Completion Reports, the President's Reports, project files and discussions with Bank staff. Prior to preparing the PAR, an OED mission visited Tanzania in July 1995 and discussed the results of these projects with Government officials and representatives of energy sector entities. Their cooperation and assistance is gratefully acknowledged. The PAR reassesses the projects' achievements and the performance of the Bank and the Borrower, not only in the execution of these projects but also with respect to the broad objectives of least-cost development for the energy sector and the scope for gas utilization. The draft PAR was sent to the Borrower for comments. The comments received from the Tanzania Electric Supply Company Limited are reproduced as an Attachment to the PAR. r 5 Basic Data Sheet PETROLEUM SECTOR TECHNICAL ASSISTANCE PROJECT (CREDIT 1604-TA) Key Project Data (amounts in Us$ million) Appraisal Actual or Actual as % of estimate current estimate appraisal estimate Total project costs 11.0 11.0 Loan amount 8.1 6.3 Cancellation 1.8 Cumulative Estimated and Actual Disbursements CYSemester Cumulative Disbursements Actual/Forecast US$ million % Forecast Actual 1986 4.0 2.4 30 1987 6.5 2.7 34 1988 7.5 2.7 34 1989 8.0 4.5 56 1990 - 6.7 84 1991 - 7.0 88 1992 - 7.6 95 1993 - 8.0 100 Project Dates Original Actual Identification 07/84 Preparation 10/84 Appraisal 12/84 11/84 Negotiation 04/85 04/85 Board Approval 06/85 06/06/85 Signing 07/85 07/11/85 Effectiveness 08/85 10/09/85 Completion 12/31/88 12/31/92 Closing date 06/30/89 12/31/92 6 Staff Inputs (staff weeks) 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 Preappraisal 20.4 26.6 Appraisal 45.4 Negotiations 1.0 Supervision 23.2 21.9 16.2 6.8 7.1 3.5 6.0 1.6 Total 20.4 73.0 23.2 21.9 16.2 6.8 7.1 3.5 6.0 1.6 Mission Data Date No. of Staff days Specializations Performance Types of (month/year) persons infield represented rating problems Identification/ 09/84 1 n.a. Petroleum Preparation Engineer Appraisal 11/84 5 21 Financial Analyst, Petroleum Engineer, Petroleum Specialist, Institution Strengthening Specialist, Gas Specialist Supervision 10/85 3 5 Economist, I Petroleum Engineer, Financial Analyst 12/85 2 2 n.a.* P 06/86 3 Petroleum n.a.* Specialist, Petroleum Engineer 11/86 1 5 Geologist n.a.* 03/87 1 n.a. Financial 1 Analyst 10/87 3 8 Economist, 1 Financial Analyst, Geologist 07/88 3 20 d/ Economist, 2 M Petroleum Engineer, 7 Date No. of Staff days Specializations Performance Types of (month/year) persons infield represented rating problems Financial Analyst 11/88 * 03/90 1 9 d/ Economist, 2 M Petroleum Engineer, Financial Analyst 07/91 2 4 Financial 2 M, L Analyst, Petroleum Engineer 10/91 3 7 d/ Financial 2 M Analyst, Petroleum Engineer, Economist 08/92 1 5 d/ Financial n.a.* Analyst Other Project Data Borrower/Executing Agency: FOLLOW-ON OPERATIONS Operation Credit No. Amount Board Date (US$ million) Petroleum Sector Rehabilitation 2202-TA 30.0 1991 Songo-Songo Petroleum Exploration S/27-TA 30.0 1980 Second Songo-Songo Petroleum Exploration 1604-TA 44.8 1981 8 SONGO-SONGO PETROLEUM EXPLORATION PROJECT (CREDIT S/27-TA) Key Project Data (amounts in USS million) Appraisal Actual or Actual as % of estimate current estimate appraisal estimate Total project costs 30.0 Loan amount 30.0 30.0 100.0 Cancellation 0.0 Cumulative Estimated and Actual Disbursements FY 81 FY 82 FY83 Appraisal Estimate 25,000 28,000 30,000 Actual 12,671 29,250 30,000 Actual as % of Appraisal 51 104 100 Date of Final Disbursement: 9/7/82 Project Dates Original Actual Identification 9/79 Appraisal 1/80 Negotiation 5/8/80 Board Approval 6/12/80 Signing 6/30/80 Effectiveness 9/80 9/12/80 Completion 6/83 9/7/82 Closing date 9/30/82 9/30/82 9 Staff Input Date No. of Specializations Performance (month/year) persons Staff Weeks' Represented rating2 Preparation I 79 3 3.2 Economist 0.1 Geologist 0.6 Consultant Appraisal 80 6 0.3 Advisor 1.4 Chief 16.4 Economist 0.5 Engineer 0.5 Financial Analyst 4.9 Geologist Negotiations 80 4 2.1 Economist 0.4 Financial Analyst 2.3 Geologist 0.1 Technical Preparation II 80 5 4.0 Chief 0.3 Deputy Chief 0.7 Economist 0.2 Financial Analyst 8.7 Geologist Lending Operations 80 2 4.0 Loan Officer Supervision SPN 1 0.4 Economist Project Admin. 81 2 2.8 Loan Officer 0.7 Operations Asst. Supervision 81 9 0.6 Loan Officer 0.4 Chemical Eng. 0.9 Deputy Chief 15.1 Economist 25.2 Engineer 0.6 Financial Analyst 4.2 Geologist 0.1 Operations 0.3 Technical Project Admin. 82 2 0.8 Loan Officer 0.1 Oper. Asst. Supervision 82 5 0.1 Asst Director 2 9.6 Economist 22.4 Engineer 0.5 Financial Analyst 2.5 Geologist Project Admin. 83 1 1.1 Oper. Asst. 2 Supervision 83 3 3.2 Economist 2 1.2 Engineer 0.1 Geologist Project Admin. 83 1 0.2 Oper. Asst. 2 Supervision 83 3 3.2 Economist 2 1.2 Engineer 0.1 Geologist Project Admin. 84 1 0.2 Oper. Asst. 2 Supervision 84 4 0.0 Advisor 2 0.1 Economist 3.3 Engineer 0.0 Financial Analyst Supervision 85 4 0.0 Economist 2 1.8 Engineer 0.1 Geologist 0.9 Engineer Supervision 86 1 0.1 Geophysicist PCR 89 1 0.1 Geophysicist PCR 89 I 1.9 Geophysicist TOTAL 153.2 1. Data inadequate to split staff time in the field and headquarters 2. 1) Problem free or minor problem. 2) Moderate problem. 3) Major problem. 10 Other Project Data Borrower/Executing Agency: FOLLOW-ON OPERATIONS Operation Credit No. Amount Board Date (US$ million) Petroleum Sector Rehabilitation 2202-TA 30.0 1991 Second Song-Songo Petroleum Exploration 1199-TA 44.8 1981 11 SECOND SONGO-SONGO PETROLEUM EXPLORATION PROJECT (CREDIT 1199-TA) Key Project Data (amounts in US$ million) Appraisal Actual or Actual as % of estimate current estimate appraisal estimate Total project costs 44.8 38.7 Loan amount 30.0 30.0 100.0 Cumulative Estimated and Actual Disbursements FY 82 FY83 FY84 FY85 Appraisal Estimate for Total 32,000 40,000 40,000 40,000 Donor Finance IDA Portion 16,000 20,000 20,000 20,000 Actual for IDA Disbursement 8,424 16,983 18,682 18,718 Actual as % of Estimate 52.6 84.9 93.4 93.6 Date of Final Disbursement: 9/10/84 Project Dates Original Actual Identification 6/81 Preparation 8/81 Appraisal 9/81 Negotiation 11/81 Board Approval 11/22/81 Signing 1/1182 Effectiveness 1/82 2/23/82 Completion 6/30/84 9/10/84 Closing date 12/31/83 12/31/83 12 Staff Input Date No. of Specializations Performance (monthlyear) persons Staff Weeks' Represented rating2 Through Appraisal Preparation 1 80 1 3.4 Loan Officer 81 3 0.0 Deputy Div. Chief 6.1 Economist 0.2 Geologist Appraisal 82 3 0.3 Asst. Vice Pres. 8.6 Economist 0.5 Financial Analyst 82 1 6.2 Loan Officer Negotiations 82 1 0.1 Economist Preparation II 82 3 1.1 Division Chief 1.2 Dep. Division Chief 2.2 Economist 82 1 0.4 Engineer Proj. Administration 82 2 0.3 Loan Officer 0.1 Operations Asst. Supervision Supervision 82 5 0.1 Division Chief 10.3 Economist 5.2 Engineer 0.1 Financial Analyst 0.6 Geologist Supervision 83 1 0.4 Engineer Project Admin. 83 2 0.7 Loan Officer 0.2 Operations Asst. Supervision 83 4 11.8 Economist 17.7 Engineer 0.5 Financial Analyst 1.3 Geologist Project Admin. 84 1 1.1 Operations Asst. Supervision 84 4 0.2 Division Chief 8.3 Engineer 7.7 Economist 0.0 Engineer Project Admin. 85 1 1.0 Loan Officer Supervision 85 1 0.2 Dep. Division Chief I Supervision 86 2 0.6 Engineer 1 0.4 Engineer PCR 89 1 1.0 Geophysicist 1 PCR 90 2 0.0 Country Economist 1 0.1 Economist Supervision 90 1 0.2 Engineer TOTAL 113.4 1. Data inadequate to split staff time in the field and headquarters 2. 1) Problem free or minor problem. 2) Moderate problem. 3) Major problem. 13 Other Project Data Borrower/Executing Agency: FOLLOW-ON OPERATIONS Operation Credit No. Amount Board Date (US$ million) Petroleum Sector Rehabilitation 2202-TA 30.0 1991 哺 15 Evaluation Summary Sectoral Context I . Tanzania has no indigenous oil supply. Petroleum accounted for 30 percent of total imports in the early 1980s, when these projects were being implemented. Consumption had risen to 0.8 million tons in 1993, but now represents less than 10 percent of the value of Tanzania's total imports. Exploration activity for oil by foreign private firms continues on a modest scale, but there have been no discoveries to date. Gas was first discovered in the Songo-Songo field in the mid- I 970s by an international oil company (IOC). The lease was relinquished because the IOC believed that it could not develop the gas profitably. Proven natural gas reserves to date amount to almost 750 billion cubic feet (bcf), but they have yet to be developed commercially. The Bank Group has been working on establishing a private sector financed gas fired power project since 1992. Projects' Objectives and Descriptions Songo-Songo Petroleum Exploration Projects 2. The objective of the First and Second Songo-Songo Petroleum Exploration projects was to assess the oil and gas potential of the Songo-Songo field and thus help Tanzania develop a domestic supply of hydrocarbons. The first project was for two wells, and the second, which was approved eighteen months later, was for another three. The five-well drilling program cost almost US$80 million. The bulk of foreign financing came from the Bank (US$49 million), while the European Investment Bank (EIB) and the Organization of Petroleum Exporting Countries Fund (OPQ together provided a further US$20 million in cofinancing. Petroleum Sector Technical Assistance Project 3. After the completion of the Songo-Songo drilling program, the Bank supported a third petroleum sector project in 1985. The initial rationale for this project was the urgent repairs needed to stem the gas leaks at two Songo-Songo gas wells drilled in the 1970s, prior to Bank involvement. These gas leaks endangered life, property and the future development of the gas field in which the Bank had been heavily involved. The inability of the Government to find alternative funding for the repairs provided the initial impetus for the project. However, the Bank staff also drew on the recently-completed Energy Assessment Report to flesh out the project by including several unrelated components with diverse objectives, including alleviating petroleum product distribution bottlenecks, inducing further petroleum exploration efforts by IOCs, and developing a strategy for the utilization of identified gas resources. 