Sierra Leone - Issues and options in the energy sector

- =:: Report No. 6597-SL Sierra Leone: Issues and Options in the Energy Sector 0(ctotb'r 19F8 bft io do pint NPIbtd Bank Energy Sector Assessment Program This d euntent hs.a amEnstricted distribution. Its contents may not be disclosed without at-ithorization from the Government the UNDP or the Wbrld Banik.  JOINT UNDP/WORL) BANK ENERGY SECTOR ASSESSMENT PROCRA!M Reports Already Issued Country Date Number Indonesia November 1981 3543-IND Mauritius December 1981 3510-MAS Kenya May 1982 3800-KE Sri Lanka May 1982 3792-CE Zim.babwe June 1982 3765-ZIlt Haiti June 1982 3672-HA Papua New Guinea June 1982 3882-PNG Burundi June 1982 3778-BU Rwanda June 1982 3779-RW Malawi August 1982 3903-MAL Bangladesh October 1982 3873-BD Zambia January 1983 41:0-ZA Turkey March 1983 3877-TU Bolivia April 1983 4213-BO Fiji June 1983 4462-FIJ Solomon Islands J.,re 1983 4404-SOL Sene,.al Jtjl Y 1983 4182-SE Sudan Jliy 1983 4511-SU Uganda JuI y 1983 4453-UC Nigeria August 1983 4440-UNI Nep,a AugL3t, 1983 4474- NEP The Gamibia N)vemrber 1983 4143-GM Peru Janu ary 1984 4677- EP Costa Rica January 1984 465,-CR L.esot.hn .Jan'.ary 1984 4676-LSO Seychelles nan ar-y 1984 4693-SEY Morocco March 1984 4157-.M0R Portug, al. Apri 1 1984 4824-PO Niger May 1984 4642-NIR Ethiopia Julv 1984 4741-ET Cape Verde August 1984 5073-CV Guinea 8issau August 1984 5083-GUB Botswana Spptember 1984 4998-BT St. Vincent and the Grenadines September 1984 5103-STV St. Lucia September 1984 511l-SLU Paraguay October 1984 5145-PA Tanlzania Novembier 1984 4969-TA Yemen Arabh kepubi)C icecemn-,er 1984 4892-YAR Liberia December 1984 5279-LBR Islamic R ep Li b t Maoritania tA)pril, 1985 5224-MAU Jam?iri C r3 1 8 5466-EM r ')rv Co ....r a ! 3O- I C 5 sen iI n . 22- 2 F N UIit ~ ~ ~ ~ ~ ~ : -.7!'" ;,';i'' ' ,1 ( 2,P  FOR OFFICIAL USE ONLY REPORT NO. 6597-SL SIERRA LEONE ISSUES AND OPTIONS IN THE ENERGY SECTOR OCTOBER 1987 This is one of a series of reports of the joint UNDP/World Bank Energy Sector Assessment Program. Financing for this work has been provided, in part, by the UNDP Energy Account, and the work has been carried out by the World Bank. This report has a restricted distribution. Its contents may not be disclosed without authorization of the Gcvernment, UNDP and the World Bank.  ABSTRACT 41tLough Sierra Leone is favorablv endowed with energ, resources, particularly forestry and hydro-power, over the past few years iL has experienced severe shortages of energy supplies which have disrupted economic activity. The major sources of difficulty are: (a) a lack of foreign exchange to pay for oil imports and maintenance materials for plant and equipment; (b) an almost total dependence of the modern sector of the economy on imported energy products; and (c) weaknesses in enerSy sector institutions and lack of technical expertise for the planning and management of energy resources. Despite these difficulties, there are grounds for optimism. The Government has initiated a program of economic reform which, if sustained, would relieve the foreign exchange shortage. At the same time, oil prices have declined in the past year, providing an opportunity for the Covernment to develop a coherent policy for the energy sector to avoid future difficulties. The assessment report focuses on the major short-term and long- term issues facing the sector. In the short-term there is a need to: (a) improve the supply o electricity and petroleum products; (b) strengthen demand management to encourage the effisient use and alLoca,ion of energy resources by reLating energy prices to the costs of supply; and (c) initiate institutional reform to improve sector coordina- tion, enhance the autonomy of the parastatal energy agencies, and increase the incentives on the part of parastatals to supply energy at least cost. Over the longer term, perhaps the most pressing issue is the extent to which it would be economic to develop indigenous energy resources and production facilities to substitute for high cost impotis. Within this context, the report considers options for meeting future power requirements, the dispcsition of the refinery, and a strategy for increasing fuelwood supplies. A priority energy investment program is outlined in the report, together with a supporting plan for technical assistance and diagnostic studies.  ACRONYMS BP British Petroleum DTI Department of Trade and Industry FIC Forest Induetries Corporation IMF International 4onetary Fund NPA National Power Authority NP National Petroleum MDEP Ministry of Development and Economic Planning MANRP " nistry of Agriculture, Natural Resources, _nd Forestry MEP Ministry of Energy and Power MIS Management Information System MOF Ministry of Finance Planning MOm Ministry of Mines MOT Ministry of Transport and Communications MTI Ministry of Trade and Industry NGO Non-governmental Organization SIERMOCO Sierra Leone Bauxite Mining Compar.y SLPRC Sierra Leone Petroleum Refining Company ABBREVIATIONS bpd barrel per day DWT deadweight ton Ci gigajoule GSP Government selling price GWh gigawatt hour ha hectare HV high voltage kcal kilocalorie kg kilogram km kilometer kV kilovolt kWh kilowatt hour 1 liter LRMC long run marginal cost LX low voltage m cubic meter Mi megajoule MV medium voltage MT metric ton SRMC short run marginal cost TOE tins of oil equivalent  CURRENCY EQUIVALENTS US$1.00 = 5.10 Leones (Le) This was the exchange rate at the time of the mission (February 1986) and is the rate used in the repcrt unless otherwise stated. The Leone has since been floated. ENERGY CONVERSION FACTORS Energy Form Calorific Value TOE/ton (mil lion kilocalories/ton) Biomass Firewood (700 kg/m3) 3.5 0.343 Charcoal 6.9 0.676 Agricultural Residues 3.0 0.294 Petroleum Products LPG 10.8 1.059 Gasoline 10.5 1.029 Jet Fuel 10.4 1.020 Kerosene 10.3 1.010 Gas Oil/Disel 10.2 1.00 Fuel Oil 9.5 0.941 Crude Oil 10.2 1.00 Electricity Calorific Value: 860 kcal/kWh (0.084 TOE/MWh) Other I metric ton of crude oil = 7.505 barrels of Nigerian Bonny Light 1 Imperial gallon = 1.2 U.S. gallons This report is based on the findings of an Energy Assessment mission which visited Sierra Leone in February 1986. The mission team consisted of: Messrs. A. Ferroukhi (Mission Leader): I. Aleem (Deputy Mission Leader); J. Boroumand (Researcher); P. Beard (Power Specialist); W. Matthews (Petroleum Refining and Distribution Specialist); R. de Lucia (Household Energy Specialist), and J. Rochet (Geologist). Mr. Aleem is the principal author of the report. Secretarial support was provided by Ms. Morrissa Young and Mr. Basha,at Ahmad.  TABLE OF COVITS SUMMARY AND RECOMMENDATIONS .............................. . i I. ENERGY A!JD THE ECONOMY ............................... * . 1 The Economy ..... .......9.....**. 1 External Trade Gap ........ ............................ 1 Recent Developments .....* *...*..................... 2 Economic Scenarios for Sierra Leone..................... 3 Energy-Economy Interactions............................... 4 Key Energy Issues ....*............................. ... 6 Energy Consumption Patterns ...... ............. ........ 7 Consumption Patterns Under Alternative Scenarios.......... 9 Energy Resource Base...... .....9999 *99999999999999999999999 11 Institutional Framework... ...........9 . ...... 9 99 12 II. PETROLEUM .................... o.......... . . . 13 Introduction ...................... , 13 Recent Developments and Subsector Organization.* ......... 13 Short-Term Issues ....................... 16 Procurement ........ .................................. 16 Demand Management and Pricing ........................... 20 The Role of the Government in Managing Shortages and Crises.. ......*s**............. 26 Long-Term Issues ..... .........00......**9*..... 28 Future Demand for Petroleum Products.................... 28 The Viability of the Refinery .......................... 30 Recommendations .................................. 33 Oil Exploration .................... 33 Subsector Organization.. ... .. 99 36 Investment and Training Requirements...................... 37 III. ELECTRICITY ................. so,.... 40 Introductio. .................... 40 Institutional Reform: Raising the Effectiveness of NPA Management ................................a....... 42 Institutional Framework .............999......99.9.99..... S., 42 Managerial and Technical Problems Within NPA............. 44 aecommendations .... *999999999999999999999999 999999999999 45 Physical Rehabilitation of the Public Power Supply ........ 47 Concerns About Existing Rehabilitation Plans ............ 51 Tariff Policy, Demand Management, and Financial Rehabilitation..* ......................... 52 Demand and Supply Side implications of Tariff Policy ...... 53 Supply Costs and Tariffs ................................ 57 Recommendations ......O-..... . .. .... 59  Lcns-Term Issues ..................... . . . ... ..... ............ *0 *e@. 60 Bumbuna vs. Other Development Options..................... 61 Development Options for Provincial Load Centers........... 71 Reconmendationso.... ... .. ........ *******. .......* *..*. 73 Investmento ..*.ooo.o.o.ooo..o..oe*ooooo.o....................... 75 Trends 1981-1986 ..' o o.$..0 .06 0 0 6 4 ..............0.................. 75 Future Requirements. . .................. ... . . . . .. 77 IV. WOODFUELS/HOUSEHOLD ENERGY.......................... 78 Reasons for Concern..... ... .o.9. .... o..o.o. .o** **ooooo * 78 Characteristics of Household Energy Use...... ........0... 79 Woodfuel Supply and Demand .....o.... ...... ......... .... 82 Household Energy/Woodfuel Strategy ............... ................. 85 Supply Side Measures... ....................o..,.... 90 Main Supply Side Recommendations........................ 9) Changes in the Inst;tutional Frameworko ............o 94 Recommendations. . ....... ... .**oGo...o........... 96 Evaluation of the Potential of Other Energy Sources ..... 96 Recommendationso.......................................... 100 Investment and Technical Assistance Program...o.oa...to... 100 V. ENERGY SECI'OR MANAGEMEN.. .... o .... ... o..... o. .o..0.0.. 102 Introduction ......... .. .............. 102 Energy Sector Coordination and Planning .....,.,,........ 102 Problems with Existing Institutional Frameworko...... too 102 Reco m mendations... o ............................ 00 103 Investment Planning/Aid Coordination...............*... 104 Availability of Funds. .. .104 Investment Requirementsq.........m.. . . ........ .. . . . . 105 Aid Coordination and Monitoring-o............... ........ 106 TABLES 1.1 Scenarios (Summary) ....................... ..... 4 1.2 Petroleum Imports and External Trade . .............. 5 1.3 Structure of Final Energy Consumption (1984) .............0 8 1.4 Final Energy Consumption, 1995, Accelerated Growth Scenario..........o.......to.........o*. .......... 9 1.5 Final Energy Consumption, 1995, Base Case Scenario ..... o...... 10 2.1 P2troleum Subsector--Basic Data ........ o ...... ....... 14 2.2 Sierra Leone Crude Oil Imports ............... ...o.......... 17 2.3 Comparison of Official Retail Prices in 1985................. 20 2.4 Sierra Leone Official Petroleum Product Prices and Economic Cost of Supply u.......... - ..... . .. 21 2.5 Comparison of Official 1985 Ex-Refinery Prices with Hypothetical Costs of Direct Imports .................... 22 2.6 Revise,i Petroleum Product Price Structure................... 24 2.7 Forecast Summary of Inland and Bunker Der,|and: Accelerated Growth Scenario with Reduced Pumbuna .... .. ..... ...... 29 2.8 Forecast Summary of Inland and Bunker Demand: Base Case Scenario with Reduced Bumbuna .................. 29  2.9 Refinery Ownership. ......................... ........... .. .0 30 2.10 Refirery Investments... .. .. . .. .. . .. . . . ... .*. .... 37 3.1 Power 3ubsector--Basic Data ... 41 3.2 Power Plants on Western Area G r i d 48 3.3 Casoi Ordered and Supplied (1985) 50 3.4 NPA Generation and Sales ................ 53 3.5 NPA long Run Supply Costs ... 57 3.6 Demand Forecast ... 61 3.7 Bumbuna Hydro Project ... 64 3.8 Comparison of Bumbuna Options ... 67 3.9 Bumbuna Surplus Energy in Wet Season (AC Scenario) ... ...... 71 3.10 Details of Sniall Hydro Schemes....... 72 3.11 Estimated Investment in NPA Assets FY1981-86 75 3.12 Estimates of NPA Investment Requirements FY1987-91: Base Case Scenario................ ..... I 76 3.13 Estimates of NPA Investment Requirements FY1987-93: Accelerated Growth Scenario ......... ..... ... . 76 4.1 Household Energy Expenditures as a Share of Monthly Income... 78 4.2 Estimated Household Energy Consumption, 1 9 8 .80 4.3 Comparative Costs of Cooking Fuel.s in Freetown, February 1986 ........... ................................. 81 4.4 Forest Resources and Annual Wood Incre ms . . .83 4.5 Woodfuel Demand (1986) ....................................... 84 4.6 ComparaLive Cooking Cost of ALternative Stoves. 87 4.7 Investment Program (1987-91): Household and Renewables Sectorse.. .. ........... 101 5.1 Energy Sector Public Investment Requirements, 1987-91 105 ANNEXES 1 Scenario Details.. ....... 108 2 Energy Balances ............................................... 112 3 Retail Petroleum Product Price3 .. 117 4 Crude Oil arid Product Price Forecast ..... 118 5 SLPRC Refinery Economics ........... ..O. 120 6 Petroleum Concessions . .... 121 7 Autogeneration of Electricity in the Area of Western Power Grid ............... 122 8 Regression Analysis . ...... ...126 9 Relationship of Load Forecast, GDP and Electricity Generation ..... 129 10 H:storical Pattern of Tariffs .............................. 131 11 Generation Cost Estimates. .................. .. ......... 132 12 Future Generation and Consumption of Electricity, by Scenari G . 139 13 Alternative Expansion Pans ...142 14 Bumbuna Capital Cost Estimates . . 149 15 Future Coal and Crude Oil Prices, by Scenario . ............... 158 16 Effect of Bumbuna Scheme on Foreign Exchange Balance.. 159  17 Future NPA Supply Costs andTariffs.......................... 161 18 Household Energy Survey Results .............................. 167 19 Retail Price Movements in Household Fuels.................... 169 20 Current Cost/Price Build-up for Fuelwood Retailed in Two Different Areas of Freetown, February 19868 6............. 170 21 Energy Putential from Availabl.e Agricultural Residues in Sierra Leone....... ................ *****....********** 171 22 Lignite Deposits it. Sierra Leone--Yema Araa 172 NAPS IBRD 20078 Sierra Leone IBRD 20079 Power System and Potential Hydroelectric Sites IBRD 20080 Natural Forest Regions  SUMMARY AND RECOMMATIONS Overview of Energy and the Economy 1. Energy consumption in Sierra Leone is dominated by Luelwood which accounts for about 80% of the energy used. The remaining 20% is supplied by imported oil. Fuelwood is the traditional form oi energy and is used almost exclusively by households for cooking and for traditional craft activities. Petroleum, on the other hand, is the most inportant source of energy for the modern sector of the economy. In recent years foreign exchange difficulties have restrained petroleum imports and re- duced the share of oil consumption in the overall energy balance. Future growth in consumption is expected to show a return to the longer term historical trend of a shift towards modein fuels, primarily oil but also hydroelectric power, a resource which is still virtually untapped. 2. Gross energy consumption (including conversion and distribution losses) is estimated to have grown between 1975 and 1984 at an annual rate of 1.7% to some 1.2 million tons of oil equivalent (toe). Over the same period GDP growth has been sluggish, averaging about 2%, so that the ratio of the growth in energy consumption to that in GDP has been less than one, which is below the average fo;- most developing countries. Petroleum consumption peaked in 1979 at around 178,000 tons and has been on a declining trend since then reaching 167,000 tons in 1984. The 1984 gross energy consumption of 1.2 million toe represents a per capita consumption of 310 kilograms of oil equivalent (kgoe); per capita con- sumption of commercial energy was about 60 kgoe, sqich is about average for countries with similar income per capita (US$330 in 1983). House- holds and transportation were the two largest users of energy in 1984, representing 86% and 6% of final dem-nd. Industry ranked third with 5% of demand. 3. Sierra Leone, with a population of 3.7 million people, is reasonably well endowed with energy resources, particularly forestry and hydroelectric power. The annual supply of woodfuels from the country's forests is estimated at around 4 million m3, about 10% of which is consumed in Freetown. Annull sustainable yields of wood are estimated at between 9 and 15 million m but a large part of this supply is not eco- nomically accessible and there are growing signs of regional shortages near urban areas. There is an extensive network of rivers and tribu- taries which provide a las-ge hydroelectric power potential conservatively estimated at 1,200 MW. Technically, and economically, the most promising site is at Bumbuna, on the Seli River, with an ultimate potential of 305 MW of installed capacity. Around 20 mini-hydro sites have also been identified but their feasibility has yet to be fully evaluated. The only development so far has been a mini-hydro plant of 4 MW capacity which is be.ng built with Chinese help. As far as petroleum is concerned, there is a potential for offshore reserves but limited exploration efforts have prevented an evalJation of this potential.  - ii - 4. The oil price shocks of the 1970's had caught Sierra Leone totally unprepared to cope with the "energy crisis." It was entirely dependent on petroleum imports for satisfying its non-traditional energy needs and had not developed any indigenous energy resources to substitute for high cost imports. At the same time, as in many other developing countries, there was practically no institutional structure, little or no technical expertise, and modest economic resources to manage this complex problem. The oil price shocks arrived at a time when the country was already facing foreign exchange difficulties because of declining mineral exports. The net result of these developments has been a severe shortage of commercial energy supplies since the early 1980s with attendant dis- ruptions to economic activity. 5. While the problems currently faced by the energy sector are severe, there are grounds for optimism. Firstly, discussions currently underway with the International Monetary Fund and the World Bank should, if an agreement is suc:cessfully concluded, help to ease the foreign ex- change difficulties, a problem at the center of many of the energy sector's difficulties. Secondly, the price of oil has declined signi- ficantly in the past year. While the future course of oil prices is uncertain, if, as expected, prices remain so over the next two to three years, then Sierra Leone has a good opportur ;y to prepare itself for any sharp increases in the future price of oil. The Government should aim to maximize the benefits to the economy from the current low oil price environment while developing a coherent policy for the energy sector that would help avoid a recurrence of the problems currently facing the sector. In particular the Government should urgently consider optionJ which would: (a) encourage the efficient utilization of resources avail- able to the energy sector (focusing on projects with high rates of economic return rather than those saving foreign exchange at low rates of return); (b) enhance the use of domestic energy resources where economic; (c) diversify the country's fuel mix away from a high dependence on oil; and (d) develop an incentive framework and institutional structure capable of adapting to and benefiting from rapidly changing conditions in internaLional energy markets. 