4. The components of this US$I I million project were: (1) support for gas development through emergency well repairs, construction of protective structures for unprotected offshore wells, a gas utilization study for Songo-Songo gas and, if feasible, a pilot project for using cornpressed natural gas (CNG) in public transport; (ii) upgrading the petroleum product distribution system with tankers, pumps and liquid petroleum gas (LPG) bottles, and a product 16 distribution study; (iii) preparation of a petroleum exploration promotion undertaking; and (iv) institutional development. Project Implementation and Results 5. The exploration projects failed to find oil, but by the time the Second Credit was fully disbursed in late 1984, gas reserves of over 700 bcf had been proven in the Songo-Songo field. Project implementation suffered from technical and logistical problems. As explained in the PCR, the deficiency became apparent under the first project, but was not corrected under the second. Instead, Bank technical staff became directly involved in drilling decisions and project management. The project management problems also had a major impact on the total cost, which was much higher than envisaged at the time of appraisal of the first project. A four-well program costing US$45 million became a five-well drilling program with a final cost of nearly US$80 million at the end of the second project. Implementation delays, poor project management, technical problems and poor sitting of wells all contributed to the higher costs. The extra funding required came from the second, follow-on project. 6. The Petroleum Sector TA Project was implemented over seven and a half years, from 1985-1993. The Bank financed the full foreign exchange cost of US$8 million. In addition to financing the emergency well repairs, which were carried out expeditiously, the project assisted in the area of petroleum exploration promotion and in implementing a study on rationalizing Tanzania's petroleum products distribution system. Geological and geophysical data that was acquired was presented to the international oil industry to seek exploration proposals. The promotion effort was successful: three production sharing contracts (PSCs) were signed with two companies during the past five years. Unfortunately, these exploration ventures failed to discover any commercially exploitable reserves. Several new production sharing contracts were signed in 1994 and 1995, after Tanzania improved the contract terms to meet stiff international competition, and further agreements are under negotiation, suggesting that its longer-term prospects are picking up. The petroleum product distribution system study was instrumental in the implementation of a followup Petroleum Sector Rehabilitation Project, which focuses on expanded private sector participation. 7. The institutional development component lacked a clear plan and was implemented in an ad hoc fashion, primarily through short term courses that, according to TPDC, were of marginal value. TPDC's long-term overseas training program, even limited in the extent that it was, appears to have had better results. The Audit mission found no evidence to support the PCR's conclusion that this component was partially successful and sustainable. Two major components of the project, protective platforms for offshore gas wells ($1.9 million), and the CNG demonstration scheme ($2.5 million), together representing 40 percent of the original project cost, were dropped entirely. Issues A. Project Management 8. Both of the Songo-Songo exploration projects suffered from the same design weakness, namely, inadequate project management arrangements. The implementing agency, Tanzania Petroleum Development Corporation (TPDC), was inexperienced, and a fully staffed project 17 implementation unit (PIU) with the necessary expatriate TA with clearly defined responsibilities was required. This was not a feature of either project. Attention to institutional development and training was also inadequate. B. Neglect of Scope for Gas Utilization 9. The projects' focus on the possibility of discovering oil reserves led to the neglect of programs related to the domestic use of gas reserves, even though it was apparent before the projects were appraised, that even if no oil were found there would still be substantial gas reserves that could be exploited. A gas market study was supposed to have been carried out under the First project, but was not carried out till mid-1982, a year after the Government of Tanzania (GOT) had signed a contract to develop a gas-based fertilizer plant. The wells drilled under the First project had confirmed that gas reserves exceeded 500 bcf but there was no oil. In view of this added information, the Second project should have included a component for the studies required to develop the field as a gas-only project, in the likely event that oil was not discovered in commercial quantities; it did not. This study was eventually incorporated in the follow-on Petroleum Sector TA project, effective in late 1985, three years after the end of the Songo-Songo drilling program. C. Gas Utilization and Power Planning 10. One of the most striking features of the Tanzanian energy sector today is the pervasive influence of both power shortages on the country's economy at a time when there exists a large reserve of unexploited natural gas that could be used for power production. The GOT, with the assistance of The Bank Group, has been working since 1992 on putting together an independent power production (IPP) project using Songo-Songo gas. The primary reason for the delay in exploiting these gas reserves for power production can be traced to the determination of the Government of Tanzania to use its gas resources for producing fertilizers.' Over the 1982-1988 period, GOT attempted unsuccessfully to raise funds for ajoint-venture US$450 million plant to manufacture fertilizer for export, but this proved impossible, given the macroeconomic situation and the associated country risks. 11. The shelving of the fertilizer project in the late 1980s, after international prices had fallen, should have made the gas available for other uses. This newly available resource was capable of fundamentally changing the country's options for power generation which were severely limited at that time. However, GOT continued to hold back over half the gas reserves for a possible future fertilizer project, thereby removing gas from the generation alternatives being studied by the power planners. The Bank did not question this fundamental policy position, and continued to support an exclusively hydro-power based power development program. 12. This lack of flexibility came at a time when Tanzania was facing a looming shortage of power generating capacity. Following the improvement in the macroeconomic situation from 1986 onwards, electricity demand began to rise rapidly. In 1989 it was clear that there was likely 1. As explained in TANESCO's comments on this PAR (see Annex), the Governments decision to use gas for fertilizer was based on its conviction that foreign exchange earning projects should take priority over other projects. 18 to be a serious shortage of generating capacity in the early 1990s, resulting from the delay in implementing preliminary studies for new hydropower generating facilities. 13. Nor were the decision makers in the Tanzania Electric Supply Company (TANESCO) and the Bank particularly interested in promoting the gas for power option in the 1980s. Even in 1991, all efforts were devoted to preparing the Kihansi hydro project, even in the face of system planning studies that underlined the urgent need for bringing new generation plants on stream in a time period that experience had shown was highly improbable with a large-scale hydro project. The key decision makers in Tanzania were apparently unconvinced of the risks of excessive reliance on hydropower and of the likelihood of an impending power shortage, so no significant efforts were made to seek funding for new rapidly implementable thermal generation facilities. 14. The 1984 Energy Assessment Report had demonstrated that a gas pipeline to Dar-Es- Salaam was justified even under very conservative demand forecasts for industrial and power gas use. Despite this report, little appears to have been done to seriously promote non-fertilizer uses of gas prior to 1988. 15. It was only as a result of the drought-induced power load shedding in 1992 that the Tanzanian authorities became convinced of the need to give high priority to the development of the Songo-Songo gas field for power generation. At least three years were thus lost from the time that the fertilizer scheme was shelved. Efforts have been underway since that time to find a way to utilize the gas for power generation with the assistance of the private sector. The Bank is currently working on putting together such a project, but a final agreement with a foreign investor has yet to be signed. Thus, with hindsight, it is apparent that if TANESCO and the Bank had begun in 1990-1991 preparing and promoting a gas-based generation plant rather than concentrating exclusively on preparation of the Kihansi Hydro project, it is likely that a gas pipeline would be close to completion today and that the country would have avoided the additional foreign exchange cost of burning imported diesel fuel for 2-3 years. Overall Assessment 16. Although the bank was extensively involved throughout the 1980s in lending operations over the entire spectrum of Tanzania's energy sector (oil and gas, coal, and hydro electric power generation), this involvement failed to lead to a national power sector development policy that took into account the risks and potential costs of excessive reliance on hydropower. The failure, throughout the 1980s, to transfer to Tanzania the lesson of the high risks of an exclusively hydro generation system, learned from the West African droughts of the early 1980s, has been costly; inadequate rainfall since 1992 has led to massive power shortages and load shedding over the past three years, which was only been mitigated at the end of 1995 with the installation of 75 MW of fuel oil-fired, gas turbine generators, financed by an emergency transfer of funds within the FY93 IDA credit for the Kihansi Hydro project. 17. Integrated energy sector analysis and planning of the kind proposed by the Bank/UNDP Energy Assessment Program did not take root with either the Borrower or in the Bank in the 1980s.2 Although the 1984 Energy Assessment Report provided a sector-wide view of problems and their solutions, individual sub-sectoral lobbies succeeded in promoting investments that were 2. In their comments on this PAR (see Annex ). TANESCO points out that it intends to carry out a full power sub-sector master plan to fully evaluate all available options to meet future power demand. 19 either uneconomic (coal) or premature (Kihansi Hydro), while gas remained unexploited. A wiser approach might have been to have had included a gas development planning agreement as an integral part of the original Songo-Songo exploration program. Such an approach might have avoided the ten-year hiatus between the completion of the drilling at Songo-Songo in 1982, and the initiation of the current effort to use the gas, which was initiated in 1992. 18. The exploration promotion campaign was useful in bringing Tanzania's petroleum potential to the attention of the international oil industry. It was a worthwhile use of the Bank Credit since the oil industry still appears to be willing to risk its own funds in exploring for oil in Tanzania. It is the only component of the Petroleum TA Project which has had sustainable benefits. Project Ratings 19. The Audit assesses the outcome of the two exploration projects as marginally satisfactory. They proved the availability of a substantial volume of domestic gas, but did nothing to ensure that this gas would be utilized. The sustainability of these projects is judged uncertain, and the institutional development is judged to be negligible. Bank performance and borrower performance are judged unsatisfactory. The Audit assesses the outcome of the Technical Assistance project as satisfactory, because it met two of its most important objectives (plugging the gas leaks in older wells and contributing to Tanzania's successful efforts to induce further oil exploration efforts by IOCs). Bank and Borrower performance for this project are judged satisfactory, but institutional development is judged to be negligible. Lessons 20. It makes little sense to spend limited investment resources on proving and developing domestic resources (natural gas or any other) without establishing at the same time an adequate institutional and legal framework and a plan for the utilization of these resources. If the public sector is to be involved in their development, then a utilization study may be needed to identify the economic benefits from alternative potential uses, including, primarily, power, domestic, and private industry. In the case of gas, such a study would focus on the netback from the most obvious uses, including power plants at various alternative locations. If the private sector is expected to use the gas (as it definitely should if the optimum use appears to be for industry), then the user with the highest offered netback payment should be allowed to proceed. 21. Energy sector planning should take into consideration all the natural resources at the disposal of the country. Adequate planning requires early preliminary studies of all alternatives from both the technical and institutional view, so that decisions are not limited by the "only available alternative within the time frame required" syndrome. Care should also be taken that technically feasible alternatives are not rejected because the Bank's task manager's and/or the national power entity's lack of familiarity and experience with these alternatives biases them against their use. 22. System balance is an important factor in system planning risk mitigation. In establishing a long term plan for power development, care needs to be taken to include an explicit risk mitigation factor for such naturally recurring problems as unusual adverse hydrological conditions (as draught) on a predominantly hydro-based power system. A series of individual discussions that favor development of the next hydro site may be individually rational but 20 collectively unsound if these decisions lead to a high probability of system-wide failure. The Bank's nine projects in the energy sector in Tanzania over the past 15 years were insufficiently harmonized. One way to impress this point onto unwilling national power planners is to provide greater cross-fertilization of experience of countries which ran into problems when they did not take these risk factors into consideration. 21 1. Introduction Macroeconomic Context 1.1 The three projects reviewed in this report were approved by the Board between 1980 and 1985, the final years of a two-decade period of socialist policies in Tanzania that ended in 1986. The early 1980s was the worst period for the Tanzanian economy since Independence. Real GDP declined, transport infrastructure and public services deteriorated, capacity utilization was very low, exports fell and the acute shortage of foreign exchange made even basic consumer goods scarce. Tanzania had no access to IMF resources during this period due to its unwillingness to devalue its currency. Nor was there a Bank-supported Structural Adjustment Program. 1.2 Since the launch of the Economic Recovery Program in 1986, Tanzania has pursued much more market-oriented policies and has dismantled many State controls over economic activity. Growth recovered while inflation fell. The exchange rate is freely floating and foreign exchange bureaus have been authorized since 1992. Real GDP has risen by about 4 percent per annum in recent years, a little above the rate of population growth. However, despite a decade of reforms, Tanzania is still handicapped by its pre-1985 legacy. There are still about 300 parastatal organizations, many involved in purely commercial activities, although a program of privatization is now gathering momentum. Foreign aid remains of critical importance to the economy as official exports cover only a third of imports. Donor funding is required for a very high proportion of the investment program, as well as to cover a significant share of recurrent expenditures. Since 1993, Tanzania's macroeconomic performance has started to deteriorate once again, as budget deficits and inflation have risen. Economic growth has also fallen, partly as a result of droughts and the ensuing shortage of electricity. Sectoral Context 1.3 Tanzania has no indigenous oil supply, and petroleum accounted for 30 percent of total imports in 1982, even though the consumption of petroleum products had stagnated from 1975 onwards at about 0.6 million tons/year due to a shortage of foreign exchange. There was a significant level of suppressed demand, controlled by rationing, despite the depressed level of industrial activity. Exploration activity for oil by foreign private firms continues on a modest scale (para. 3.8), but there have been no discoveries to date. 1.4 Gas was first discovered by an international oil company (IOC) in the Songo-Songo area, about 150 km south of Dar-Es-Salaam, in the mid-1970s. However, the IOC decided to relinquish its rights because it was not interested in investing in a gas project based on local consumption. In 1979, GOT requested Bank assistance to appraise and develop the field. Today, proven natural gas reserves amount to almost 750 bcf, but, they have yet to be developed commercially. Probable reserves are also estimated to exceed a further 700 bcf. The Bank Group has been working on establishing a private sector financed gas-fired power project since 1992. 