6. Within the above setting, this report focuses on the main short- and long-term issues in the sector, evaluating the options open to the Government and making recommendations as appropriate. Short-Term Issues 7. The assessment mission found the energy sector in a serious state of deterioration which has resulted from, and contributed to, the prolonged stagnation in the economy. While oil imports have been restricted for some time, the cutback in 1985 was particularly severe with net petroleum imports declining about 40% below 1984 levels. This development was accompan ed by additional limitations on the imports of spares by the power utility. The resulting disruption to the cconomy  - iii - from the lack of fuel was significant, especially in rural areas. Trans- portation services were greatly curtailed, and electricity supplies limited to a few hours a day with some areas going without electricity for as long as a month. Even the output of diamonds--a major foreign exchange earner--was reported to be declining sharply because fuel was not available to power mining equipment. Fuel shortages have also led to a rampant black market in petroleum products, and encouraged inter-fuel substitution: urban households have increasingly switched over to the traditional fuels, enabling fuelwood prices to rise 150% in real terms between 1984 and 1g8', thereby imposing a heavy burden on low income groups. Similarly, to obtain more reliable electricity supplies, large industrial, commercial, and well-off residential consumers have been importing diesel generating sets financed partly from foreign exchange purchased in the "parallel" market, outside the banking system. Over the past three years, the cost of importing these generating sets is conser- vatively valued at about $10 million--further exacerbating foreign ex- change shortages. 8. The most urgent short-term needs are to improve the supply of electricity and petroleum products and strengthen demand management and institutions. Shortage of Electricity and Petroleum_Products 9. In the petroleum subsector, supply shortages due to foreign exchanoe difficulties have been exacerbated by inefficiencies in the procedures for procuring crude for the refinery. There are two problems associated with crude supplies: 'a) the supplies are not at least cost, and (b) ad hoc arrangements have led to delays in allocating foreign exchange for crude purchases. 10. The mission found that Sierra Leone has been paying an "excess" margin between c.i.f. and f.o.b. values of about $2.50/barrel on crude supplies, equivalent to about $10 million over the past three years (1983-85). Excessive margins of similar magnitude also appeared to have been paid in the last two years (1981-82) of an earlier agreement with the oil company shareholders of the refinery. This is in large part due to the use of open-ended contracts at fixed terms with one supplier. Such contracts are non-competitive and invite monopoly rents whether the supplier is an oil company or, as at the time of the mission, a trading agent implementing the shipping arrangements for the Government's state- to-state deal with Nigeria. Currently, however, the ability of the Government to obtain more competitive terms for oil supplies is compro- mised by the non-payment of the outstanding debt (totaling $58 million) to the oil company shareholders of the refinery for past deliveries of crude: the presence of the debt limits the number of suppliers prepared to supply crude. II. Despite the excessive margins paid for crude supplies over recent years, deliveries have not always arrived when needed. Foreign exchange problems have resulted in a last minute scramble to find suffi-  - iv - c,ent funds for each cargo. There is no system for budgetitig or allo- cating foreign exchange for purchasing oil supplies on an annual basis and then properly programming this allocation to each shipment. 12. The mission recommends that the system for procuring crude oil supplies should be reorganized and made more competitive. The state-to- state deal with Nigeria is in Sierra Leone's interest and should be retained; this recommendation is based on considerations of proximity (the refinery's economic viability, as discussed below, is closely tied to the availability of Nigerian crude as it provides transport cost advantages over other crudes), and the suitability of the crude to local market demand and assumes that Nigeria will continue, as in the past, to offer the crude at competitive terms relative to the international mar- ket. However, international bids should be invited for the management of crude supply and shipping arrangements At the same time, the system for allocating available foreign exchange should be reorganized so that delays in oil payments can be avoided. These measures together with an increase in cargo sizes (to 30,000 tons from the current level of 20,000 tons) should comfortably reduce margins by $3/barrel, equivalent to more than $5 million a year, and enhance the reliability of oil supplies. However, effective implementation of such an arrangement will require at least a demonstrated government willingness to address the problem of the refinery's existing debt. 13. In the power subsector, the public utlity is in a state of physical and financial disrepair, reflecting years of neglect, poor management and a shortage of foreign exchange. At the time of the mission, less than half the installed generating capacity was in service, between a third and a half of the electricity generated was being lost in technical and non-technical losses, and consumers were experiencing frequent and prolonged supply interruptions. These problems intensified further in the mont[s following the mission, reaching crisis proportions in September 1986 when only 6 MW of capacity was available to service an estimated peak demand of 25 MW. While supply has improved since that time and demand of about 18 MW is currently being met, the fact remains that the National Power Authority (NPA) is unable to meet its responsi- bilities of providing adequate and reliaole power supplies. NPA's current difficulties are the latest in a long history of managerial, technical, and financial problems at the utility during the last twenty years. 14. The major short-term issue relates to the institutional, physical, ana financial prerequisites for a successful rehabilitation program. These include: (a) an institutional framework which would allow the utility to operate autonomously, give adequate compensation to skilled staff, and ensure the effectiveness of senior management; (b) re- pairs and improvements to existing generating facilities; and (c) changes in tariffs and access to required levels of foreign exchange. Components (a) ani (c) of the rehabilitation programn are discussed below under institutional reform and demand management. In terms of physical rehabil:.tation, the operational deficiencies associated with the decline  v in power supplies arise out of two related factors: a lack of preventive maintenance and a chronic shortage of spare parts due to foreign exchange limitations. 15. The main priority is for arrangements to be made for over- hauling and repairing all the generating plant at the Kingtom station near Freetown, along with the importation of spare parts to maintain the system in a good state of repair. Enhancing the reliability of power supplies will save the country an estimated $4 million annually by discouraging the importation and use of small (and less efficient in terms of fuel consumption) private generating sets. These savings would help to pay for a major part of the estimated $10 million of necessary rehabilitation expenditures. At the same time, an independent review should be carriei out urgently to determine the details of the rehabilitation program, including the technical and economic viability of repairing the (three) MAN units at Kingtom which are nearing the end of their normal working lives. Contingency plans should also be made for new generating capacity to replace these units in case they cannct be economically repaired. Institutional Reform 16. Institutional deficiencies represent a major obstacle to the efficient and reliable supply of energy. These deficiencies include: (a) the absence of a single ministry for formulating and coordinating energy sector policy, and (b) the absenc' of incentives on the part of parastatals to supply energy at least cost. In the case of both electri- city and refining the last problem is further complicated by a blurring of the distinction between government and parastatal roles in operations and investment. The measures proposed below complement the components of an economy-widp restructuring of the public enterprise sector proposed as part of the Structural Adjustment Loan program, currently under discus- sion between the Government and the World Bank. In addressing these problems, the mission evaluated (i) the scope for coordinating policy in the energy sector, (ii) the effects of lack of autonomy on the function- ing of the parastatals, especially NPA, and (iii) the scope and options for reducing costs in the parastatal companies. 17. Absence of Coordination. There is no single ministry or autonomous body with the political mandate and status to coordinate the activities of the various ministries and develop a coherent energy policy. At the moment, there are five ministries dealing with specific energy subjects in an uncoordinated manner. This lack of coordination has, in turn, contributed to the chaos experienced during fuel short- ages. The Ministry of Energy and Power (MEP) is in theory responsible for coordinating all energy activities. In practice, the MEP's main preoccupation has been with electrical power and water distribution; the MEP does not, in effect, have the status or the political mandate to carry out a sector coordination role. To strengthen sector management, it is recommended that the MEP should be given the status and support necessary for coordinating activities in the energy sector. In addition,  - vi - to make the proposal effective, the institutional responsibility for crude oil and product imports, refining, and oil product marketing which currently rests with the Ministry of Trade and Industry (MTI) should be transferred to MEP. At the same time, MEP should assume responsibility for overseeing conservation policy and programs. For implementing these changes, it is recommended that (a) appropriate resources should be made available to MEP; (b) MEP should draft a detailed proposal of what its enhanced role would entail and how it plans to carry out its increased responsibility; and (c) consensus should be established at the highest levels of Government (i.e., within the Cabinet of Ministers) for such a change, especially as it will involve, as envisaged, a transfer of responsibilities for an important portfolio, namely petroleum supply, refining and distribution. 18. Separation of commercial and administrative functions. There is both political interference with, and lack of management autonomy among, the parastatal energy agencies (NPA, NP, the Refinery). This problem applies particularly to NPA, which operates as a department of MEP. The General Manager )f NPA lacks a.,tonomy and is subject to daily interference from the MEP. The 1982 NPA Act is intended to provide NPA with a high degree of autonomy through an independent Board. In practice this is not the case, and the Board itself is highly susceptible to MEP interference. The lack of autonomy undermines the effectiveness of NPA Management. To increase the effectiveness of NPA, the NPA Act should be reviewed and modified to give the Authority greater autonomy in its management. This legal separation should be carried out as part of an action program for changing the institutional framework within which NPA operates before the implementation of any physical rehabilitation measures. Unless such a change takes place, physical and financial rehabilitation will not bring about a sustained improvement in the public power supply, and which should be the objective of the rehabilitation program. More recently, there have been improvements in the relationship between NPA and MEP with the appointment of a new Board. While these changes are in the right direction, the basic problems of autonomy and lack of representation from the businesb community remain: (a) the Board is still over-represented by incumbent or ex-civil service per3onnel; (b) it does not have members from the private sector, which would provide the Board with the benefits of sound commercial expertise; and (c) the limited autonomy that the Board may enjoy currently is dependent on the good will of the incumbent Minister of Energy and Power. The Government, while agreeing to the recommendation in principle, has indicated that pulitical constraints will, at present, prevent changes tn the NPA Act. However, it has agreed to these changes as a target for the future. 19. Other components of an action program should include: (a) a mechanism to ensure that (i) NPA can regularly set tariffs at levels which allow it to recover its costs on an agreed basis, and (ii) it has access to sufficient foreign exchange for importing spare parts and carrying out routine equipment maintenance;  - vi.i - (b) a requirement that in return for greater legal and financial autonomy, NPA should be run on a ccmmercial basis achieving a target rate of return and be fully accountable for its perfor- mance in providing reliable power supplies; (c) the focusing of MEP's efforts on power sector policy formula- tion and coordination--one of its first tasks should be the development of a policy statement outlining the Government's strategy for the sector in the short and longer term; and (d) an increase in the remuneration of skilled NPA staff to make them competitive with the private sector. (This will require a premium to be paid above public service scales as a compensa- tion for technical skills.) 20. Cost effectiveness and competition. The cost effectiveness of the operations affecting the parastatal companies, the Sierra Leone Petroleum Refining Company (SLPRC) and the power utility (NPA) can be enhanced through greater market competition combined with a strengthuning of the monitoring and regulatory functions of Government. 21. The possibility of reducing crude procurement costs by putting up supply contracts to international tender has already been touched upon above. A rela.ed area, under the supervision of the Department of Trade and Industry (DTI), where current arrangements are not conducive to least cost supply is that of petroleum product supplies. Because SLPRC has a monopoly over the supply of refined products to the domestic market, the country cannot take advantage of favorable conditions in international markets by importing products when they are cheaper than the refinery's output. The economic losses associated with these arrangements may on occasion be considerable and are not likely to be entirely eliminated by measures for more efficient use of the refinery as proposed below. DTI (or, if responsibilities are transferred, MEP) should therefore consider the possibilities for introducing competition into product supply arrangements. One possibility is to open up domestic product market to international tender. Another alternative is to require SLPRC to supply refined products from the cheapest possible source, combined with a linking of ex-refinery prices to spot markets. 22. NPA has a monopoly on national electricity supply. At present, the role of private electricity suppliers is confined to autogenera- tion. NPA exercises this monopoly in the provinces by managing and coordinating the operations of 22 widely dispersed provincial stations. This coordination is a costly overhead and NPA should consider other options for increasing cost effectiveness. These should include: (a) decentralization of operations so that they can be managed on a local basis, (b) selling the Provincial Systems to the private sector or local community groups, when such possibilities arise, and (c) establishing joint ventures with the mining companies for generation facilities that supply both the mines and local communities.  - viii - 23. Cost effectiveness has also been weakened by deficiencies in the monitoring and regulatory functions of gcvernment. The regulatory role normally pleyed by government is reduced because of a severe lack of resources (Forestry Department), lack of access to petroleum market intelligence (MTI), and organizational deficiencies such as the abselce of planning cells. The full impact of this deficiency is difficult to quantify but, as an example, US$5 million per annum in excess payments for oil imports over the past five years, as already discussed above, could have been avoided if the DTI could have monitored the oil market nore effectively. It is, therefore, important that more resources are made available to the DTI (or, if responsibilities are transferred, MEP) and the Forestry Department to carry out their regulatory functiens effectively. Demand Management 24. In conjunction with supply side measures the problem of severe energy shortages should be addressed Lhrough better demand management, primarily through pricing policies which reflect economic cost. In particular, prices for eLectricity and petroleum products need to be raised significantly to cover the opportunity cost of supplies at realistic exchange races. The prices of petroleum products in Sierra Leone are well below their opportunity cost to the economy; even after the recent increases announced in the April 1987 budget, retail prices are about 25% below the level which the mission estimates should be charged to reflect (a) the opportunity cost of oil products, based on crude oil prices of $20 per barrel, and (b) a tax structure which is an average by West African standards. The Government lacks a cohesive demand management policy. Both the level and str'icture of prices need to be reviewed immediately to eliminate pricing distostions and reduce the subsidies that magnify the financial troubles of the Government and subsector organizations such as the refinery. Beyond this pricing review, the Government should reconsider the objectives and policy underlying the establishment of ex-refinery and retail prices, parti- cularly at this time of uncertainty in international oil markets. Furthermore, a pricing review should be carried cut in conjunction with an evaluation of product supply strategy (as discussed above in the con- text of institutional reform at the refinery) and not independent of it. 25. Together with a strategy for rehabilitating the supply of electricity, the problem of power shortages should also be addressed through better demand management including changes in power tariff policy. In the long run, tariff increases would help to alleviate shortages by dampening demand, but because of the magnitude of suppressed demand, the more immediate impact of tariff increases is likely to come from an improvement in supplies through a strengthening of NPA's finances. At the time of the mission, tariffs were well below both the short and long run costs of providing power and thus contributed to NPA's precarious finances.  - ix - 26. In November 1986 the Government, in discussions with the Bank, agreed to raise the tariffs by 300% to Le 3.75/kWh (or US15c/kWh, at an exchange rate of Le 25/$). This is a move in the right direction. However, given the recent rapid decline in the value of the Leone (in March/April 1987 the official exchange rate was fluctuating in the range f Le 45-<5/$) further increases will be necessary to bring tariffs into line with 'he financial and economic costs of supply. Tariffs need be raised to about US20C/kWh to allow full cost recovery and an 8% return on revalued assets (as agreed with the Bank). Tariffs, if they are set at these levels, will be at the upper end of the range (UScIO-20/kWh) for West African countries. In part, this reflects rhe relatively high level of technical and non-technical losses in Sierra Leone and, hence, the urgent need to improve the efficiency of public power supplies. While losses are likely to decline in future as a result of efficiency improvements, it is expected that they will remain at sufficiently high levels (15-20%) to maintain the pressure on long run supply costs; the long run marginal costs of supply to customers on the Western Area grid has been estimated by the mission to be within the range of US16.5- 23.8c/kWh. 27. In addition to the issue of tariff levels, there are various distortions and anomalies in NPA's current tariff structure. In addition to measures recently taken, the mission recommends that NPA should com- plete a tariff study to design a more logical tariff structure that would help to meet NPA's financial, objectives, reduce the existing anomalies and distortions, and make suggestions on how tariffs should be changed during a possible transition to a hydro-based system. Long-Term Issues 28. For the long run, the focus of the report is on the efficient development of energy resources to meet the growing demands for different types of energy in the !conomy beyond the turn of the century. The long term issues relate to )ptions for meeting future power requirements, the disposition of the refit,ory and fuelwood supplies. Options for Meeting Future Electricity Demand 29. There are two aspects of the long-run power issue. The first concerns the viability of the Bumbuna hydroelectric scheme in relation to other options to meet future demand in the Western Area. The second relates to development options for isolated provincial centers. 30. Western Area. For meeting future demand for the Western Area, there are a number of options open to Sierra Leone. These include alternative hydroelectric schemes as well as the possibility of a thermal system based either on coal or oil located at Freetown. The hydro- electric schemes which are technically viable include a 180 MW plant on the Mano River and several versions of the Bumbuna scheme on the Seli  x River. An evaluation of these options, to meet future demand to the year 2020, suggests the following two general conclusions: (a) The rehabilitation of Kingtom and Falconbridge plants should provide sufficient capacity to meet demand in the Western Area well into the 1990s. The chances for an earlier commissioning of new generating facilitivas would rise if the older (MAN) units at Kingtom cannot be restored to full service. (b) A reduced form of the Bumbuna scheme (followed by a pure thermal development) with an installed capacity of 47 MW is the least cost solution within the set of thermal and hydro options examined, over a wide range of load forecast and oil price assumptions. This development will also allow diversific,tion of the country's energy supplies and provide some insurance against oil supply disruptions. The-e may, however, be some room for improvement in the design and sequencing of the project, which as currently envisaged by the consultants is not optimal. A preliminary estimate of the balance of payments impact of the Reduced Bumbuna scheme suggests that the outflows for debt servicing will be partially offset by the gains from the decline in oil imports due to the use of the hydroelectric scheme. The results show that over the life of a 12 year loan (at a 10% interest rate) the Reduced Bumbuna scheme will lead to greater net outflows than the pure thermal scheme, but the level of excess outflows is relatively small--on average of $1- 2 million per annum. Beyond the loan repayment period, the investment will save between $5-10 million--depending upon oil prices--per year in foreign exchange through the reduction in oil imports. The balance of payments impact will be more favorable to Sierra Leone if it can obtain better terms for financing the investment than has been assumed in the analysis. While the reduced Bumbuna scheme will not add significantly to debt servicing compared to thermal alternatives, initial financing of such a lumpy investment could create problems, given the constraints on foreign exchange resources (including borrowing) over the next five years. However, financing constraints by themselves shuuld not be a sufficient rationale for cancelling or significantly delaying the project if it is economically and technically sound. The financing aspects of the problem need to be looked at more carefully including: (i) the economic rate of return on the project relative to those on other investments which also have claims on foreign exchange resources; (ii) the magnitude of suppliers credit available for the project (which would represent additional resource), a ' (iii) options available to the Government for better management of the financial problems caused by the investment in Bumbuna.  - xi - 31. Provincial systems. There are a number of mini hydro schemes, identified in recent studies, which could meet a high proportion of the r.ining sector demand with consequent savings in fuel imports. The via- bility of these schemes has not been fully studied, but two schemes, Singimi Falls (7.2 MW) and Benkongor Falls (10.8 MW), merit further evaluation. Initial assessment of the schemes, which are not run of the river, suggests that they will yield high economic returns. 32. There are a number of possibilities for interconnect;ng isolat'2d power systems. These should be carried out if economic. However, a policy should be identified for future interconnections to prevent haphazard development and to ensure that the design meets national standards. 33. Main recommendation. NPA should attempt to resolve a number of unanswered questions before making a final decision on the Bumbuna scheme and the development pian for the Provincial Systems. Among other things, NPA should (a) review the design, sequencing, and financing of the reduced Bumbuna project; (b) investigate the prospects of exporting surplus energy from Bumbuna to Guinea; (c) evaluate the viability and optimum sequence of development of the more promising mini hydro schemes; and (d) define technical standards to be adopted in interconnecting isolated Provincial Systems. The Economics of Refinery Operations 34. The refinery, run entirely by Sierra Leone nationals, is operationally and managerially sound even though it is suff-ring from crude shortages, attendant shutdown and start up of operations, and is technically bankrupt from assuming the cumulative interest on the Government crude debt to the oil companies. There is a case for retaining the refinery as an operating entity. Based on current and projected long term crude and product prices, it is marginally viable; however, it also constitutes a center of managerial and technical excellence in Sierra Leone, a hydrocarbons testing laboratory, and provides employment and training spin-offs. 35. At present, the refinery operates on a cost plus basis providing little incentive for improving technical and economic effi- ciency or for exploiting market opportunities. As discussed in the section on institutional reform, there is a need to make the refinery more responsive to market conditions either directly through more competition, or indirectly through a pricing scheme which takes account of conditions in international markets. Even if the refinery is economic on a long run basis there is no rationale for operating the refinery if price relationships (say through dumping of products) are such that the gross margin is negative (i.e., variable costs are not being covered). If these conditions persist for a significant period of time, it would make sense to shut down the refinery's operations without firing the staff. Such a policy would require close monitoring of the gross  - Xii - margin. By ensuring that the refinery does not operate when the gross margin is negative, losses, if incurred, could be restricted to about the level of the refinery's fixed costs, $1.5 million. This policy is more likely to be effective if the refinery is allowed to run as a commercial entity. 36. The main factors which support the refineries' viability are: (a) low operational cost due to reLatively efficient operations and the absence of (expensive) expatriate staff; (b) freight advantage of transporting crude relative to products due to Sierra Leone's location-- the cost of transporting crude from Bonny in Nigeria (the usual source of crude supplies) is about half the cost of transporting products from Rotterdam, the base location for determining product prices, to Freetown; and (c) sunk capital costs of existing faczilities and the relatively modest needs for future investment over the next 10 to 15 years. The refinery's long term viability is assured as long as it is run effi- ciently, no major changes in refinery configuration are contemplated and Sierra Leone continues to obtain crude at competitive terms from Nigeria rather than from a more distant source. As the refinery's economics are finely balanced, a significant change in any one of the above three variables, while unlikely, could make the refinery an uneconomic opera- tion and the Covernment should be aware of this sensitivity. 37. Although profits are likely to be modest, the refinery's financial viability can be sustained as there are no major investment requirements anticipated over the foreseeable future. The start up of the Bumbuna scheme could reduce the refinery's profits by about $0.5 million but would not alter its economics significantly. These losses reflect the penalty to be paid on the sale of -urplus fuel oil at export parity prices. 38. In terms of financial resuilts, the performance of the refinery has been adversely affected in the recent past by (a) increased fuel consumption and losses associated with frequent start up/shut down of operationis due to crude supply interruption, (b) delays in adjusting ex- refinery prices following increases in the Leone price of crude supplies, and (c) the continued inclusion of accrued interest on the Government debt to the oil companies in the refinery accounts. 39. Main Recommendations (a) The refinery should be retained as an operating unit. However, a least cost strategy should be implemented requiring the ref'inery to operate only when variable costs are covered. Its legal statutes and pricing policies should be reviewed to ensure that it becomes more responsive to international market conditions.  - xiii - (b) It should operate as a receiving terminal when prices make operations less attractive. (c) The validity of charging accrued interest on oil company debt to the refinery's accounts should be teviewed. Household Energy/Woodfuels Stratety 40. Estimates by the mission have confirmed the Government's concerns "hat risi:ig fuelwood costs are absorbing a significant propor- tion of the income of urban households (as much as an estimated 40 of monthly expenditures). In addition rising prices are a symptom of growing regional shortages and deforestation near urban areas. Without some measures for increasing supplies or reducing demand, especially in the Peninsula Area as well as in the vicinity of Bo, Makeni, Koidu, and Kenema, the cost of fuelwood supplies is expected to grow. 41. The most promising strategy fcr reducing woodfuel costs is the increased use of improved stoves, whose efficiencies are about twice as high as conventional woodstoves currently available in the market. These stoves, which would be locally manufactured, could reduce wood demand by 40-60% while incurring modest investment costs on the part of the con- sumer. A program of demand management centered on the utilization of improved stoves should be initiated. No expenditures should be required from the Government other than those for coordinating, testing, and promoting private production and marketing of improved stoves. As the stoves are made from locally available scrap metal, there should also be no drain on foreign exchange resources. 42. Supply side measures should atso be initiated as part of an integ.-ated approach to provide household energy at least cost. However, there are a number of problems associated with supply side measures involving new fuelwood plantations. These include: (a) uncertainty about costs and competitiveness of the wood from these plantations; (b) resource constraint faced by the Forestry Services who will implement these measures; and (c) uncertainty about the link between demand growth and deforestation. The mission recommends that the Government review plans for fuelwood plantations. Lower cost options should be examined, including agroforestry. Investments in the next few years should be limited to a few demonstration plantations for obtaining better infor- mation on yields and costs under different modes of management. 43. A higher priority on the supply side than plantations should be given to better utilization and management of existing resources. Such a program could include the following components: (a) studying the poten- tial for increasing supplies from mangrcves and bush fallow agriculture, (b) natural forest protection, and (c) reviewing possibilities for better controlling supplies to urban areas, for example through, city gate  - xiv - cAlIection of taxes, building of feeder roads to surplus areas, and increasing the efficiency of traditional charcoaling techniques. Investment Priorities 44. The following projects should receive the highest priority in the public sector investment budget for the period 1987-1991: Power: (a) Rehabilitation of Kingtom and Falconbridge power plants ($6.5 million). (b) Strengthening of transmission and distribution networks in the Western Area to meet expected load growth following rehabil.tation of power plants ($3.3 million). (c) Completion of Bo/Kenema transmission liaik ($1.8 million). (d) Improvements to Provincial Systems, including spare iarts and communications ($3 million). (e) Projects defined by priority investigations listed below under technical assistance. Petroleuim: There are no major petroleum investments to be financed by the public sector. A small amount of investment ($1.0 million) will be required for the Government's share of building security stock of crude oil to 30 days of consumption (about 20,000 tons). (Another $0.55 million of public investments may be required for establishing retail outlets for kerosene and diesel in the rural areas and for the Government share of investments in ernergy efficiency improvements at the refinery, if these investments are considered viable.) Woodfuels/Household Energ: Demonstration fuelwood plantations ($1 million). Deferred Investments 45. The following projects should be deferred and reviewed subsequent to other recommended action. (a) Major new generation and transmission projects pending completion of Master Plan Study and review of the viability of rehabilitating existing generating plant.  - xv - (b) Heavy fuel generating plant investment ($7 million) planned for Bo, under assistance from Danida, pending review of the viability of the scheme. (c) Major fuelwood plantation investments pending completion of a review of the viability of such schemes based otn the experience of ongoing and proposed demonstration projects. IPriority Investigations and Technical Assistance 46. The following investigations and technical assistance should be initiated during 1986/87. Power: (a) Consultancy services for evaluatin,g details of planned .ehabilitation, including repairs to MAN units, and improvements needed in management practices and financial control at. NPA ($0.1 million). (b) Various studies incltuding those related to the Bumbuna scheme, mini hydro sites, and tariffs ($0.2 million). (c) Power loss reduction study ($0.1 million). Petroleum: (a) Review of petroleum pricing structure and associated supply strategy ($50,000). (b) Design and development of accounting and information systems at the DTI for monitoring and regulating domestic market ($200,000). (c) Exploration promotion or! offshore relinquishments ($0.5- 1.5 million). (d) Evaluation cf the viability of (i) establishing retail outlets in rural areas to be operated by cooperative/local groups, and (ii) investing in energy efficiency improvements at the refinery ($50,000). Fuelwood: (a) Preparation of a program for testing and promoting private production and marketing of improved stoves for household cooking ($600,000).  - xvi - (b) Strengthening of forestry department in areas of agroforestry and extension services as well as in the setting up of an Improved Stove Unit in the MEP for coordinating a program for improved stoves ($500,000). (c) Technical assistance (from donors) to eialuate the scope and benefits of establishing regulatory control over the supply of woodfuels to urban areas and improving the efficiency of charcoal production ($50,000). Energy Sector: (a) Provision of technical assistance to MEP for advising on and implementing improved energy sector coordination and planning ($200,000). 47. While all the recommendations merit prompt consideration, perhaps the most important of these concern the need for technical assistance to increase Sierra Leone's capacity for developing and implementing a strategy for meeting its future energy needs.  I. ENERGY AND THE ECONOMY The Economy 1.1 Sierra Leone is a relatively small (72,000 km2) tropical country along the West African coast. It has a population of 3.7 million people with v per capita income in 198" of US$330 and low levels of social development: life expectancy is only 38 years, with infant mortality at 20%, 88% of the population has no access to safe water and about 80% of the adults are illiterate. 1.2 The country has good agricultural, marine, mineral, and hydro- power potential but is one of the least developed countries in the world. GDP grew by about 4.3% per year in the 1960s but at an average rate of only 1.9% between 1973 and 1983. With the population growing by about 2.5% annually, real per capita income was declining over the latter period. 