22 1.5 Coal resources are estimated at about 2 billion tons, of which 300 million tons are proven. Domestic coal production from two underground mines takes place on a small scale, with total annual sales of under 50,000 tons. 1.6 Despite plentiful potentially developable hydroelectric resources, estimated at 3800 MW, less than 250 MW had been developed by the early 1980s. Annual per capita electricity consumption in 1982 was only 36 kWh, low even by standards of sub-Saharan Africa. Hydroelectricity accounted for 80 percent of supply. Electricity demand rose rapidly after macroeconomic conditions improved in the late 1980s. Per capita electricity consumption had risen to 53 kWh by 1993. Hydro generation capacity has since risen to 382 MW at the end of 1994 and accounted for over 90 percent of electricity produced. However, Tanzania has acted rapidly to ameliorate this situation. It installed 34 MW of diesel powered gas turbines in 1994, financed with bilateral funds, and in late 1995 installed another 75 MW of gas turbines, financed through a shift of funds from the transmission component of the FY93 Kihansi Hydro project (Credit 2489). 1.7 Because the Tanzanian power supply system was overwhelmingly hydro-based, it was vulnerable to the effects of drought. Poor rainfall since 1992, combined with demand growth and delays in bringing new generating plants into service resulted in load shedding with damaging consequences for the economy. 3. Only about 60 MW of thermal capacity in 1993, much of which was workable. 23 2. Project Objectives and Description Songo-Songo Petroleum Exploration Projects 2.1 The main objective of these two closely-interrelated projects was to assess the petroleum potential of the Songo-Songo field and thus help Tanzania develop a domestic supply of hydrocarbons. In effect, they were two phases of a single exploration program which were implemented back-to-back over the 1980-84 period. The two projects consisted essentially of the drilling of five wells in the Songo-Songo field to evaluate the size of the gas reserves and to confirm the presence of oil. They also contained small training and technical assistance components. At a time when oil prices were very high in real terms and Tanzania, as a non-oil producing developing country, was having to devote an inordinate share of its export earnings to cover its oil import requirements, little attention was given to the risk that only gas reserves might be proven. Petroleum Sector Technical Assistance Project 2.2 After the completion of the Songo-Songo drilling program, the Bank supported a third project in the sector with diverse objectives. According to the President's Report (PR), the project was expected to alleviate petroleum product distribution bottlenecks, assist GOT to develop a strategy for the utilization of its gas resources and to keep up the petroleum exploration effort by the private sector. However, the main rationale for the project was the urgent repairs needed to stem the gas leaks at two Songo-Songo gas wells drilled in the 1970s, prior to bank involvement. 2.3 The components of this US$11 million project were: (i) support for gas development through emergency well repairs, construction of protective structures for unprotected offshore wells, a gas utilization study for Songo-Songo gas and, if feasible, a pilot project for using compressed gas (CNG) in public transport; (ii) upgrading the petroleum product distribution system with tankers, pumps and LPG bottles, and a product distribution study; (iii) preparation of a petroleum exploration promotion undertaking; and (iv) institutional development. 2.4 The overriding motivation for this project was the urgent repairs needed to stem the gas leaks at the two Songo-Songo wells, even though these wells had not been financed by the Bank. The fact that they endangered life, property and the future development of the Songo-Songo gas field in which IDA had been heavily involved, and the inability of the Government to find alternative funding for the repairs provided the initial impetus to design a third project in the same sector in five years. 2.5 The scope of the project included a second, more detailed gas utilization study, even though a similar study had been carried out in 1982-1983 under the second exploration project, and the issue of gas utilization had also been examined in some detail by the 1984 Energy Assessment Report. The project also envisaged a feasibility study of a gas pipeline to Dar-Es- Salaam and a study of the institutional arrangements needed to produce, transport and distribute gas, arrangements which had not been agreed upon, despite the 1981 GOT joint venture agreement to set up a gas-based fertilizer plant.  25 3. Implementation Songo-Songo Exploration Projects 3.1 The total final cost of the projects is not known due to poor record-keeping for local expenditures, but was estimated at over US$75 million in the combined Project Completion 4 Report (PCR). The bulk of foreign financing came from IDA (US$49 million), while the EIB and OPetroleum EngineerC Fund together provided a further US$20 million in cofinancing. The projects failed to find oil, but by the time the second credit was fully disbursed in late 1984, gas reserves of over 700 bcf had been identified and proven. 3.2 Drilling Program: Implementation of the five-well drilling program is described in considerable detail in the PCR. Two wells were completed successfully and can serve as future production wells. The PCR describes in detail the technical and logistical difficulties encountered during drilling which contributed to the substantial cost-overruns under the first project. The PCR also clearly presents the problems arising from the lack of a properly functioning Project Implementation Unit (PIU). Two foreign consulting firms were employed for drilling management and interpretation of exploration results, but with overlapping authority. The deficiencies of this approach became deeply apparent under the first project, but were not corrected under the second. Instead, Bank staff became deeply involved in drilling decisions and project management. The lack of an effective PIU also resulted in an excessive degree of involvement by GOT staff (up to and including the Minister) in operational decisions that properly belonged to the Tanzania Petroleum Development Corporation (TPDC). 3.3 The project's management problems also had a major impact on the total cost, which turned out to be substantially higher than anticipated. What was envisaged at the time of appraisal of the first project as a four-well program costing US$45 million became a 5-well program with a final cost of nearly US$80 million at the end of the second project. Implementation delays, poor project management, technical problems and poor sitting of wells. all contributed to the higher costs. The extra funding required came from the second, follow-on project. 3.4 The Audit agrees with the PCR's assessment that the wells could have been drilled less expensively with better project management. Furthermore, some unsuccessful drilling could probably have been avoided by taking more care with the original geophysical interpretation. The PCR also correctly concluded that, given the prevalence of gas in the discovery basin, it would have been appropriate to investigate gas uses and economics in greater depth, before executing the drilling program. 3.5 Training: The training impact of the projects was minimal, possibly because training was limited to overseas courses of short duration. While these may have been useful, TPDC feels that its needs at that time were to send its staff for longer-term training. It financed these longer-term training needs out of its own budget. Since the time of these projects, TPDC has built up a skilled pool of exploration staff which handles exploration promotion efforts on an 4 Report No. 9611, dated May 1991 26 ongoing basis. However, this institutional building appears to have gone on outside of, and independently from, the projects under review. Petroleum Sector TA Project 3.6 The TA project was implemented over seven and a half years, from 1985-1993, three and a half years more than originally expected. The critical component, the emergency well repairs, was completed within 12 months of Board approval. The Bank financed the full foreign exchange cost of US$8 million out of a total project cost estimated at US$11 million. 3.7 Well Repairs. TPDC feels that the well repairs were executed satisfactorily, even though new leaks re-occurred in 1992.5 TPDC once again requested IDA to finance the repairs from the unused balance of the Credit. Although the Bank declined the request, TPDC was able to plug the leaks on a temporary basis. Workover repairs of the other (Bank-financed) wells drilled from 1980-1983 are also urgently needed now, and are scheduled to be carried out in the near future as part of the field development program but with a joint-venture partner.6 3.8 Exploration Promotion. In addition to the well repairs, the project contributed to the development of Tanzania's hydrocarbons sector in the area of petroleum exploration promotion. Geological and geophysical data was acquired and interpreted and the results were presented to the international oil industry in late 1989 to seek exploration proposals. As a result, three production sharing contracts (PSCs) were signed with two companies during the past five years. Unfortunately, these exploration ventures were unsuccessful; three dry wells were drilled and the acreage has now been relinquished. The PCR does not offer any explanation for the lack of greater interest on the part of IOCs. The Audit mission was able to ascertain that the reasons are outside Tanzania's control: stiff competition from other more promising oil provinces, particularly in the former Soviet Union, low world oil prices and risky geological conditions in Tanzania. Faced with this failing interest, the Government has attempted to enhance the contractual terms offered to IOCs to stimulate renewed interest on their part. A new PSC has recently been signed and another is at an advanced stage. GOT's positive action in this area is encouraging. 3.9 Rehabilitation of Product Supply Installations. The utilization of funds for product distribution equipment and spares was much lower than envisaged; only half of the US$1 million allocated to this category was used by the product marketing companies. The reason (which is not discussed in the PCR) was the increased availability of the needed equipment in the local market, as the scarcity of foreign exchanged eased. The component was designed at a time when economic conditions in Tanzania were at their worst, with acute shortages of all kinds of goods. By the time the credit was effective, foreign exchange was less scarce and the availability of spares, etc. in the local market improved steadily throughout the project's implementation period. It would appear that the oil marketing companies preferred to use their own local currency to directly buy the equipment and supplies they needed, rather than procure them via TPDC and the Bank credit. The Audit mission concluded that while the rationale for this component was sound at the time of appraisal, it was nevertheless unsuccessful because the improved macroeconomic climate rendered it largely superfluous. Despite the poor take-up rate by the oil marketing 5. The PCR for this project (Report No. 12638) makes no mention of these new leaks. 6. The onshore wells repair work has been completed and an offshore workover contract is being negotiated. 27 companies, a similar but larger component was included in the follow-on Petroleum Sector Rehabilitation Project (Credit 2202-TA), which was designed to assist greater private sector participation in product distribution. 3.10 The provision in the credit for the procurement of Liquid Petroleum Gas (LPG) bottles was also not drawn upon. Here the reason was that the Bank project did not ensure that LPG would be made available from the local refinery. Utilization of flared LPG depended on related investments at the Tiper oil refinery, which were not included in the project and which were not done.7 The failure of this component was a result of poor project design. 3.11 Studies. The studies supported by this TA project were, on the whole, highly successful. Various studies relating to the development of the Songo-Songo gas field were carried out at a late stage (1992) to take advantage of some unused funds. These studies were a major factor in the three-year delay in completing the project; their results were important for the sector because they helped to define the institutional and financial arrangements for the development of a gas- based power generation project with the private sector playing a predominant role. The studies concluded that a guarantee of foreign exchange convertibility of dividends by a multilateral lender like the Bank was a precondition for any private foreign investor to consider developing the field. The study on the cost of delivery of petroleum projects formed the basis for the subsequent private sector focused Petroleum Sector Rehabilitation project. 3.12 Training. The institutional development component of the project, which was designed to strengthen TPDC and the Ministry of Water, Energy and Minerals (MWEM), was modest in size (less than US$0.5 million). It was limited to short-term training courses, some computer equipment and software and the continuation of some ongoing technical assistance. Overall, this component lacked a clear plan and was implemented in an ad hoc fashion, with results of marginal value. As a result, the project provided little in the way of support for institutional strengthening. 3.13 Other Components. Two major components of the project, protective platforms for offshore gas wells ($1.9 million), and the CNG demonstration scheme ($2.5 million), together representing 40 percent of the original project cost were dropped entirely. The protective platforms were found to be unnecessary after temporary navigational aids were fitted. This component was an important risk minimization element of the project, but which could be dropped once an alternative solution was found to alleviate the risks. 3.14 The CNG scheme was dropped because it was premature to promote CNG as a transport fuel in the absence of a gas pipeline to Dar-Es-Salaam. Given the experimental nature of CNG as a transport fuel, where old vehicles would have to be converted at high cost, it is unlikely that this component would have been successful if it had been implemented. 7. The PCR for this project (Report No. 2638) does not comment on the success or failure of this component. 8 One of the studies also recommended establishing a separate gas transmission and distribution company to allow for greater flexibility, mixed ownership and a reduced regulatory burden.  29 4. Issues Project Management 4.1 Both of the Songo-Songo Exploration projects suffered from the same design weakness, namely, inadequate project management arrangements. The implementing agency, TPDC, was inexperienced and needed to be supported by a fully staffed PIU, with the necessary expatriate TA and with clearly defined responsibilities. This was not a feature of either project. 4.2 Bank staff were heavily involved in micro-management of the Songo-Songo drilling program. Yet they had neither the responsibility nor the accountability for the decisions they took. This close participation of Bank staff in project management was not unique to Tanzania. It may have been justified at the time, given the inexperience of the Borrower and the lack of a properly staffed PIU. Although only a second best solution, it was a practice of the Bank's petroleum staff at that time in projects with inexperienced state-owned oil companies. Choice of Lending Instrument 4.3 Stopping the gas leaking from the Songo-Songo wells was both important and urgent, but it was not necessary to create a US$8.1 million project to accomplish this goal. The Bank could have approved a supplemental credit of US$1.5 million, attached to the preceding Credit S1199-TA, which was still under implementation when the leaks were first reported in early 1993. In fact, Bank staff had initially hoped to be able to use savings from this credit to carry out the repairs, but the hoped for savings did not materialize due to the appreciation of the dollar against the SDR. 4.4 A supplemental Bank credit would have been a better choice of lending instrument. It would also have provided a faster response to the problem: the risk of a costly and dangerous explosion hung over the Songo-Songo area for nearly three years prior to the leaks being plugged in early 1 986. Project records show no trace of this option having been considered. Perhaps this was another case of lending program pressures. Neglect of Scope for Gas Utilization 4.5 The exploration project's focus on finding oil led to the downplaying of the planning process for oil gas utilization even though it was apparent even before the project was appraised, that substantial gas reserves existed even in the absence of oil. A gas market study was supposed to have been carried out under the First project, but was not carried out until mid-1982, a year after GOT had signed a contract to develop a gas-based fertilizer plant. The wells drilled under the First project had confirmed that gas reserves exceeded 500 bef, so the design of the Second project could have included a component for the studies required to develop the field as a gas- only project, in the event that oil was not discovered in commercial quantities. Such a component was eventually incorporated in the follow-on Petroleum Sector Technical Assistance project, but this latter credit did not become effective until late 1985, three years after the end of the Songo-Songo drilling program. 30 Gas Utilization and Power Planning 4.6 Twenty years after the Songo-Songo gas was identified, it still remains undeveloped. It is only since 1992 that the Bank has undertaken a serious effort to assist in its development for use in the power sector. During those twenty years, power planners seem to have been oblivious to the risks created by Tanzania's excessive reliance on hydropower as the basis for its electricity generation, risks that had been apparent for well over a decade, and had been demonstrated in similar national energy supply configurations in other African countries in the mid 1980s, when extended drought had led to costly power outages. This single-minded, hydro dominated approach to power development in Tanzania, coupled with the failure to attempt to resolve the problems surrounding the use of Songo-Songo, has led to a situation in Tanzania in which there are massive power shortages at the same time that there are unexploited natural gas reserves. Yet, as explained below, gas-based power generation made economic sense for the early 1980s, even when there were competing alternative uses for fertilizer production. The GOT and the Bank must share the blame equally for this failure to implement an optimizing resource utilization policy in the power sector. 