1.3 The sectoral composition of GDP in 1984/85 is estimated at 35% agriculture, 20% transport and communication, 11% trade and tourism, 10% mining and quarrying, 6% manufacturing and handicrafts, and 18% other services. Foreign exchange is earned mainly by exporting minerals (i.e., diamonds, bauxite, rutile, gold, iron ore). Mineral exports have accounted for about 70% of export earnings in the 1980s. 1.4 The sluggish growth of GDP experienced in the 1970s, together with a declining trend in per capita income, has continued with an average annual rate of 2.1% from 1979/80 to 1984/85. Underlying the poor performance of the economy in the past 15 years are two distinct trends. The first is a cyclical one caused by declining export earnings, in turn the result of fluctuations in commodity prices and the wor.d economy. The second is a secular trend marked by the gradual depletion of Sierra Leone's major mineral resources of alluvial diamond and iron ore. In recent years the recorded production of diamonds, by far the most important source of foreign exchange, has fallen to less than one- tenth its level of a decade ago, while the production of iron ore ceased altogether in 1976 and was resumed for only two years on a smaller scale in 1982. Exports of other commodities have not yet substituted for this decline. External Trade Gap 1.5 The Government has not successfully adjusted to either of the adverse trends. With increasing imports and escalation of oil prices, the trade balance deteriorated in the late 1970s and early 1980s. External funds borrowed to finance the rising trade and current account deficits accumulated into a public debt of over $400 million, equivalent to 27% of GDP, by end of fiscal year 1983. While the size of the trade  gap has diminished since 1983, the overall balance of payments position has remained under pressure: ir spite of a 20% improvement in exports during FY85, scheduled debt service payments in that year accounted for 25% of !xport earnings, the current account deficit was $87 million, about 6% of GDP, and external arrears had increased to $285 millior, more than double the level of exports. In conjunction with mounting arrears and a rising debt service burden, most of Sierra Leone's foreign exchange earnings have, for a few years, been flowing outside the banking system through the "parallel" market where the exchange rate has been much higher than the official rate. The combined effect of these developments has beeli an acute shortage of foreign exchange at the disposal of the Government. Recent Developments 1.6 By the time of the mission in February 1986, the economic and financial situation had become precarious, with a deteriorating balance of payments, a thriving parallel market in foreign exchange, widening budget deficit, and inflation accelerating to a level in excess of 100% per annum. To recover from this economic crisis, the Government, on June 27, 1986, announced a bold package of reforms as part of its FY87 budget. The major elements of this program are: (E) a flexible exchange rate; (b) the removal of all subsid;es on petroleum and rice; (c) a more liberalized trading regime with almost all commodities freely importable; (d) substantial increases in the producer prices of major export crops; (e) removal of controls on the prices of all commodities except monopoly public utilities; and (f) structural adjustment changes to be implemented in the areas of public sector management, allocation of public expendi- tures, agricuitural policy, and industrial incentives. These reforms, still in various stg,ges of implementation, are consistent with the Government's dialogue with the International Monetary Fund (IIMF) and the World Bank. 1.7 The macroeconomic environment, which is still adjusting to the sharp changes in policy, shows no sign of stabilizing as uncertainty persists regarding the Government's resolve to fully implement announced reforms. A stand-by agreement reached with the IMF, and approved by its Board in November 1986, collapsed soon afterwards as agreed targets on budget deficits and credit ceilings could not be met. The collapse of the stand-by arrangement with the IMF has coincided with a worsening of the economic situation: shortages of petroleum products have increased, inflation has accelerated sharply, reaching the 150% per annum rate in the last six months of 1986, and the currency has depreciated rapidly. The official exchange rate was fluctuating in the Le 45-55/$ range in March/April 1987 compared with Le 5.10/$ at the time of the mission, one year earlier. A major factor underlying the rapid depreciation of the Leone has been excessive money creation to fund the increasing fiscal deficit; the basic problem is that too many Leones have been chasing too few dollars. At the same time, the amount of foreign exchange surren- dered to the banking system has declined due to uncertainty surrounding the implementation of the IMF program, thereby adding  -3- pressures on the Leon;.. 1.8 More recent discussions, held during March 1987, between the Government, the Bank, and the IMF, have led to the definition of a "shadow" program outlining new short term targets for fiscal deficits, credit creation, and the removal of subsidies on rice and petroleum products. Sierra Leone's performance under the "shadow" program will provide a vital test of the Government's commitment to economic reform; a successful implementation of the program is expected to pave the way for a new stand-by agreement with the IMF and a structural adjustment loan (SAL) from the Bank during FY1987-88. Economic Scenarios for Sierra Leone 1.9 In evaluating the issues in the energy sector and the implica- tion of different options, two scenarios--alternative views of what the world would look like in the future--have been devel. ped. The key difference between the scenarios relates to the timing and extent of the adoption in Sierra Leone of the economic reforms being currently discussed with the IMF and the Bank. It may be mentioned that on several occasions in the past the Go-vernment, for domestic reasons, has been unable to follow through on prior agreements with the Bank and the Fund. Hence the two scenarios would remain valid even if current discussions lead to a signed agreement. 1.10 In one scenario--Accelerated Growth (AG)--the proposed reforms (including significant increases in the prices of products to bring them in line with economic costs) are implemented to a large extent, if not completely, over the next two years, paving the way for significant capitai inflows to the country. In the second scenario-- Base Case (BC)- -the reforms are implemented in a very limited way and they come too late--late 1980s or early 1990s--leading to prolonged economic stagnation. 1.11 The external environment, including oil prices, is assumed to be the same in both scenarios, excep: for the analysis of the power expansion plan where two oil price profiles are proposed for each scenario. In effect, therefore, there will be four swenarios for the evaluation of power expansion options including the Bumbuna scheme. 1.12 Details of these scenarios are give-i in Annex 1 both in terms of qualitative descriptions and quantitative values. Table 1.1 summarizes the GDP and oil price assumptions underlying the two scenarios.  Table 1.1: SCENARIOS (SUMMARY) 1986-90 1990-95 1095-2000 GOP Growth (% p.a.) Base Case (BC) 0 2 2-3 Accelerated Growth (AG) 3 5-6 4-5 1986 1990 1995 2000 Base oil prices a/ (S/barrel) Current prices 20 23 39 48 Constant 1986 prices 20 18 25 25 Alternative oil prices b/ (S/barrel) Current prices 10 13 30 40 Constant 1986 prices 10 10 19 21 a/ Prices common to both scenarios. b/ Alternative profile used for evaluating sensitivity of power expansion plans to oil prices. Source: Bank projections, mission estimates. Energy-Economy Interactions 1.13 Economic activity and energy sector developments have been caught in a downward spiral. Economic stagnation, together with balance of payments problems, have curtailed the demand and supply of energy. At the same time shortages in energy, as well as the oil price shocks of the 70s, have disrupted economic activity. Energy-economy interactions are most apparent in considering the constraint posed by the balance of pay- ments. Lack of foreign exchange has reduced petroleum imports, disloca- ted transportation, and limited the operation of petroleum based elec- trical power plants. Furthermore, shortages of imported spare parts have restricted essential maintenance, increasing energy losses and reducing operating capacity; "firm" electrical generating capacity has less than halved in the previous three years. In turn, the heavy claims of petro-  - 5- leum imports on foreign exchange--these increased from 8% of export earnings in 1972 to over 50% in 1982--(among the highest share as com- pared to neighboring oil importing countries) has constrained non-oil imports and acted as a drag on the economy. 1/ Once the economy recovers from its current malaise, a central issue is whether and how it will meet its growing energy requirement under the prevailing balance of payments constraint. Table 1.2: PETROLEUM IMPORTS AND EXTERNAL. TRADE (million current Leones) 1980 1981 1982 1983 1984 1985 Exports 224.1 176.9 136,6 201.8 331.0 627.7 Crude Oil Imports 67.3 91.4 61,3 88.1 153,1 166.1 Petroleum Product imports 4.9 6.3 21.2 4.2 5.4 0.0 a/ Petroleum Exports and Bunker Sales 14.6 18.5 10.0 18.7 19.8 29.0 a/ Net Petroleum Imports 57.6 79.2 72.5 73.6 138.7 137.1 Net Exports (1) - (4) 209.5 158.4 126.6 183.1 311.2 598.7 Petroleum imports as % of Net Exports (5) * (6) 27.5% 50.0% 57.3% 40.2% 44.6% 23.5% Imports Financed from Outside the Banking System 227.5 194.8 277,4 9.8 152.0 466.8 Adjusted Exports b/ (6) + (8) 437.0 353.2 404.0 192,8 463.2 1075.5 "True,, Ratio of oil imports to exports (5) - (9) 13.2% 22.4% 17.9% 38.2% 29.9% 12.7% a/ Exciudes values of fuel oil exported in exchange for importation of gasoHine. b/ Assuming imports financed outside Banking system are equal to the total export earnings retention in the same year. Source: Bank of Sierra Leone, IMF statistics, mission estimates. 1/ It is important to note that if the official exports are adjusted for unofficial earnings through the "parallel" market, then the proportion of foreign exchange earnings claimed by petroleum imports is reduced (see estimate in Table 1.2). The burden is nevertheless still significant.  - 6 - Key Energy Issues 1.14 An overview of total and sectoral energy consumption in Siert; Leone is presented in the 1984 energy balance (Annex 2). Primary energy supply was about 1.2 million (toe) in 1984, or 310 kilograms of oil equivalent (kgoe) per capita, which is about average for countries with similar income and among the lowest levels in the world. About 80% of primary energy supplies were provided by domestic biomass energy in the form of fuelwood. The remaining 20% of primary energy requirements are met by imported crude oil and its products, of which about 30% are used is Sierra Leone's existing thermal-only power generation system. These major sources of primary energy, fuelwood and petroleum imports, are both associated with supply and financing problems which have serious consequences for the economy as a whole. 1.15 There is evidence of increasing fuelwood shortages near the urban centers combined with a deterioration in the size and the quality ,f forest resources. The long term environmental effects of the deple- tor of forest cover and overcutting are difficult to estimate and could be significant. However, the more direct and obvious problem is the steadily rising costs of fuelwood consumption: as revealed by surveys carried out by the mission, shortages .n urban areas have led to large real term increases iii the price of fuelwood over the past few years and fuelwood consumption now absorbs a relatively high proportion of house- hold income (see Chapter IV). The situation is particularly worrisome in the Peninsula Area, as well as in the vicinity of the towns of Bo, Makeni, Koidu, and Kenema. 1.16 Yet, in spite of the urgency of corrective action, investments in the forestry sector have been minimal so far, Less than 2% of extern- ally financed investment, which accounts for around 75% of all fixed capital investments in Sierra Leone, was allocated to projects for afforestation and charcoal production during 1981-84. Over the same period, the electricity subsector, which represents only 2% of final energy demand, received over five times as much financing as the forestry subsector. While the importance of electricity to the economy is likely to be much greater than that revealed by its share in final encrgy consumption (in part because the end-use efficiency of using electricity is five to six times higher than that for traditionaL fuels), there is a need for the Government to review the composition of energy sector investments in general and the allocation to forestry and fuelwood related investments in partic lar. 1.17 Imported petroleum products are the only source of primary energy supply to the modern sector other than a limited amount of char- coal and fuelwood used by industry. As discussed above, petroleum im- ports are placing a significant burden on the balance of payments and the  -7 - Government has been restricting imports since the early 1980s. Despite these restrictions, the proportion of foreign exchange absorbed by oil imports has exceeded 40% since 1981, except for 1985 when the share dropped to 23%, largely due to a 40% cutback in the volume of net petroleum imports. The resulting disruption to the economy from the lack of fuel has been severe, especially in rural areas. 1.18 There are three issues which require the immediate attention of the Government. The first one relates to the existing shortages in the supply of electricity and petroleum products, and the supply and demand side measures that should be undertaken to alleviate them. The second one involves weaknesses in the coordination and management of the govern- ment ministries and parastatal agencies dealing with energy. These weaknesses represent a major obstacle to the efficient and reliable supply of energy and exacerbate the disruptions caused to the economy from supply shortages when and as these have occurred. There is a need for the Government to evaluate options for making these agencies more effective. The third short-term problem relates to distortions in energy pricing which, as described in later sections of the report, have contri- buted to financial problems for the Government, the power utility, and the refinery and have led to inefficient choice of fuels by consumers. Policy changes to remove these distortions and alleviate concomitant problems should be urgently evaluated. 1.19 There are three long-term issues which focus on the development of the energy sector over the next 10 to 15 years. First, the optimal long-term strategy for developing the power sector must be determined. Within this context, the need for investment in hydroelectric schemes over the next 10 years should be assessed. Here the advantages of reduced dependence on imported oil have to be balanced against the increase in the economy's financial burden which the large and lumpy investment in hydro-schemes implies. Second, the economics of the refinery need to be examined, since it may under specific circumstances be cheaper to import products directly. Third, fuelwood supply at the national and local levels should be studied to determine if there are any impending shortages of this resource and what can be done, in terms of both an investment program and policy initiatives, to improve fuelwood supplies and reduce household energy costs through more efficient use of the resource. Energy Consumption Patterns 1.20 An overview of present and future energy consumption patterns is provided by the energy balance for 1984 as well as projections for  -8- 1995, by scenario (see Annex 2). The implication of these balances for alternative policies is discussed below. 2/ 1.21 Balances for 1984. Final energy consumption patterns based on the 1984 energy balance are presented in Table 1.3 below. Final energy consumption by sector is dominated by the residential sector (including home based craft activities). Households consume 86% of net domestic energy supplies, while Lhe balance is divided up between transport (6%), industry (5%), mining (2%), and agriculture (1%). This distribution reflects the dominant role of traditional energy, mainly fuelwood. This is used, as already discussed, almost exclusively by households for cooking and craft activities, and generally with very low combustion efficiencies. The same comment applies to the use of charcoal. rable 1.3 STRUCTURE OF FINAL ENERGY CONSUMPTION (1984) (%) Sector Fueiwood Charcoal Petroleum Power Total Agriculture - - 6 - I Mining - - 6 38 2 Industry 3 10 15 40 5 Transport - - 49 - 6 Households 97 90 24 22 86 100 100 100 100 100 735 39 118 19 911 ('000 TOE) Source: Mission estimates. 1.22 The sectoral distribution of final consumption of imported energy presents a totally different picture. Transport is the major consumer of petroleum products, accounting for 49%, followed by households (24%), industry (15%), and agriculture and mining 6% each. 2/ Annex 2 also provides an energy balance for 1975 for reference purposes. Household energy use figures for 1975 are based on the assumption that the consumption per capita was the same as estimated for 1984.  - 9 - This distribution reflects the low level of industrialization in Sierra Leone and the insignificant use of petroleum products currently used by households for cooking and lighting. 1.23 Industry (including commercial enterprises) stands out as the biggest electricity consumer with 40% of sales, while mines and house- holds consume 38% and 22% respectively. In energy terms the modern industrial sector in Sierra Leone consumes almosc the same amount of fuelwood and charcoal, totaling 26,000 toe, as it does of electricity and petroleum combined (although the delivered energy in the case of fuelwood is much lower because of conversion losses). Consumption Patterns Under Alternative Scenarios 1.24 Final energy consumption patterns for 1995 under the AG and BC scenarios are presented in Tables 1.4 and 1.5. 1.25 Under an AG world, the net effects of a restructured economy, significant and feasible improvements in the efficiency of energy use, and the commissioring of the Reduced Bumbuma scheme are reflected in the energy balance for 1995. Comparing Tables 1.3 and 1.4, the consumption pattern is still dominated by household energy. However, there is some shift toward transportation and industry, whose share in total final energy consumption goes up from 5% and 6% respective'Ly in 1986 to 8% in 1995. This limited increase in the share of economically active sectors, in spite of fairly rapid growth, reflects both the large weight of tradi- tional energy in current consumption and the emphasis in government policy under AG for growth based on expansion of agriculture rather than industry which is more energy intensive. The sectoral distribution of imported petroleum products also signals the emphasis on agriculture; the share of petroleum products consumed by agriculture increases from 6% to 8%. Table 1.4: FINAL ENERGY CONSUMPTION, 1995, ACCELERATED GROWTH SCENARIO (S) Sector Fuelwood Charcoal Petroleum Power Total Agriculture - - 8 - 1 Mining - - 7 20 2 Industry 2 8 18 50 8 Transport - - 47 - 8 Hoiseholds 98 92 19 30 82 100 100 100 100 100 ('000 TOE) 895 51 198 31 1174 Source: Mission estimates.  - 10 - 1.26 Industry, however, continues to be the biggest consumer of electrical power with 50% of sales. The share of the mining sector declines significantly from 38% in 1984 to 20% due to a reduction in mining activity. Households increase their share of power to 30% partly because of steadily rising incomes. 1.27 In spite of GDP growth in the 4-5% range, total energy consumption rises only at a modest rate of 2.4% p.a. from 911,000 toe to 1,174,000 toe. This largely reflects the slow rate of growth in fuelwood demand (1.8% p.a.); the rate of growth of fuelwood consumption is tem- pered by the use of more efficient stoves by a significant proportion of the population, as discussed in Chapter IV. 1.28 When comparing fuel energy consumption under AG with BC (as given in Table 1.5) the advantages of a restructured economy, and the more efficient use of resources becomes even more apparent. In spite of stagnating economic activity (with GDP growth rates averaging 1% p.a.), total energy consumption averages a growth rate of 2.1% p.a. compared with 2.4% under AG. In the case of commercial energy (power plus petroleum) the growth rate under BC is 2% p.a. (with an average CDP growth rate of 1% p.a.) compared with 5% under AG (with an average CDP growth rate of 4.5% p.a.). If the same ratio between GDP and energy consumption growth had been maintained in AC as in BC, then commercial energy consumption would have increased at a 9% p.a. rate: instead of doubling, the estimated consumption of imported petroleum products (and of electrical power) would have increased three-fold between 1984 and 1995. The effect of policies in AC have significantly lowered the level of energy consumption under that scenario, with attendant effects on fuel imports. Table 1.5 FINAL ENERGY CONSUMPTION, 1995 BASE CASE SCENARIO (X) Sector Fuelwood Charcoal Petroleum Power Total Agriculture 6 - % Mining - 8 25 2 Industry 1 6 20 44 6 Transport - - 46 - 6 Households 99 94 20 31 87 100 100 100 100 100 ('000 TOE) 929 50 148 21 1,147 Source; Mission estimates.  - 11 - Energy Resource Base 1.29 Sierra Leone is reasonably well endowed with energy resources. Its main domestic energy resources are woodfuels and the potential for hydroelectric power. Fuelwood is by far the largest single domestic source of energy in use. 1.30 Anual supply of fuelwood is currently estimated at around 4 million m , 10% of which goes to Freetown. High forests which once covered most of Sierra Leone have been reduced to 4% of the land area through shifting cultivation, commercial logging and cutting for fuelwood. Most of the country is now covered with secondary forests in various stages of degradation, or derived savannah. As fuelwood is obtained from both primary and secondary forests, the reduction in the size of the former does not on its own give a complete picture of the available supply of fuelwood. Reforestation programs have yet to be implemented in any significant way. The issue of future supplies and likely scarcities of fuelwood, both at the regional and national level, is discussed in greate.r detail in Chapter IV. 1.31 There is a large, and virtually untapped, hydroelectric power potential conservatively estimated at about 1200 MW. Technically and economically the most promising site is at Bumbuna with an ultimate potential of 305 MW installed capacity and 1460 GWh of electricity gene- ration. (This compares with existing installed capacity of 134 MW, 98% of which is oil fueled 3/ and a total output in 1984 of 194 GWh.) After initial work, as part of a staged development of the Bumbuna project, further investment has been deferred awaitinig an improvement in economic conditions. Around 20 mini-hydro sites have also been identified for serving isolated load c-enters but their feasibility has not yet been fully evaluated. The only development so far has been a mini-hydro plant (4 MW capacity) on the Goma river which is being built with Chinese help. The viability of, and steps needed for, developing Sierra Leone's hydroelectric potential are considered in detail in Chapter III. 1.32 As yet no crude oil has been discovered in Sierra Leone although an oil company, Amoco, is currently involved in offshore drilling. There is a potential for petroleum offshore but limited exploration activity has prevented an evaluation of this potential. 1.33 There are other domestic energy resources which could provide significant levels of energy inputs in the future, yet their viability has not been examined. These include: crop residues and solar energy for uses such as fish and crop drying where they could substitute for diesel and kerosene; alcohol from biomass such as cassava or molasses, 3/ A seasonal hydro-site exists in the Goma Valley with an installed capacity of 2.4 M. However, it has been out of operation since 1983.  - 12 - which can be used as a gasoline extender or for industrial purposes; lignite deposits, whose size and quality has not been properly surveyed since their identification in the 1920s, and which could substitute for oil in power generation and as a supplement/substitute for fuelwood in the household sectors. The potential for developing other energy resources is evaluated in Chapter IV. Institutional Framework 1.34 A number of ministries are directly responsible for energy sector functions. The MEP is responsible for coordinating all energy related activities. It also establishes and implements policy in the power subsector, including the supervision of the NPA, the electrical power utility. The importation, refining, and distribution of petroleum products is regulated by the MTI, while exploration activities for petroleum (as well as for lignite) fall within the jurisdiction of the Ministry of Miies (MOM). The Ministry of Finance (MOF) has institutional responsibility for allocating foreign exchange for oil imports. The same ministry sets petroleurn product prices in collaboration with the MTI. The manda'. for the management and development of forestry and other woodfuel resources rests with the Ministry of Agriculture, Natural Resources and Forestry (MANRF). There is no ministry with the legal mandate to develop, research, and promote the use of alternative energy resources such as solar energy or agricultural wastes. The focal point for studies and research in this area, as well as on improved stoves and more efficient charcoaling techniques, has to date been the Fourah Bay College of the University of Sierra Leone. 1.35 An important influence on some energy sector activities is exerted by ministries which do not have a direct responsibility for these functions. This is particularly true of hydroelectric development and othier major additions to the power system, since these projects absorb significant resources. Their priority in the national economic plan is reviewed and established by the Ministry of Economic Planning and Na- tional Development. Coordination is required on these projects between the MEP, NPA, and the planning ministry, as well as with a number of special purpose authorities. Other ministries with linkages to the energy sector include Transport and Communications, and Social We'lfare, the latter for its role in trial programs for the dissemination of improved stoves. 1.36 In the private and parastatal sectors the principal energy trading entities are (a) the oil companies in various wholly private or joint vencure arrangements with the Government of Sierra Leone; (b) NPA, the state-owned power utility; and (c) the Sierra Leone Petroleum Re- fining Company, the parastatal which operates the refinery. Fuelwood cutting, transport, and marketing are almost entirely in the hands of small private operators. Over the past few years various Non-government Organizations (NGOs), including YWCA, PLAN International, and Bois de Feu, have become involved in experimental, non-commercial, fuelwood related programs both on the demand side (more efficient use of fuelwood) and the management of supplies (village tree plantings).  - 13 - II. PETROLEUM Introduction 2.1 The modern sector of the Sierra Leone economy is highly dependent on imported petroleum products for its fuel and power requirements. In recent years the economy has been unable to obtain regular and adequate supplies of petroleum and the ensuing shortages have adversely affected the performance of almost all sectors. 2.2 The immediate reason for inadequate supplies has been the country's difficult foreign exchange situation--a problem that is already being addressed at the macro level in discussions underway between the Government, the IMF, and the Bank. However, petroleum shortages, as well as their consequences have been exacerbated by inefficiencies in procure- ment arrangements, price distortions, and the inability of the Government to manage the shortages. The first aim of this chapter is to focus on complementary micro level measures which will help to reduce shortages as well as the cost of supplies. A second objective is to consider the options open to the Government in the longer term to reduce the country's dependence on petroleum and encourage the supply of petroleum products at least cost to the economy. 2 3 This chapter is composed of four sections. Section 1 provides a brief overview of the subsector describing recent developments in supply, demand, and organizational structure. Section 2 looks into short-term decision making in three separate areas: (a) supply procure- ment arrangements; (b) demand management; and (c) institutional coordination. In each area options are examined and recommendations made for raising efficiency and improving the supply demand balance. Section 3 evaluates components of a long term strategy for the subsector in- cluding: (a) the viability of the refinery; (b) measures for enhancing offshore exploration; and (c) establishing a balance between public and private sector activities. Section 4 outlines an investment and tech- nical assistance program for the subsector. Recent Developments and Subsector Organization 2.4 FoLlowing a 3% per annum growth rate during the 1970s, consump- tion of oil products has been on a declining trend sinGe 1979-80 (Table 2.1). The decline in consumption mainly reflects restrictions on imports necessitated by foreign exchange and procurement problems, rather than the effec,t of domestic price changes or the sluggish performance of the economy. 4/ Petroleum supply shortages reached their peak in 1985 when 4/ Domestic prices for products declined in real terms over 1980-85, and GDP is estimated to have grown at a 2.1% rate.  - 14 - net imports of crude and products were restricted to 132,000 tons, 38% below the 1984 level of 212,000 tons. Total product sales in 1985, about 152,000 tons, were below the levels of a decade earlier arn represented a per capita annual consumption of 41 kilograms (kg), one of the lowest for West African countries. The low level of consumption in 1985 helped to save foreign exchange, with oil imports dec.ining in value from $55 million to $33 million, or from 42% to 21% of total export earnings. However, the cutback led to the most serinliq shortfalls to date causing widespread load shedding, excessive service station queues, high black market premiums for petroleum products, and disruption of trarnsportation and other economic activities. Approximate estimates of suppressed demand by the mission suggest that gasoline and kerosene purchases by consumers in 1985 were about 45% below what they would be if sales were unrestricted. These developments underscore the immediate need to improve the reliability of oil supplies and secure adequate foreign exchange for oil purchases. Tjble 2.1: PETROLEUM SUBSECTOR--8ASIC DATA 1'000 tons) 1980 1981 1982 1983 19e4 1985 Petroleum trading Crude imports 222 224 160 181 236 158 Product imports 12 8 39 16 14 9 Product re-exports 51 63 33 46 38 55 Net imports 183 169 6'6 15 212 112 Ref in ng Crude processed 216 223 162 194 221 170 Product output: L=PG I 1 1 1 1 Gasoline 29 29 25 29 35 30 Kerosene 43 44 30 40 41 25 Gasoil 73 77 56 62 78 59 Fuel oil 63 64 44 53 56 47 losses (i) 3.4 3,7 4.0 4.5 4.6 5.1 Growth Product sales Rate (%) 1980-85 LPG I 1 I 1 I 1 +2 Gasoline 41 39 38 34 38 34 -4 Kerosene 28 24 27 27 28 22 -4 Gasoil 70 69 60 73 74 74 +1 Fuel oil 32 30 34 29 27 22 -3 Total 172 163 160 164 168 153 -8 Source: SLPRC, Marketing Companies; mission estimates.  - 15 - 2.5 Products obtained from processing imported crude at the small Kissy refinery do not match the requirements of the 'ocal market. To balance refinery output with market requirements, surplus products have been exported and those in deficit imported. There has been a consistent and significant surplus of fuel oil (Table 2.1) which has either been exported or sold as bunker fuel for use in ships. Other products are approximately in balance but emergency imports have been increasingly frequent since 1982 as refinery operations have been disrupted by lack of crude supplies. One effect oF the refinery's intermittent operations has been a steady increase in its fuel use and losses since 1981. The re- finery is operated by the SLPRC which was until recently owned 50% by the Government and 50% by British Petroleum (BP), Shell, Mcbil, and Texaco. In July 1985, however, the equity ownership of 10.2% held by BP was bought by the Precious Metal Marketing Corporation, a private company, in which the Government holds a 35% share. In effect, the Government holds more Lhan a 50% share in SLPRC, which places the latter in the category of a parastatal. 2.6 At the time of the mission the management and administration of storage, transportation, distribtition, and retail marketing of petroleum products in Sierra Leone was handled by four marketing companies. Three of these are subsidiaries of major oil companies--Shell, Mobil and Tex- aco--the fourth, National Petroleum (NP) is owned by the Government (60%) and private Sierra Leone interests. NP was formed from BP's marketing facilities following the closure of the latter's operations in Sierra Leone during 1985. Shell, Mobil, and Texaco each own product storage and road tanker loading facilities at Kissy. Looking at the positive side, there are no major operational, management and infrastructure problems in marketing and distribution, which are independent of the foreign exchange problem. The physical system has not deteriorated significantly in spite of limications on imported z-eres but this situation cannot be q'ustained indefinitely without a drop in operational standards. 2.7 Turning to subsector management, there is no single institution responsible for the whole petroleum subsector. The institutional respon- sibility for oil imports, refining, and marketing rests with the MTI, with the Director of Trade and Industries (DTI) being the responsible official. Exploration activities, on the other hand, fall within the jurisdiction of the MOM. 2.8 The responsibility for the setting of petroleum prices is shared jointly by the MTI and the MOF. The review and establishment of both ex-refinery and retail prices is carried out by a Price Advisory Committee (PAC) which is composed of representatives from both MTI and MOF, as well as from the oil marketing companies and the refinery.  - 16 - Short Term Issues Procurement 2.9 Procurement Arrangements. The DTI plays a lead role in negotiating terms fcr state-to-state crude procurement agreements. However, responsibility for implementation of the agreement, including the management of shipping and financia'L arrangements, was assigned to an agent--CEVIL Trading--at the time of the mission. This agent also handles the direct import of finished products to supplement the refinery's output, as well as the export of surplus fuel oil. The international affiliates of the domestic oil marketing companies are permitted to arrange certain direct imports, such as jet fuel for international bunkers and diesel oil for the large mining companies, payments for which are made in offshore dollars. 2.10 Prior to 1982, crude sales to the refinery were arranged on a rotational hasis by the oil company sha-eholders under terms defined by the Principal Agreement with the Government. The basic terms included 90-day credit, a "commercially competitive f.o.b. price...having regard to sales by major crude oil purchasing companies to non-affiliated buyers," and freight rates which reflected market conditions at the time of delivery. Compensation for crude deliveries was made by the refinery in Leones to the Bank of Sierra Leone which in turn converted them into dollars for payment to the creditors. Following exchange rate difficulties in the early 1980s, the Bank if Sierra Leone stopped converting the Leones it was receiving from the refinery into dollars. By early 1982, overdue dollar payments exceeded US$40 million and the companies refused further crude deliveries except on a dollar cash basis. No arrangement has yet been made for repayment of the debt, now amounting to UCS58 million with interest, and crude supplies from the companies remain suspended. 2.11 In the meanwhile, the Government negotiated a state-to-state contract with Nigeria for the supply of Bonny Light crude to SLPRC and arranged for a "third party," CEVIL, to handle shipment, scheduling, and financing of deliveries to FreeLown. Contractual details of this arrangement were not made available to the assessment mission. However, it is known that in February 1986, the f.o.b. price paid by Sierra Leone was equivalent to the Nigerian Government's Official Selling Price for term contracts. Effective February 18, 1986, a "net-back" pricing arrangement has been used. Credit terms under the present arrangement require a cash payment in dollars for half the cargo at the time of loading, and for the remaining half 180 days later. 2.12 The infrastructure for importing and exporting petroleum products is located at the main shipping port at Kissy, about four miles along the coast from Freetown. The Kissy jetty can accommodate tankers up to 46,000 deadweight tons (DWT). There is sufficient flexibility in lines and connections to enable the off-loading of crude oil and all  - 17 - types of products (except LPG) to storage tanks located either at the refinery or the marketing company terminals. This flexibility also allows all grades of product for exports to be loaded into tankers from refinery or marketing company tankage. 2.13 In the past, crude shipments were delivered in monthly cargoes of about 20,000 tons. The shortage of 20,000 DWT tankers in the region and lack of planning has necessitated part loading of larger vessels in the 40,000-60,000 DWT range ith attendant dead freight charges. To avoid these charges, the Government is making plans to switch to eight shipments per year, each of 30,000 tons, which the mission estimates to be more economic. Existing storage capacity, about 40,000 tons of which is assigned to crude oil and 140,000 tons to products, is adequate for handling the proposed increase in the size of shipments. Some reassignment of storage between crude and products may, however, be necessary to accommodate security stocks, as discussed below. 2.14 Supply Costs and c.i.f/f.o.b. Margin. In Table 2.2 an annual comparison has been made between the actual landed cost (c.i.f.) of the crude and two alternative f.o.b. values--the official Nigerian Government Selling Price (GSP) and average spot market values. This comparison suggests that margins were excessive both during the last two years of the deliveries under the Principal Agreement with the oil companies (1981-8') and over the period 1983-85 under the eximting contract with CEVIL Trading. Table 2.2: SIFRRA LEONE CRUDE OIL IMPORTS Average Yearly FOB CIF/FOB Differential Year Quantity Actual Landed cost GSP "SPOT" GSP "SPOT" "> (Tons) (5'000) (S/bbl) (S/bbl) (S/bbl) (S/bb') (S/bbl) 1977 188,139 21,059 15.15 14.39 14.20 0.77 0.95 1978 221,002 25,400 15.55 13.93 14.10 1.62 1.45 1979 234,601 40,640 23,25 20.87 31.66 2.38 (8.41) 1980 221,794 64,170 39.15 36.30 37.67 2.85 1.48 1981 224,029 76,136 45.93 39.20 36.66 6.73 9.49 1982 159,881 49,496 41.44 35.60 33.50 5.84 7.94 1983 181,216 48,882 35.94 30.73 29.80 5.21 6.18 1984 235,705 60,894 34.42 29.69 28.8K 4.73 5.57 1985 158,008 40,274 33.96 28.59 27.8v 5.37 6.16 Note: Type of crude: Over this entire period, virtually all the imported crude has been Niger- ian. The only significant change is in the ratio of (Bonny) light to medium crudes: 1974-78, 80% to 20%; 1979, 90% to 10%; 1980, 80% to 20%; 1981, 85% to 15%; 1982, 95% to 5%; 1983-85, 100% light. Source: Mission estimates.  - 18 - 2.15 Under the Principal Agreement with the oil companies the f.o.b. price was supposed to be a "commercially competitive" value representa- tive of third party sales. There is no exact measure of FOB third party sales prices. However, a comparison of actual landed cost with f.o.b. prices based on spot and GSP indicates the opportunity cost to Sierra Leone of importing crude under the Agreement. Table 2.2 suggests that these costs were relatively reasonable over the 1977-1980 period. The tanker market progressively tightened during these years and freight rates for the Bonny-Freetown voyage rose to $2.00/barrel. With most deliveries coming from Nigeria, c.i.f./f.o.b. differentials of about $2.50/barrel were not out of line with actual costs during that period. During 1981 and 1982, however, the landed cost per barrel was US$4-7 higher than the f.o.b. opportunity costs. This excessive margin cannot be attributed to either high freight rates or a tight oil market: while still relatively high, freight rates were not significantly above 1980 levels and spot prices had fallen below GSP since crude was in plentiful supply by then. 2.16 The margin for the crude landed under the state-to-state arrangement, 1983 through 1985, also appears excessive. The average c.i.f./f.o.b. differential over the period was $5.06/barrel. An "iefficient" margin, including freight, insurance, crude losses, working capital costs, and a 3% commission for handling charges to CEVIL, would have been at most $2.60/barrel. This implies that Sierra Leone paid an "'excess" margin of at least $2.40/barrel, amounting to approximately $10 million over the three years. 