4.7 The Bias for a Fertilizer Project. The major reason that Songo-Songo gas has remained unexploited for nearly twenty years is that the Government wanted to establish a domestic 9 fertilizer industry. This interest was realized when TPDC signed an agreement in mid-1981 with a foreign chemical company to set up a joint-venture US$450 million plant to manufacture fertilizer for export, even before completion of the Bank-financed reserve proving drilling program. The desirability of this contract, which was reviewed by the Bank only after it had been signed, was a hotly debated issue within the Bank. The Energy Department believed that GOT had to bear a disproportionate share of the project's costs and risks,10 and that the gas price received by the GOT was too low to justify these risks. On the other hand, the Industry Department believed that the gas price was reasonable and risks were acceptable because prospects for selling surplus fertilizer production in the Gulf and the Indian sub-continent were good and the price projections used to estimate cash flows were conservative. Nevertheless, the netback to gas used in fertilizer production was never more than US$2.00 per thousand standard cubic feet, substantially lower than the netback in power production. In the face of these differing and partly conflicting positions on the gas for fertilizer issue, the Country Department opted for neutrality. Country Department management never made any official comments or recommendations to GOT on either the terms of the fertilizer scheme or on alternative uses for the gas. 4.8 Over the period 1982-1988, the GOT worked to get the fertilizer scheme off the ground. During the early part of this period IFC also expressed some interest in, and support for, the project. These efforts were to no avail: given Tanzania's macroeconomic situation and the associated country risks, it proved impossible to mobilize the necessary funding from commercial sources. After the 1986 fall in oil prices, international fertilizer prices fell sharply, private investor interest also declined, as the netback to the gas fell to at, and then below the 9. In their comments on the PAR (see Annex), TANESCO points out that it intends to carry out a full power sub-sector master plan to fully evaluate all available options to meet future power demand. 10. TPDC was to finance the gas field and pipeline investment at an estimated cost of US$100 million, contribute 74 percent of the equity in the fertilizer plant and guarantee the entire debt financing (over US$400 million), while its private partner was contributing a mere USS24 million in equity. 31 agreed gas price. The foreign investor finally pulled out in 1988, effectively shelving the project. 4.9 This shelving of the fertilizer project in the late 1980s should have fundamentally changed fuel supply options for power generation. It did not. Under the fertilizer project agreement, 510 bcf of the 725 bcf of Songo-Songo's proven gas reserves had been reserved, leaving little for other purposes. However, rather than allowing the entire 725 bcf to be made available for alternative uses after the fertilizer project was shelved, the GOT decided to continue to hold back 395 bcf for a possible future fertilizer project. Whatever the rationale for this decision (which has proven impossible to ascertain), it had far-reaching consequences for decisions on power generation, since it artificially removed gas-fired generation from the alternatives being studied by the power planners. Yet the netback to gas used in power was in the range of US$4 per mscf, double that in fertilizer use, even when fertilizer prices were high." 4.10 The Bias for Hydropower. Tanzania repeatedly opted for developing hydropower rather than gas-based generation and the Bank's Africa Projects Department continued to support this choice throughout the 1980s and early 1990s. In 1981, IDA processed a loan to Tanzania for a hydro power plant.12 This hydro project was an appropriate choice. The hydro project was at an advanced stage of processing and the results from the drilling program were still incomplete. 4.11 However, the issue of the use of gas in power should have been reviewed again in 1985, when TANESCO's results from long-term power system plan models indicated that new generation capacity would need to be commissioned by 1992. There was some question as to whether a new hydropower project could come on stream in time since hydroelectric projects take a long time to build. Following the improvement in the macroeconomic situation from 1986 onwards, electricity demand, which had stagnated in the first half of the decade, began to rise rapidly, and Tanzania was beginning to face the prospect of a significant shortage of power generating capacity. 4.12 Nevertheless, very little preparatory work was carried out on new generation plants required to meet the rising demand by late 1988, when TANESCO met with its principal financiers at a Donors' Meeting to consider funding requirements for new investments in power supply. Hence, it was agreed that the 1985 power planning study would be rapidly updated under Bank financing. This study revised the load forecast upwards and demonstrated quite clearly that there was likely to be a serious shortage of generating capacity in the early 1990s, as a result of the delay in starting work on new hydropower generating plants. However, gas was again rejected as a potential alternative solution. Demand forecasts for non-power use of gas had been revised upwards, and the power planners assumed (or were instructed to assume), that all other potential users would take priority over power generation in the gas allocation process. The use of these planning parameters inevitably led the Power Sector Development plan to reject the possibility of using gas for power generation because "gas can no longer be considered to be available for power use in generation, except for the small quantities which might be consumed for bridging power and for standby and emergency use after the bridging period."'3 I1. See OED's study, A Review of World Bank Lending for Natural Gas, Report No. 10828, June 30, 1992. 12. The Fourth Power Project (Credit 1405-TA) to finance the 80 MW Mtera hydro plant, was appraised in late 1981. 13. Review of the 1985 Power Sector Development Plan, Acres International, July 1989. 32 4.13 In March 1990, the planning parameters for gas availability were altered again, when the feasibility study for the gas pipeline to Dar-Es-Salaam, completed in late 1989, came up with a lower industrial gas demand forecast, freeing up more gas for power generation. TANESCO's consultants were instructed to redo their 1989 calculations in time for the October 1990 IDA pre- appraisal mission for the Kibansi Hydro project. This report14 concluded that "use of gas in the power sector is strongly economic and can delay further hydropower projects by about 6 years." In fact, the cost of generating electricity from natural gas (using combined cycle technology, which produces an energy conversion efficiency of over 40 percent), was estimated to be about US$0.05 per kWh, when gas is priced at US$2 per mscf. And this cost was for firm supply. Nevertheless, the consultants recommended that the 180 MW Kihansi Hydro project proceed with priority, on the grounds that insufficient work had been done on developing a gas-based project, so that "...Before the (gas) resource can be developed, many important issues must be addressed and far-reaching decisions taken regarding policy matters, institutional arrangements and gas supply construction... Security of power supply is most important and Lower Kihansi is the only generating station that can reliably be commissioned by 1996/97." However, no recommendation was forthcoming on initiating this work to develop gas in parallel with the planning for Lower Kihansi. 4.14 This conclusion was not surprising, considering that TANESCO's top management at that time had an overt preference for hydro generation plant, since this was a technology that they understood how to operate after it was built, and were also skeptical about the ability of TPDC to produce and deliver gas.15 Nor was the recommendation challenged by the Bank, since the project staff dealing with the Tanzanian power sector were prepared to support further expansion of hydropower rather than turning to gas based generation. In 1991, all efforts were being devoted to preparing the Kihansi project, and an Engineering and Technical Assistance Credit was rapidly processed (Credit 2330-TA) to finance the detailed engineering of Kihansi. The continued importance of hydropower was borne out by the US$140 million IDA credit provided in March 1993 (Credit 2489-TA) for building the first phase of the Kihansi Hydro project.16 4.15 The arguments used to reject the gas-based power option were based on the lack of government action on complex institutional arrangements and gas pipeline construction issues that had not been studied. But they had not been studied because during the 1980s neither the Bank not the GOT were interested in developing a program of gas utilization for power. These arguments were then used to justify the Bank's decision to support and assist in the financing of the construction of the Kihansi hydropower plant. The process culminated in the Power VI Project, for which the IEPS was approved by Management in 1990 and a US$200 million Bank 14. Use of Gas in the Power Sector, Acres International, October 1990. 