2.17 Weaknesses of Current Procurement Arrangements and Possible Options. Despite the excessive margins paid for crude supplies over recent years, deliveries have not always arrived when needed. Foreign exchange shortages, together with the absence of a budgeted allocation of funds for 'petroleum imports, have resulted in a scramble to find sufficient foreign exchange for each cargo. This has apparently involved the Ministry of Finance itself in parallel market dealings from time to time. Delays in crude arrivals have led to severe shortages of products and undoubtedly escalated landed costs as vessels originally scheduled for loading in Nigeria have had to be cancelled on occasion. 2.18 Most of the problems associated with procurement arise from foreign exchange difficulties. Because of its inability to pay off the large outstanding debt of the oil companies, the Government's ability to obtain more competitive terms for oil supplies is compromised: the presence of the debt limits the number of international crude suppliers which the Government can approach for a supply contract. At the same time there is room for improvement in the way the Government is coordinating the activity. First, there is no system to allocate foreign exchange for purchasing oil supplies on an annual basis and then properly programming this allocation to each *-hipment. Second, the use of an open-ended contract at fixed terms with a single party, whether it is CEVIL Trading or an oil company, is non-competitive and invites monopoly rents.  - 19 - 2.19 Options and Recommendations for Changing Procurement Arrangements. There are several options open to the Government for ( repaying oil companies for past deliveries of crude; (b) developing a system for programming available foreign exchange to match the timing of pre-arranged shipments; and (c) improving the efficiency of procurement arrangements. A detailed study of these options cannot be undertaken here. However, some comments can be made on each. 2.20 For repaying the oil companies, there are a number of options ranging from repudiating the debt to immediate repayment. Somewhere in between lies the possibility of making arrangements to repay the debt over a long term horizon at concessionary interest rates, possibly in the form of a premium on the prices of future oil supplies. The Government should attempt to resolve this problem as soon as possible through a negotiated rescheduling of debt repayments. 2.21 The appropriate system for programming foreign exchange requirements to purchase crude depends on the state of the foreign exchange market. A totally free floating market, if it emerges, will not require any allocation system at all. However, if the market does not clear (i.e., shortages remain), or if the market is finely balanced such that the large payments (US$4-5 million) required for oil purchases cannot be easily accommodated, then the Government will have to find a way to allocate the funds required to purchase shipments of 20,000 tons about once a month (or two shipments of 30,000 tons every quarter). The Government should budget the requirement for funds within the constraint of foreign exchange availability and give suitable priority to petrolE n imports. An alternative approach would be to arrange a revolving credit facility with a major commercial institution to support oil imports. The trade-off would be in the form of costs for arranging the facility. 2.22 Finally, there are a number of options open for improving the efficiency of procurement arrangements. The feasibility of these options depends on the progress made in settling the oil company debt issue and the future mode of refinery operations. Assuming that the Government shows its intention to solve the debt problem, and the present supply mode is continued, 5/ the main options are to: (a) maintain the current state-to-state agreement with Nigeria and implement the agreement with an agent like CEVIL, but review the formulae for the determining the CIF/FOB margin; (b) invite international bids for crude supplies CIF Freetown; and (c) maintain the state-to-state agreement for Nigerian Bonny crude but invite international bids for managing procurement arrangements. The mission recommends that the Government adopt the last option because: (i) Bonny crude is a good match with the country's market in terms of both quality and product yields (discussed in the refinery section); (b) its close proximity reduces freight charges; and (c) procedures for 5/ Crude procured for local refining with imports/exports of deficit / surplus products.  - 20 - regularly arranging and deciding on international bids would tax the capacity of current institutional structure--delays in making decisions could more than offset the advantages of such a procedure. It is diffi- cult to estimate the savings from each of these options. However, the analysis above suggests that the CIF/FOB margin could comfortably be reduced from $5/barrel to about $2/barrel by introducing more competition and negotiating deliveries in fully laden 30,00n ton DWT vessels. A savings of $3/barrel would lead to annual savings of more than $5 million for crude imports of 236,000 tons at 1984 levels. Demand Management and Pricing 2.23 In conjunction with measures to improve the supply of petroleum products, the problem of severe shortages should be addressed through better demand management, primarily through pricing policies. The prices of petroleum products in Sierra Leone are well below their opportunity cost to the economy, and the Government lacks a cohesive demand manage- ment policy. Both the level and structure of prices need to be reviewed immediately to eliminate pricing distortions and reduce the subsidies that magnify the financial troubles of Government and subsector organizations such as the refinery. Beyond this pricing review, the Government should reconsider the objectives and policy underlying the establishment of ex-refinery and retail prices, particularly at this time of uncertainty in international. oil markets. 2.24 Price Trends. A review of the time series of average annual retail prices of petroleum products between 1975 and 1985 (Annex 3) reveals that prices in real terms generally remained stable over the late 1970s, but have declined significantly since 1980. In 1985, and aE.- the time of the mission, real prices had fallen to about half of their 1975 levels. To make a meaningful comparison of prices with neighboring countries is difficult because most West African countries have a distorted price/exchange rate pattern. Senegal is an exception in this regard, and Table 2.3 shows that pump prices in Sierra Leone were well below those in Senegal during 1985, both at official and parallel ex- change rates. Table 2.3: COMPARISON OF OFFICIAL RETAIL PRICES IN 1985 (USS/US Gallon) Sierra Leone Senegal at 5.4 Le/S at 10 Le/S (official rate) (parallel rate) Gasoline 1.23 0.67 2.93 Kerosene 0.62 0.33 1.99 Gasoil 1.08 0.59 1.89 Source: Mission estimates.  - 21 - 2.25 Distortions in Ex-Refinery Prices. The official price structure existing at the time of the mission is outlined in Table 2.4. The single most important influence in the price structure is exerted by ex-refinery prices. These are based on the landed cost of crude in Leones, anticipated operating costs, taxes, interest payments, and profits for the ensuing twelve months. These costs are distributed on an Imperial Gallon (IC) basis over the total expected volume output of the refiner for the same period. Product price differentiation is introduced to take account of political and commercial considerations, but with the proviso that the weighted average price structure should allow a complete recovery of the refinery's costs. Table 2.4: S!ERRA LEONE OFFICIAL PETROLEUM PRODUCT PRICES AND ECONOMIC COSTS OF SUPPLY, 1985 Gasoline Kerosene Gasoil Fuel Oil Fuel oil to NPA (Le/IG) (Le/IG) (Le/IG) (Le/IG) (Le/Long ton) (1) Ex-Refinery Price 6,65 3.22 5.90 3.00 720.00 (2) Duty 0.74 0,27 0.63 - - (3) Total Mktg Co Margin 0.46 0.36 0.32 0.34 81.29 (4) Fr-eetown Area Transport a/ 0.05 0.05 0.05 0.04 8.71 (5) Dealer Commission 0.10 0,10 0.10 - - (6) Sel ing Price, Freetown b/ 8.00 4.00 7.00 3.38 810.0 (7) Economic cost of supply in Freetown c/ 9.81 10,74 10,23 8.16 1860.0 a/ The transportation charge covers deliveries to service stations within a 20 mile radius; for trucking costs from main Kissy depot to up country locations there is an additional levy equivalent to approximately Le 0.20/100 miles. In the case of fuel oil tf;e charge is based on jetty and barge costs from Kissy to Kingtom station. b/ Effective April 15, 1985. c/ Estimated using the CIF costs of direct product imports (see Table 2.5) based on 3 shadow exchange rate of Le IO/Si. To the c,i,f. costs has been added a terminalling charge of $3/tonne and margins for marketing, transportation and distribution identical to lines x, 4, and 5 in the table. Source: SLPRC, Marketing Companies; mission estimates. 2.26 Table 2.5 compares official ex-refinery prices ill 1985 with the hypothetical cost of direct imports for the same year. Average West European spot prices are used as a starting point, to which are added freight and related charges. As shown, the structure of import parity prices was different from that of official ex-refinery prices: both kerosene and fuel oil were being sold considerably below their oppor- tunity cost and, in effect, being subsidized by gasoline, whose price was about 20% above import parity. At the same time, ex-refinery kerosene prices were well below those for gasoil, as compared to import parity prices where the opposite situation existed. The need for cross subsi-  - 22 - dies should be carefully reviewed by the Government. Further, the differential between kerosene and gasoil should be eliminated; this is essential to prevent adulteration of gasoil by kerosene, especially in road transport uses. The price structure is inefficient for two reasons. Firstly, the cost-plus pricing arrangements means there is no incentive for the refinery to improve efficiency. Secondly, to the extent the refinery considers these product prices as given, its produc- tion decisions are based on artificial differentii.ls which are not consistent with the country's opportunity costs. 2.27 The comparison in Table 2.5, based on official exchange rates, masks another important distortion: ex-refinery prices in Leones were well below the economic cost to the country. It is estimated that the shadow exchange rate in 1985 was 10 Le/$. If import parity prices are converted to Leones at that rate, ex-refinery products will be about half their economic costs. This distortion, which is due to an over valued exchange rate, represented an explicit subsidy during 1985 since most purchases of crude in that year were financed by the Government through parallel markec purchases of the dollar. Given that net imports in 1985 were equivalent to $33 million, the subsidy paid by the Government to oil consumers was approximately $10-15 million. The effect of subsidizing ex-refinery prices is also reflected in retail prices. As shown in the last line in Table 2.4, the economic costs of supply were well above retail. prices existing in 1985. These distortions have continued in spite of recently introduced changes in product prices as discussed below. TatIe 2. 5: COMPARISON OF OFFICIAL 1985 EX-REFINERY PRICES WITH HYPOTHETICAL COSTS OF DIRECT IMPORTS Gasoline Kerosene Gaso'I Fuel oil Ex-Refinery Price Le/IG 6.65 3.22 5.90 3.00 at official exchange rate, a! S/ton 376 164 284 126 Hypothetical 1985 13nded Cost NW Europe SPOT 266 265 239 152 Freight *ins,loss, S/ton 12 12 12 12 Imp Duty*Turntax 10.0%, S/ton 28 28 25 6 Terminalling, S/ton 3 3 3 3 rotal Landed Cost of Imports 309 308 279 183 Ex-refinery as a % import price at official exchange rate 121% 53% 102% 69% al Average official exchange rate during 1985 of 5.4 Le/dollar. 2.28 Tax Relativities. An important element of retail prices is the tax or government "duty" on petroleum products. The total "take" under  - 23 - the existing structure amounts to an average of 22% of c.i.f values on all light products (gasoline, kerosene, and gasoil). This breaks down as follows: 8% import duty (on crude), 2% turnover tax on the refinery, 2% turnover tax on marketing, and a 10% duty in the price structure. The more or less uniform 22% tax on light products is on the low side when compared with values for several other West African countries (Ivory Coast, Senegal, Togo, Burkina, and Benin) as shown below: Total Tax as a Percentage of CIF Low Average High Gasoline 40 80 120 Kerosene 0 20 40 Gasoil 20 40 70 It is a general practice not to tax fuel oil (or gas oil) sold for power generation. 2.29 Revised Structure of Prices. An alternative structure of prices is proposed in Table 2.6 based on a range of crude oil prices varying from $10 to $20 per barrel and an exchange rate of 48 Le/$, the average for March/April 1987. Ex-refinery prices are set by import prices (c.i.f.), which are derived from the two crude prices using the assumption that the product prices will allow a world scale refinery to be economic. In essence, these prices represent LMRC to the country for two different prices of crude (See Annex 4). In practice spot prices for products will be different from these long run values but they still should fall within this range, as was the case in the first quarter of 1986; at the time, c.i.f. prices per ton were, based on spot, $185 for gasoline, $228 for kerosene, $200 for gasoil, and $106 for fuel oil. The table suggests that large increases are required in the domestic prices of all products. 6/ The table should be considered as indicative of the options for changes in product prices. 6/ Despite recent increases, official selling prices of petroleum products in Freetown during April/May 1987 (kerosene, 45 Le/IC; diesel, 47 Le/IG; and gasoline, 55 Le/IG) were still well below the upper end of the range estimated in Table 2.6; the upper end of the range is appropriate for comparison as crude oil prices were touching $20/barrel in this period.  - 24 - Table 2.6: REVISED PETROLEUM PROOUCT PRICE STRUCTURE World crude: S20,00/bbl Exchange rate: 48.00 Le/S Gasoline Kerosene Ga3oil Fuel Oil to NPA (Le/IG) (Le/IG) (Le/IG) (Le/IG)(Le/long ton) CIF Freetown (S/ton) S/ton 233 236 215 119 Le/tonne 11184 11328 10320 5712 Terminal Throughout Charge Le/tonne 144 144 144 114 Total Import Parity Le/tonne 11328 11472 10464 5856 Le/IG 38.4 42.5 41.2 24.8 ,Ex-Refinery Price 38.4 42.5 41.2 24.8 ;855 Duty (as in footnote) 31.8 8.5 16.5 - - Total Mktg Co Margin (6%) 2.3 2.6 2.5 1.5 360 Freetown Area Transport 0.8 0.8 0.8 0.3 20 Dealer Commission 1.5 1.5 1.5 - - Selling Price, Freetown 74.8 55.9 62.5 25.6 6235 Sensitivity at S10/bbl CIF Freetown (S/ton) S/ton 134 135 124 66 Selling Price, Freetown 44 33 37 15 3586 Note: Tax structure assumed is an average by West African standards: Gasoline, 80%; Kerosene, 20%; Gasoil, 40%; Fuel oil, 0%. Transportation and marketing margins set at levels which reflect economic costs and oil industry practice in LDCs. Source: Mission estimates. 2.30 Pricing Policy and its Objectives. If fiscal revenue objectives of pricing are set aside, the nurpose of government policy should be to ensure that prices signal the economic cost of supplies to the consumer. However, this is not simple to implement in practice, as a number of policy options are open to the Government in choosing prices to reflect economic costs. 2.31 First, at e time of uncertainty in international oil markets and exchange rates, a choice has to be made between allowing prices to reflect the recent decline in oil prices, or raising them to a level consistent with the estimated long run cost. To sustain pricing reform will require the resolution of short term and long term pricing trade- offs. This is a complex issue which cannot be fully addressed here. Nevertheless, a balance must be reached between the advantages of price stability and sustained conservation efforts resulting from pricing based on long run costs, and the economic benefits of passing on lower short run prices to industry and final consumers of energy. At the moment domestic prices are lower than they should be to reflect the recent changes in oil prices and exchange rates, or short run marginal costs.  - 25 - Therefore, immediate priority should be given to bringing these prices at least into line with current opportunity costs. 2.32 Second, the issue of pricing cannot be examined independent of the refinery viability issue. If the refinery is assumed to be viable in the long run and the Covernment strategy is to retain it as an operating entity, then a pricing policy where ex-refinery prices are determined by import parity (-ather than cost plus pricing) could lead to large fluctuations in the financial position of the refinery. Large financial losses sustained over a number of years, if spot prices remained depressed over that period, could bankrupt the refinery, unless supported by tihe state. Hence, in determining domestic pricing policy, explicit recognition needs to be given to product supply strategy. 2.33 Pricing Policy Options. The above analysis suggests that there are ambiguities and weaknesses, both in terms of pricing policy objectives and the actual level and structure of prices. In the mission's view, the Covernment has three main options for strengthening policy, as defined below: (a) Continue with the present approach in which ex-refinery prices are based on a cost plus pricing scheme (allowing the refinery's costs to be covered). (b) Link ex-refinery prices to import parity with regular reviews and let the refinery operate as any other commercial entity. In such a scheme, the refinery could go bankrupt or operate only when viable. In this option, prices would reflect short run economic costs. (c) Link ex-refinery prices to import parity, but use a stabiliza- tion fund to minimize the fluctuation in ex-refinery prices due to changing costs. A mean stabilized price would initially be established based on an assessment of long run costs. The refiner or product supplier would then deposit to or withdraw from the fund when import prices were below or above the stabi- lized ex-refinery price. Only when the stabilization fund hit a predetermined floor or ceiling would a review of the stabi- lized price be required. This scheme attempts to stabilize the ex-refinery prices consistent with LRMC. 2.34 After the issue of policy objectives has been settled, options for adjusting the level and structure of prices can be reviewed to establish prices whicn signal appropriate economic costs. 