15. As explained in its comments on this PAR (see Annex) insistence on implementing Kihansi was based on decision to proceed with a project that had firm financing the place. The prospect of abandoning the project in favor of a new project which had not yet gone through the long decision-making process with various authorities including financiers before it could reach the same firm financing stage as Kisansi was simply not palatable for a gas fired power plan confirms that judgment. 16. It must, however, be emphasized again that the economic comparison of gas and hydro at this point in time favored the use of gas for power generation. At Mtera, the cost of generation, based on average year output was US$0.053 per kWh; at Kihansi it was around US$0.07. These costs must be compared with the cost of firm power (available also during drought conditions, when hydro- power is not available) of about US$0.05 per kWh from gas generation when gas is priced at about US$2.00 per mscf (based on combined cycle technology, which has an energy conversion efficiency of over 40 percent). For details see OED's Report Lending for Electric Power in Sub-Sahara Africa, Report No. 14961, October 16, 1995. 33 Credit was approved by the Board in April 1993. However, despite the confidence expressed in the SAR that risks of delay relating to a hydro project were less than with gas, the construction of Kihansi is proceeding slower than expected and 2000 is now the earliest expected commissioning date. Work on developing a project to use Tanzania's unutilized gas reserves for power generation began in 1992. 4.16 The Costs of these Biases. The Audit finds this approach hard to justify in retrospect. The earlier system planning studies had underlined the urgency of bringing new generation plant on stream, which could not be accomplished by another large hydro project, but because the key decision-makers in Tanzania were unconvinced of the merits of gas-based generation and of the risks of impending power shortages, little was done to seek funding for new thermal plant to provide the "bridging" power needs cited earlier. This Rroblem has been noted before: the PAR of the Fourth Power and Power Rehabilitation Projects (under implementation in the late 1980s) observed that the rehabilitation of existing old thermal plants that was envisaged under these projects was repeatedly postponed as they were felt to be of "low priority." 4.17 Both TANESCO's power planning consultants19 and some Bank staff were, however, aware of the risks and potential impact of a prolonged drought. The Bank had already witnessed the damaging effects of the 1983-1984 drought on Ghana and Cote d'1voire, both of which had all-hydro power systems at the time. The advantages of "thermal complementation" to such hydro systems had been demonstrated in reservoir simulation models developed by the same consultants, but TANESCO was either unfamiliar with or unconvinced by this analytical work. 4.18 In the view of this Audit, the need to respond urgently to the rapidly increasing power demand and the lack of prior preparatory work on Kihansi should also have convinced the Bank of the importance of pursuing more rapid solutions to the looming generation shortage, such as installation of gas turbines. Unfortunately, it was only as a result of the drought-induced load shedding in 1992 that the Tanzanian authorities became convinced of the need to add more thermal and gas turbine capacity, and to give full priority to the development of the Songo- Songo gas field. At least three years were thus lost from the time that the fertilizer scheme was shelved. 17. Annex 4.1 of the SAR for the Power VI Project states: "Reflecting Tanzania's greater experience with hydro development and possible delays in gas development, the Government has decided to commission the Lower Kihansi project first in 1998 and to introduce gas generation to the power system thereafter as soon as it would be available in Dar-Es-Salaam. IDA has endorsed this decision." 18. ReportNo. 14503. 19. The same power planning consultants were advising VRA, the Ghanaian generation company throughout the 1980s. 20. As early as 1984, the Energy Assessment Report (para. 4.39) stated that "a substantial thermal capacity would allow the application of more liberal reservoir operating rules and thereby increase average hydroelectric energy production." 34 4.19 The delay in developing gas-based generation has proven to be extremely costly to the economy, which has suffered from load shedding since 1992. According to recent IDA financed studies, in 1994 alone, the failure to meet one third of the existing demand due to an extended period of low rainfall and the resultant loss in production capacity, are estimated to have cost Tanzania US$170 million, or about eight percent of its GNP. 35 5. Conclusions Project Outcome and Impact 5.1 The exploration promotion campaign was useful in bringing Tanzania's petroleum potential to the attention of the international oil industry. Even though oil prices have fallen precipitously since that time, the effort was clearly a worthwhile use of Bank resources, since it induced the IOCs to make a large investment in further exploration, and the oil industry still appears to be interested in exploring for oil in Tanzania. The petroleum products distribution study also was useful in that it prepared the grounds for the follow-on rehabilitation project. 5.2 Even though the Bank has been extensively involved across the whole spectrum of the energy sector in Tanzania during the past fifteen years through nine IDA credits,21 Tanzania has been unable to avoid the planning failures that have led to power shortages and load shedding over the past three years (1993-1995). Nor has Tanzania been able to benefit from the gas resources discovered in the beginning of the 1980s. The Songo-Songo wells, drilled 15 years ago, are deteriorating, and gas is unlikely to reach Dar-Es-Salaam for at least another 3 years. The Bank must share part of the responsibility for this inadequate energy sector resource utilization strategy. Bank and Borrower Performance 5.3 The Bank had serious (albeit officially unexpressed) reservations about the fertilizer project, given its size, the additional debt burden and the disproportionate share of risks to be borne by GOT, as well as its supposed priority as the highest value, and therefore optimum, user of gas. The Bank could have taken a more overtly critical position on the project in its dialogue with GOT in the early and mid-1980s, but instead decided to take an informal approach to encouraging the GOT to make available some of the gas for power and other uses. This soft approach failed to bear fruit. Thus, even though the 1984 Energy Assessment Report had demonstrated that a pipeline to Dar-Es-Salaam was justified even under very conservative industrial and power gas demand forecasts, and made a strong recommendation to begin the feasibility study as soon as possible, little was done to seriously promote non-fertilizer uses of gas prior to 1988. 5.4 Although the Bank lent Tanzania US$50 million to help explore and prove up the gas reserves it waited almost a decade before it undertook to promote a commercially viable gas utilization program. The Bank usually insists that the commercially viable projects it finances in the energy sector generate revenues to reimburse the project debt incurred by the Borrower. Yet Bank staff only began to consider the use of gas for power from 1989 onwards, once it was clear to all concerned that the fertilizer project was out of the question. Serious efforts to develop the Songo-Songo gas field for power began much later, in 1992, and a final agreement with a foreign investor has yet to be completed. 21. Four each in petroleum and power plus one for coal. 36 5.5 The Bank, as the main sponsor of the Songo-Songo gas field development program, could have done more to ensure that the gas reserves were used, by ensuring that the necessary gas market and institutional studies were carried out as part of the initial exploration program. Instead, institutional and regulatory issues were only examined in 1992. In the Bank, the administrative separation of gas sector staff from those in charge of power projects may have played a role in this lapse. Sub-sectoral lobbies within the Bank succeeded in promoting investments that were either uneconomic (coal exploration) or premature (Kihansi Hydro), while gas remained unexploited. None appear to have taken a global view of the sector and its macroeconomic links. Even within a single energy subsector, electricity, there was a disturbing lack of recognition of the risks of over reliance on hydro, and, as a result, Tanzania failed to learn from the West African droughts of the early 1980s. 5.6 Borrower performance was not significantly better than the Bank's. The GOT decided that the best use of gas was for the production of fertilizers well before any serious analysis of the options had been made. It rushed into the fertilizer option and stuck with this position long after the worldwide economic environment for fertilizer production had changed and the original rationale and justification had disappeared. TANESCO also maintained its single-minded, hydro-oriented power generation focus throughout this time period, in contradiction to the conclusions of the studies it had commissioned. Clearly, in Tanzania, as in the Bank itself, strong subsectoral lobbies were able to promote uneconomic investments (fertilizer) and premature investments (Kihansi Hydro), at the expense of other alternatives, even if gas remained unexploited. 5.7 If TANESCO and the Bank had devoted some efforts to preparing and promoting a gas- based generation plant, it is likely that a gas pipeline would be close to completion today and that the country would have avoided the additional foreign exchange cost of burning imported diesel fuel for 2-3 years. A gas production and pipeline project in the public sector could probably have been executed even earlier if GOT and the Bank had tried to do so in the mid 1980s. Such projects in other countries have received Bank support, notably in Thailand and in Bangladesh Project Ratings 5.8 The Audit assesses the outcome of the two exploration projects as marginally satisfactory. They proved the availability of a substantial volume of domestic gas, but did nothing to ensure that this gas would be utilized. The sustainability of these projects is judged uncertain, since the recently developed possibility for establishing a private sector project to use the gas in power generation still has many obstacles to overcome before it can become a reality. The project's role in institutional development is rated as negligible. Bank performance and borrower performance are rated as unsatisfactory. Well-reasoned and supported recommendations of the Bank's Energy Assessment Report were ignored by both. 5.9 The Audit assesses the outcome of the Technical Assistance project as satisfactory and its sustainability as likely, in that it met two of its most important objectives: plugging the gas leaks in older wells and contributing to Tanzania's successful efforts to induce further oil exploration efforts by IOCs. It also layed the foundation for a subsequent IDA project to improve the efficiency of petroleum product distribution through private sector participation. Bank and Borrower performance for this project are rated as satisfactory. But the project's institutional development achievement is rated as negligible. 37 6. Lessons 6.1 It makes little sense to spend limited investment resources on proving and developing domestic resources (natural gas or any other) without establishing at the same time an adequate institutional and legal framework and a plan for the utilization of these resources. If the public sector is to be involved in their development, then a utilization study may be needed to identify the economic benefits from alternative potential uses, including, primarily, power, domestic, and private industry. In the case of gas such a study would focus on the netback from the most obvious uses, including power plants at various alternative locations. If the private sector is expected to use the gas (as it definitely should if the optimum use appears to be for industry), then the user with the highest offered netback payment should be allowed to proceed. 6.2 Energy sector planning should take into consideration all the natural resources at the disposal of the country. Adequate planning requires early preliminary studies of all alternatives from both the technical and institutional view, so that decisions are not limited by the "only available alternative within the time frame required" syndrome. Care should also be taken that technically feasible alternatives are not rejected because the Bank's task manager 's and/or the national power entity's lack of familiarity and experience with these alternatives biases them against their use. 6.3 System balance is an important factor in system planning risk mitigation. In establishing a long term plan for power development, care needs to be taken to include an explicit risk mitigation factor for such naturally recurring problems as unusual adverse hydrological conditions (as draught) on a predominantly hydro-based power system. A series of individual discussions that favor development of the next hydro site may be individually rational but collectively unsound if these decisions lead to a high probability of system-wide failure. The Bank's nine projects in the energy sector in Tanzania over the past 15 years were insufficiently harmonized. One way to impress this point onto unwilling national power planners is to provide greater cross-fertilization of experience of countries which ran into problems when they did not take these risk factors into consideration.  39 ATTACHMENT COMMENTS FROM THE BORROWER SONGO-SONGO PETROLEUM EXPLORATION PROJECT (CREDIT S-027-TA) 2ND SONGO-SONGO PETROLEUM EXPLORATION PROJECT (CREDIT 1199-TA) PETROLEUM SECTOR TECHNICAL ASSISTANCE PROJECT (CREDIT 1604-TA) DRAFT PERFORMANCE AUDIT REPORT WB MARCH 21, 1996 Comments: The report is a good account of and informative on the background to Songo-Songo Natural Gas to Power Generation Development Project currently being negotiated between Tanzanian parties and foreign private investors. The report traces the original objectives of World Bank involvement in the project and how these objectives appear to have been influenced by various players in the decision-making process over the years. It concludes that both the World Bank and Tanzanian authorities failed to implement recommendations made in various reports by experts for optimal use of Songo-Songo natural gas reserve. The reports referred to in the draft recommended use of natural gas for power generation as the least-cost alternative after Mtora Hydropower Project implementation. Instead the Government position was, until recently, to use Songo-Songo natural gas in industry and, in particular, for fertilizer production. The following are general observations and remarks on the draft: The Government's decision to take that position was based on its conviction that foreign exchange earning projects should take priority over other projects. It was hoped fertilizer would be exported to earn the much needed foreign exchange, albeit, at marginally lower economic returns. This being the case, the power sub-sector had to take the least cost alternative among various sources of power generation, including the much reduced amount of natural gas at Songo-Songo available for power generation. The failure of the fertilizer project to take-off should not, therefore, be construed as inability of the Tanzanian parties to utilize gas in an optimal and timely way. This is typical of Projects requiring foreign financing with typically complex decision-making processes and divergent perceptions of various issues concerning such projects. The report is critical on the way the Tanzanian parties and the World Bank have handled integrated energy planning for the country resulting into promoting investments which were either uneconomic (coal) or premature (Kihansi). It is possible the report refers to promoting studies for coal use. Otherwise we are not aware of any substantial investments made in coal utilization in the country. As far as studies are concerned it is of great advantage to the energy sector in general and the power sub-sector in particular to have as many studies as possible to enable decision-makers to make rational choices of investments in the sector. The lack of timely and detailed feasibility studies will always create a possibility to implement sub-optimal projects. That is why TANESCO is keen to carry out a full power sub-sector master plan to review major economic changes which have taken place since the last master plan and to fully 40 ATTACHMENT evaluate any options available to meet future power demand in the country. The aim of the study is to identify the least-cost power expansion plan for the country for the next twenty years or so. The issue of Kihansi Hydro Project is not put in its full perspective. The insistence of implementing this project even after Songo-Songo natural gas is available for power is not so much a question of being single-minded on the part of TANESCO but rather a question of sticking to a decision to proceed with a project with firm financing in place given the fact that valuable time has slipped in the past to bring status of the project to the present stage. The prospect of abandoning the project in favor of a new project which unfortunately has to go through the long decision-making process with various authorities including financiers was simply not palatable to TANESCO. Indeed, notwithstanding its economic competitiveness for power generation, Songo-Songo Project is now in its third year of negotiations with various parties. The prospect of reaching an agreement in the near future is not certain. As a result TANESCO has to run the recently installed gas turbines on expensive imported fuel to avoid the unpopular economy-wrecking power rationing in the country which has been in force for the past four years. CORPORATE PLANNING DEPARTMENT TANZANIA ELECTRIC SUPPLY COMPANY LIMITED APRIL 1996.    IMAGING Report No: 15802 Type: PPAR