2.35 Recommendations. The mission recommends that an independent study be carried out as soon as possible to examine pricing policy and structure. The study should make recommendations on pricing policy and the approach for establishing and reviewing prices, both at the retail and ex-refinery level, taking into account the effects of continuously varying exchange rates and oil prices. This study should be carried out  - 26 - in conjunction with the establishment of a petroleum product supply strategy and not independent of it. In the interim period, the Government should give priority to changing ex-refinery and retail prices so that they, at least, reflect opportunity costs to the country. The Role of the Government in Managing Shortages and Crises 2.36 There are major weaknesses in the way the Covernment has been managing the shortages of petroleum products. These lie in the areas of procuring crude, establishing prices, and managing the crisis. There is an urgent need to strengthen the Government's monitorir.g and regulatory functions in these areas. 2.37 Crude Procurement Arrangements. The DTI is limited in its ability to coordinate crude and product procurement arrangements by a variety of factors. These include: (a) lack of international oil market intelligence; (b) absence of a system for monitoring supply-demand balances in the local market which would warn of impending shortages; (c) absence of regular inter-ministerial coordination with the MOF for scheduling foreign exchange to cover anticipated deliveries; and (d) political pressure for using a single implementation agency, such as CEVIL, to implement procurement arrangements and an apparent absence of a mardate to review and question orocurement costs. 2.38 Procedures for Establishing Official Prices. The procedures for establishing prices are cumbersome and time consuming, and are not flexible enough to respond to current conditions in oil markets. The result is that prices are inconsistent with costs and give incorrect signals to consumers. More specifically, there are no provisions for a regular review of prices or a pre-established formula for adjusting prices. Instead, requests for changes in ex-refinery prices or marketing margins have to pass through various committees. In the case of the ex- refinery prices this includes the SLPRC Board, the PAC and, ultimately, the Ministerial Cabinet. Delays are inevitable. Simplifying this procedure or deregulating prices (at least partially) are two of the options for making prices more responsive to market conditions; the present decline in prices offers a good opportunity to implement changes. 2.39 Managing Crises. Irrespective of the DTI's best efforts, shortages of products will occur from time to time. Under conditions of severe shortages, the market will not be able to balance demand and supply at a politically acceptable price level. Managing demand under these conditions may require quantitative restrictions and rationing. The Government should have a contingency plan to determine allocation  - 27 - priorities during such an emergency. This would help to allocate supplies smoothly on a pre-arranged basis and avoid the chaos caused by shortages. The use of ration cards, or priority schedules worked out before hand with market ng companies are two options which the Government should examine. The Government should also maintain a minimum level of crude oil in storage. One month's security storage should be more than sufficient in a situation where the voyage time to the normal source of supply (Nigeria) is only four to five days. A minimum stock level of about 20,000 tons (equivalent to approximately 30 days of consumption) should therefore be maintained. To accommodate both the security storage and the proposed larger cargo sizes of 30,000 tons, the amount of tankage assigned to crude would need to be raised from 40,000 to 50,000 tons. This can be done through reallocation of existing tanker serv,ces and without additional investment. 2.40 Distribution of Products in Rural Areas. Facilities for distributing petroleum products in rural areas are grossly inadequate. The distribution companies are reluctant to extend such facilities due to the small volume of trade and the h.gher cost and difficulties of trans- portation. This position is unlikely to change in the foreseeable future. Local traders, if at all interested, tend to expect considerably higher prices. In times of crisis, as was the case itn 1985, product prices in rural areas climb to levels three to four times higher than in Freetown and often supplies are not available at all. Shortages of diesel are, under normal circumstances, serious during planting and harvest time. 2.41 The Government should attach priority to extending a minimum level of marketing facilities to agricultural areas. One alternative to be examined is to set up a number of small rural outlets to market kerosene and diesel. These outlets could be operated by farmers' cooperatives. While the return from these outlets is unlikely to be commercially attractive, their wider impact on agricultural output and the welfare of rural poor could be significant. 2.42 Recommendations. The ability of the DTI (or of the more appropriate section within the MEP, which this report recommends as the more appropriate institutional body for regulating activities in the oil subsector, see Chapter V) to monitor and regulate the market should be strengthened by: (a) creating a statistics and planning cell which would include the monitorinig of international oil markets as one of its functions; (b) providing training to staff assigned to planning cell;  - 28 - (c) developing an information system to compile, integrate, and analyze data on exports, imports, refinery output, and sales; (d) more clearly defining organizational objectives, accounta- bility, and staffing needs; (e) developing an emergency plan for allocating available products during severe shortages in collaboration with the marketing companies; (f) simplifying procedures for establishing prices and making them more responsive to market conditions; (g) evaluating the viability of establishing retail outlets in rural areas to be run by farmers' cooperatives; and (h) raising the level of security supplies for crude oil to 30 days of supply. Long-Term Issues 2.43 The long term issues pertain to the economics of the refinery, the strategy for exploration promotion, and subsector organization. As a background to a discussion of these issues, the future demand for petroleum products are outlined in the form of two scenarios. Future Demand for Petroleum Products 2.44 The long term issues which the Government must address in the petroleum subsector should be evaluated against the future demand fer petroleum products. Demand projections have been prepared both for inland consumption and international bunkers. Two scenarios, "Acceler- ated Growth" (AG) and "Base Case" (BC) were developed for both demand categories and the combined results are summnarized below in Tables 2.7 and 2.8. The key aspects of the sceaarios are that: (a) suppressed demand for kerosene and gasoline is assumed to be diminishing over the next few years; (b) the higher growth of demand under AG is driven by rising income and not by the increase or introduction of energy intensive activities--this is largely due to the assumption that recovery is based on agricultural growth rather than industrialization; and (c) a reduced form of the Bumbuna Hydroelectric Scheme is introduced. Under these assumptions, demand grows by about 1% under BC and 4.1Z under the high growth scenario. Because of the emphasis in agriculture, the product mix or "cut of the barrel," does not differ much between the scenarios. The use of thermal power plants, instead of Bumbuna, will generate more fuel oil and gas oil consumption, as is shown in the same tables. In 1995, the introduction of Bumbuna would reduce the level of oil imports by approximately $8.5 million in the AG scenario.  - 29 - Table 2.7: FORECAST SUMMARY OF INLAND AND BUNKER DEMAND ACCELERATED GROWTH SCENARIO WITH REDUCED BUMBUNA ('000 Tons) AG-Thermal a/ 1985 1986 1987 1988 1989 1990 1995 2000 1995 2000 Gasoline 34 47 46 48 51 53 62 79 62 79 Jet/Kerosene 37 40 40 42 45 46 54 78 54 78 Gas oil 75 82 78 78 82 88 111 142 113 142 Fuel oil 37 44 49 52 55 60 39 34 97 95 Total 183 213 214 220 233 247 266 333 326 394 Total Market Product Mix (%) Gasoline 18.4 21.9 21.7 21.9 21.9 21.6 23.3 23.8 19 20.1 JetAKerosene 20.1 18.9 18.8 19.1 19.3 18.5 20.3 23.6 16.5 19.8 Gas oil 41.1 38.6 36.4 35.3 35.1 35.5 41.7 42.5 34.8 36.0 Fuel oil 20.4 20.6 23.1 23.7 23.7 24.4 14.7 10.1 29.7 24.1 Total 100.0 100.0 100.0% 100.0 100.0 100.0 100.0 100.0 100 100 a/ Petroleum demand with a pure thermal scheme instead of Bumbuna. Source: Mission estimates. Table 2.8: FORECAST SUMWARY OF INLAND AND BUNKER DEMAND BASE CASE BASE CASE WITH REDUCED BUMBUNA ('000 tons) Actual BC-Thermal 1985 1986 1987 1988 1989 '990 1995 2000 2000 a/ Gasoline 34 48 46 47 46 46 47 53 55 Jet/Kerosene 37 41 '9 39 40 39 41 53 53 Gas oil 75 79 74 72 72 73 82 83 93 Fuel oil 37 44 46 48 48 49 59 26 76 Total 183 2T2 205 206 206 207 229 215 277 Total Market Product Mix (%) Gasoline 18.4% 22.3 22.3 23.3 22.4 22.2 20.7 25.2 20.0 Jet/Kerosene 20.1% 19,2 19.2 19.3 19.5 18.7 17.9 24.3 19.3 Gas oil 41.1% 37.7 35.9 35.2 34.9 35.4 35.7 38.2 33.3 Fuel oil 20.4% 20.8 22.5 23.2 23.2 23.7 25.6 12.3 27.4 Total 1OO.OS 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 a/ Petroleum demand with a pure thermal scheme instead of Bumbuna. Source: Mission estimates.  - 30 - The Viability of the Refinery 2.45 Although it is suffering from crude shortages, attendant shut- down/start-up of operations, and is technically bankrupt from assuming the cumulative interest on the Government's crude debt to the oil companies, the refinery appears to be operationally and managerially sound. A key long term issue for the petroleum subsector is the viability of the refinery and how it fits into the least cost strategy for supplying petroleum products. 2.46 The oil refinery is located at Kissy, and is owned and operated by the SLPRC. SLPRC in turn is owned by the Government of Sierra Leone, private Sierra Leone interests, and three major international oil companies in the proportions given in Table 2.9. Taole 2.9: REFINERY OWNERSHIP Shareholders Share (%) Government of Sierra Leone 50.0 Precious Metals Marketing Corporation (PMMC) a/ 10.2 Shell 17.7 Texaco 10.8 Mobil 11.3 Total 100.0 a! The Government holds a 35% interest in PMMC. SLPRC is controlled by a board of directors consisting of ten shareholder representatives. The Minister of Finance serves as ex-officio Chairman of the Board. The refinery has a tot3l staff of 134 nationals. 2.47 Current Operations. The refinery was commissioned in 1969 with a design capacity of 10,000 barrels/day of light Iranian crude. It is a 3imple crude distillation unit with LPG recovery facilities and asso- ciated equipment. It has been running in recent years on a diet of predominantly Nigerian light crude with 10% to 20% medium crude blended in occasionally to meet product yield requirements. There is no gasoline octane enhancement process such as a reformer, and until 1982, about 12,000 tons of reformate, or high octane blending component (HOBC), were imported annually and blended in with straight run gasoline to meet premium grade octane specifications (93 RON). Because of foreign ex- change difficulties, reformate imports have been discontinued and only one grade of straight run gasoline is produced with maximum lead, 3.0 grams/gallons, blended in to achieve a RON rating of 85-87. The re- sulting reduction in performance, which is being tolerated by the public, has required engine and spark plug adjustments.  - 31 - 2.48 Operational Efficiency and Costs. The plant appears to be a neat, well configured facility in excellent condition. The staff, all nationals, are trained refinery professionals and have been following good maintenance and safety practices. They also have proper laboratory facilities for quality control and testing. Specialized professional assistance is obtained at reasonable cost from BP under a technical service agreement for annual maintenance and procurement of some materials and equipment. At the current level of operations, there are no major equipment requirements or deficiencies. 2.49 The operational statistics bear out the positive impression of the plant. Fuel and loss figures have been kept low (3.5%), at least until recently. Between 1982 and 1985 this figure has crept upwards, reaching 5.1% in 1985. The increase is mainly caused by intermittent operations due to lack of crude, and greater use of self-generated power, hence higher fuel use. More reliable supplies of crude and electricity from the NPA should restore fuel and loss figures to their former 3.5% level. With modest investments in energy conservation, the loss figure could be reduced to 2.5%. Operational costs are also modest compared with other small refineries. In 1984, total costs were $2.7 million, or $1.63/barrel, converting Leone costs at the official exchange rate. For 1985, the estimated figure is $1.58/barrel. Even these modest figures are an overestimate because of an over valued exchange rate and restric- tions on throughput due to import restrictions. Assuming a rate of 12 Leone!$ and normal throughput of 240,000 tons, operational costs for 1985/86 are estimated at $0.93/barrel or $7/ton. 2.50 Economics of Refinery Operations. An evaluation of the refinery's economics in Annex 5 suggests that The refinery should be able to run at break even or even incur a moderate profit if products from the refinery are priced at import parity. 2.51 The results show that even for the 1983-85 period--a difficult one for world refining--the Freetown refinery would have almost broken even on gross margin and its economic loss would have been about US$2.1 million per year. In the first quarter of 1986, the refinery had an economic surplus and it is projected to maintain one through 1986 and in the year 1990 if prices remain at their trend levels. In addition to whatever sutplus the refinery generates, it provides other benefi s in the form of a center for technical and managerial excellence, a hydrocarbons testing laboratory, and employment and training spin-offs. 2.52 The main factors which support the refineries current viability are: (a) low operational costs due to relatively efficient operations and the absence of (expensive) expatriate staff; (b) freight advantage of transporting crude relative to products due to Sierra Leone's location-- the cost of transporting crude from Nigeria (Bonny) to Freetown is about half the cost of transporting products from Rotterdam, the base location for determining product prices, to Freetown; and (c) sunk capital costs of existing facilities and the relatively modest needs for future invest- ment over the next 10-15 years. The refinery's long term viability is  - 32 - therefore assured as long as it is run efficiently, no major changes in refinery configuration are contemplated, and Sierra Leone continues to obtain crude at competitive terms from Nigeria. As the refinery's economics are finely balanced, a significant change in any one of the above three variables, while unlikely, could make the refinery an uneconomic operation and the Government should be aware of this sensitivity. 2.53 Short and Long Run Viability. As discussed in the pricing section, even if the refinery is viable in the long run, there may be prolonged periods when, due to product dumping, for example, spot prices remain depressed and create large financial losses in the refinery. Under these circumstances, the refinery's operations should be closed down or a stabilization fund put into effect. Regardless of whether the refinery is run as a purely commercial outfit or as a state enterprise, economic rationale suggests that its operations should be halted if the gross margin is negative for a prolonged period. In other words, if variable costs are not being covered, it would make sense to stop operations, without firing the staff, until prices return to more favorable levels. During this period, the refinery could operate as a receiving terminal for products. By following this approach, the maximum loss sustained could be limited to the level of fixed costs ($1.4 million in 1985/86). For such a strategy to work in practice, the gross margin would have to be carefully monitored. This approach might be more likely to work if the refinery were operated on a purely commercial basis. The risks to a commercial venture are that, even if the gross margin were positive, it might not be high encugh to cover fixed costs, resulting in some losses. However, a commercial organization might be able to sustain this level of losses for a couple of years with some help from the Government, if warranted. For this strategy to work in practice, ex- refinery prices would need to be linked to spot prices. This would create an incentive for efficiency at the refinery that would be missing in the regime based on a stabilization fund or the current cost plus pricing scheme. 2.54 EffecL of Bumbuna. According to Tables 2.7 and 2.8, fuel oil requirements increase until the start-up of Bumbuna in 1993 (under the Accelerated Growth Scenario), or 1997 (under the Base Case Scenario). Until this point in time, the economics of the refinery will improve as relatively less fuel will be exported at prices below import parity. In either scenario, the effect of Bumbuna coming on-stream would be to divert between 30,000 and 40,000 tons of fuel oil to export markets. With the estimated penalty at about $15/ton, the refinery's profits would decline by about $0.5-0.6 million when Bumbuna would start up. 2.55 Financial Situation. CLPRC has a policy of showing interest accrued on the $58 million debt to the oil companies on its profit and loss account. Re-evaluation of the debt (in Leones) after a currency de- valuation adds to SLPRC's balance sheet liabilities and so far has rendered it technically bankrupt; net worth in 1984 was a negative Le 17.9 million. While the Government makes arrangements to repay the  - 33 - debt, the legal basis for carrying it on the refinery's books should be reviewed. In addition to the continued inclusion of the Government debt to the oil companies in the refinery's accounts, the financial perfor- mance of the refinery has been adversely affected in the recent past by: (a) increased fuel consumption and losses associated with frequent start up and shut down of opevations due to crude supply interruptions, and (b) delays in adjusting ex-refinery prices following increases in the Leone price of crude supplies. Both these problems have already been discussed above under short term issues together with suggested solutions. Their early resolution--which should result in improvements in crude supplies and in the procedures for establishing prices--would significantly improve the refinery's finances. Recommendations (a) The refinery should be retained as an operating entity. However, a least cost strategy should be implemented, requiring that the refinery be utilized only when crude and product prices allow it to operate with a positive gross margin. When this is not possible, products should be imported directly. (b) As the economics of the refinery are finely balanced, the importation and use of Bonny Light Crude is strongly recommended to keep freight costs at a minimum. The use of other crudes could well make refining uneconomic. (c) To permit efficient monitoring of the refinery's financial condition, ex-refinery prices should be equated to import parity and the refinery should be allowed to operate as a commerc.al entity. (d) Any financial support which the Government provides to the refinery should be limited aLd non-recurring, and only granted under exceptional circumstances. (e) Investments should be made in energy conservation measures to reduce fuel use and loss to 2.5% of crude through-put. (f) The validity of charging accrued interest on oil company debt to the refinery's accounts should be reviewed. Oil Exploration 2.56 As already indicated, exploration activities fall within the jurisdiction of the MOM. Negotiations on petroleum licensing agreements are coordinated by the Director of Mines through a technical committee which includes representatives from the MOF, the Bank of Sierra Leone, and the Customs Service. The Geological Survey, which is part of the Directorate of Mines, is in charge of monitoring ongoing exploration act iVi ties.  - 34 - 2.57 This organization of exploration activities does not appear adequate enough to exploit the opportunities available. There is no formal organization for negotiating with the oil companies; the decisions to conclude agreements depend on the office of the Head of State. There is no legal framework for exploration contracts. The country appears ill-equipped to carry out the sort of independent seismic investigations and reassessments of past survey data that are necessary to actively promote exploration activities. To address this last problem there is a 1 million ECU project by the EEC designed to strengthen the staff and laboratory facilities of the Geological Survey. 2.58 So far there are no proven oil or gas resources in Sierra Leone. The Government is interested in evaluating all areas of the country which may have any potential for hydrocarbons. Geologically, only about 28,000 square miles of the country's offshore areas are of interest. These can be divided into two main areas: (a) the North West stable shelf which is geologically related to Guinea, and (b) the South Eastern shelf, related to Liberia (see Annex 6 for map). Most of the exploration activity in the last six years has been concentrated in the southern area, involving about 7,000 km of seismic and two wells, as discussed below. 2.59 Exploration efforts to date have been undertaken exclusively by private foreign oil companies operating under license. Little exploration activity was carried out before 1979 when the Oxoco-Aracca- Sundance group signed a five-year exploration agreement with the Government. The Oxoco group "farmed out" the area to Mobil, which drilled the first well in Sierra Leone in 1982 and later abandoned it as a dry well. The area was subsequently "farmed out" to Amoco, whose first well, drilled in late 1985, also turned out to be dry. Amoco is currently ccncentrating its efforts in the southeastern area and has renewed its earlier agreement by signing a six-year contract in May of 1986. This contract covers exploration over the entire 28,000 square miles of Sierra Leone's offshore areas. The agreement is divided into three two-year periods requiring Amoco to give up a third of the conces- sion in May 1988, a further third in May 1990, and the remainder at the expiration of the contract in 1992. Amoco's interest in this area is due to its activities in neighboring Liberia, where its investigations have revealed good petroleum prospects beyond the continental shelf. Geolo- gically the same prospects may exist beyond the Southeastern Shelf in Sierra Leone, and Amoco appears to be keeping its options open while it is drilling its fourth well in Liberia. 2.60 Over the next few years exploration activities may increase in Liberia and Guinea and the results there will influence the evaluation of prospects in Sierra Leone. Clearly the 28,000 square miles of offshore areas are under-explored with only two wells drilled to date. The existing seismic and geological data show that there is a petroleum potential though the geology suggests that possible oil "traps" are difficult and will increase the risks for exploration activity. However, comparing the offshore with similar basins, a fair evaluation of the  - 35 - basin would require more wells to be drilled. Once Amoco relinquishes its offshore concessions in Sierra Leone, the Government should promote exploration, attract more companies to the country, and open up the area to competitive bidding. The Government should be prepared to exploit these opportunities. Due to the close geological relationship with Liberia and Guinea, it is also advisable that interpretation work and promotion activities be coordinated betweeii the three countries under a regional project. These effo:ts need to be strengthened, particularly at this time of decline in global exploration activity. While the idea of a regional project has been floated before, without coming to fruition, there is a case for trying again to establish such a project. 2.61 Recommendations. The Government should d&aw up a promotion strategy to be implemented over the next 5-6 years to attract more oil companies for exploration. This strategy should include: (a) additior-1 equipment and assistance for improving the storage of technical data, including sample cores; (b) interpretation of existing data, including any generated by Amoco, to provide the Government with an independent evaluation of the petroleum potential; (c) review of petroleum legislation and Model Contract; (d) technical collaboration with Guinea and Liberia for a regional evaluation and promotion so as to provide a greater incentive to oi companies; and (e) the development of a training program for the Directorate of Mines and the Geological Survey. 2.62 The above program should be prepared and planned in detail by a preliminary mission comprising: (a) an experienced petroleum explorationist (about two months) (b) an expert in petroleum laboratory and storage of information and data (about one month). The cost of this preliminary mission is evaluated at about $30,000. The CosL of the total project over a three years period is estimated at 0.5- $1.5 million, depending on the amount of additional seismic data acqui- sition which certainly will be needed in the Northern part of the basin. It is highly advisable that parts of this program (specially equipment, offices, and training) be coordinated with the EEC project. This would save time and money.  - 36 - Subsector Organization 2.63 An important role for the Government is to ensure the right balance between private and public sector activities and improve the efficiency of public sector institutions. In this regard, one issue that the Government has to address in the longer term is the role of NP, in which the state holds a majority share, 60%. NP, which took over BP's marketing facilities during 1985, now has around a 40% share of the market, with the remainder shared about equally between Shell, Texaco, and Mobil. NP's share may already be considered excessively high from the point of view of a competitive market. There are indications that NP intends to buy out Texaco. A take over would increase NP's share of the market to 60%, giving it a dominant position and significantly reducing the level of competition. It is important that the Government should carefully review NP's role, in general, and the costs and benefits of permitting NP to take such a dominant position, in particular. 2.64 There is always the need, particularly in a small LDC like Sierra Leone, to strike a balance between the efficiency imparted by rationalization or "streamlining" of the industry's facilities, staffing, overheads, etc., and the competition imparted by increasing the number of firms. One area where rationalization could be justified in Sierra Leone is that of storage terminals and tankage. An NP-Texaco buyout would not, however, achieve anything in this area since NP has no terminal facili- ties of its own at present. The only possible advantage to the economy of a buy out is in terms of rationalization of the service station network, as there are undoubtedly some NP-Texaco competing locations at present. 2.65 With regard to competition, it is not clear even with four or more companies how competitive a small market such as Sierra Leone is or can be under the most favorable circumstances: each marketer receives volumes from the common supply pool based on its historical market share; ex-refinery prices are controlled; pump prices are controlled so there can only be non-price competition; this leaves only the bulk consumer business where there is some flexibility to give a discount off pump prices based only on slight differentials in total marketing and trans- port costs--the competition here is largely non-price as well. Having said this, it is clear that there is competition between the companies, even if it is non-price, from which the consumer benefits in terms of service and the quality of the product. An NP-Texaco buyout would signi- ficantly reduce whatever little competition exists without many benefits accruing from the rationalization (as discussed in the previous para- graph). 2.66 Recommendation. It is strongly recommended that NP should not be permitted to buy out Texaco or any other marketing company. In addi- tion, the Government should review the longer term balance between private and public sector involvement in the marketing and distribution of oil products. Specifically, the Government should explicitly outline the objective and role of a parastatal like NP: for example, is it meant  - 37 - to operate as a commercial company on an equal footing with other marketing companies, or, is it expected to operate on a different basis including the provision of products and services (such as to the rural areas) which, while important to the country, are not commercially viable? Investment and Training Requirements 2.67 Investments. The refinery investment requirements have been quite modest over recent years, as follows: Table 2.10: REFINERY INVESTMENTS (USS '000) Year Amouint 1979 308 1980 118 1981 100 1982 258 1983 154 1984 50 1985 236 1986 (projected) 224 2.68 Over this period, vehicles and on-site houses for certain safety staff have accounted for more than half of the capital budget. Plant equipment only amounts to about 10% of the total. These figures indicate the low level of investment that has been required to keep the plant operational. Investments in an energy conservation program at the refinery will cost an estimated US$0.5 million of which the Government's share would be US$0.25 million. However, before such an investment is carried out, its economic viability should be studied (see technical assistance section below for prefeasibility study). An additional US$1.0 million of public investment will be required for the Government's share of building security stocks to 30 days of consumption (about 20,000 tons). 2.69 At present, there are a few small items required to enhance LPG recovery but they amount to less than $50,000 and have already been requested in the foreign exchange pipeline. 2.70 There are no significant investment requirements for either the rehabilitating or expanding of marketing company installations. Texaco has plans for a small jetty repair and improvement project at a cost of $60,000. An estimated $0.3 million of public investment will also be  - 38 - required for establishing 30 retail outlets on an experimental basis for kerosene and diesel in rural areas. As in the case of the refinery investment, a prefeasibility study should be carried out to evaluate the vabi1.ity of the investment (see technical assistance section). 2.71 Private truckers, financed by the marketing companies, are in immediate need of some new vehicle stock. The estimated 28 x 1800 IC trucks required within a year or two would cost about $1 million free of taxes or duties. The financing of this investment should occur naturally, with ro need for outside support, if the transport tariff is adjusted to economic, cost-recovery levels. 2.72 Technical Assistance and Training. The refinery is well supported by its BP contract in the area of process and maintenance technical services. Some of the senior staff would benefit, however, from broader, more economic and management-oriented courses. 2.73 There are staffing weaknesses also in the area of international crude and product supply, economics, and pricing. The most efficient and economic way to upgrade personnel skills in this area is to run workshops or seminars in Sierra Leone on the subject. Suggested attendees would be personnel designated for sector pricing/monitoring functions, central energy statistics, and planning personnel and refinery staff. 2.74 The marketers appear to have their training and technical assistance needs well in hand through support from their international affiliates. The training needs of NP staff, however, should be addressed by the Government. 2.75 As evidenced by the problems experienced on the mission in obtaining data, there is a pressing need to develop a petroleum accounting and information system. Although many of the individual sources of data are quite good, there is a serious lack of integration and reconciliation. Direct imports outside of the normal refinery/marketers system are difficult to determine. A system, possibly computer based, could be developed, debugged, and made fully operational over a period of 15 to 18 months. The cost of such a system would be approximately $200,000. 2.76 The cost of the recommended exploration promotion package is estimated at $0.5-1.5 million. A preparatory mission, for designing and planning the promotion, would cost an additional $30,000. 2.77 For refining, marketing, and distribution, the following areas need to be investigated in greater detail either separately, or prefer- ably as components of a single study: (a) the economic and technical viability of energy conservation investments at the refinery;  - 39 - (b) pricing policy both at ex-refinery and retail levels, including a review of transportation margins for deliveries outside Freetown; and (c) the need and strategy for reorganizing distribution of petroleum products in rural areas. The combined cost of all three study components is estimated at $100,000, with about half needed for reviewing pecroleum pricing structures and associated supply (b).  - 40 - III. ELECTRICITY Introduction 3.1 The public power supply system in Sierra Leone is relatively modest in size. Electricity is an expensive form of energy in relation to household incomes and the purchasing power of the traditional craft and trading sector in Sierra Leone. At the same time the modern sector, including mining, industrial, and commercial activities, is small and major establishments have set up generation facilities independent of the public supply system. In fiscal year 1986 total sales of the power 3.2 The public utility system consists of the Western Area Grid centered on the capital, Freetown, and fourteen isolated Provincial systems in the rest of the country. About 90% of the electricity supplied by the whole system is consumed in the country's four main cities: Freetown (82%), Kenema (3%), Bo (3%), and Makeni (2%). Two thirds of the approximately 30,000 residential connections to households are served by the Western Area Grid; the rest of the residential consumners are linked to the Provincial Systems. These residential consumers represent some 6% of the population that has direct access to electricity supplies. 3.3 The public supply is managed by the NPA, a commercial parastatal with a national monopoly on electricity generation, trans- mission, and distribution. In principle, any type of private generation requires an exemption from the NPA. However, because of deteriorating supplies, all classes of customers have purchased generating sets and are operatirig them with or without NPA's exemption. Based on a survey carried out by the mission, a total of about 30 MW of private generation capacity is now in place in the Western Area. These sets run on gasoline or gasoil (diesel fuel) and generate a combined ouitput of about 40 GWh annually. The output from these small producers, combined with private generation at the mines located in the provinces, exceeded NPA's total generation of 93 GWh during FY86 (Table 3.1).  -41- Tab.e 3.1: POWER SUBSECTOR--BASIC DATA National Power Authority Private generation Western Area Isolated Provin- Total Small Grid cial Systems Mines Producers Total Installed Capacity (MW) FY 1986 58 13 71 33 30 134 Generation (GWh) FY 1971 76 17 93 70 a/ 163 FY 1981 132 30 162 72 a/ 234 FY 1986 85 8 93 64 40 197 Losses (%) FY 1981 27 35 28 10 - Fv 1986 27 40 29 10 a/ Not available, assumed negligible. Source: NPA, Mission estimates. 3.4 The public power supply system is in a state of physical and financial disrepair, reflecting years of neglect, poor management, and a shortage of foreign exchange. At the time of the mission, less than half the installed generating capacity was in service, between cne-third and one-half of the electricity generated was being lost in .echnical and non-technical losses, and consumers were experiencing frequent and prolonged supply interruptions. These problems intensified further over the second half of 1986, reaching crisis proportions in September 1986, when only 6.2 MW of capacity was available to service an estimated maximum demand of 26 MW. Since then, supplies have improved somewhat but the situation remains precarious with the system highly vulnerable to breakdown. -I effect NPA, is unable to meet its responsibilities of providing adequate and reliable power supplies. NPA's current diffi- culties are the latest in a long history of managerial, technical, and financial problems at the utility during the last twenty years. These problems have been compounded over the more recent past by shortages of imported spare parts and fuel (for its smaller provincial units) due to a lack of foreign exchange allocation. The major short term issue relates to the institutional, physical, and financial pre-requisites for a successful rehabilitation program. These include: (a) an institutional framework which would allow the utility to operate autonomously, give adequate compensation to skilled staff, and ensure the effectiveness of senior management; (b) repairs and improvements to existing generating facilities; and (c) changes in tariffs and access to required levels of foreign exchange.  - 42 - 3.5 Meanwhile, government planners must also look at the options available for meeting electricity demand in the longer term. The dominant use of imported energy (oil products) to activate current generation capacity makes it desirable to look into options for develop- ing hydroelectric resources which would both reduce this dependence and be easier to maintain. At the same time, there is a need to examine the viability of and strategy for interconnecting isolated Provincial Systems as part of an integrated long term expansion plan. 3.6 This chapter discusses the above short and long-term considera- tions in five sections. rhe first three focus exclusively on the rehabi- litation program. Section 1 evaluates the institutional and managerial problems at the NPA and outlines the changes needed to successfully implement any program of physical and financial rehabilitation. Section 2 considers options, including existing rehabilitation plans, for repair- ing and improving existing facilities and the level and reliability of supplies. Section 3 looks at the issue of tariffs and the changes needed to strengthen the financial position of the NPA and manage demand. In Section 4, the longer-term options for expanding the Western Area grid and the Provincial Systems are examined. The last section outlines an investment program for implementing the proposed rehab;litation measures and initiating the least cost expansion program. Institutional Reform: Raising the Effectiveness of NPA Management 3.7 A major factor underlying the current crisis at NPA is the existence of a near paralysis in planning and decision making. This has eroded management effectiveness and lowered morale and overall perfor- mance. A fundamental restructuring of the organizational processes at the NPA are essential even to ensure that a rehabilitated systermi will work. Unless such a change takes place, physical and financial rehabili- tation (described in later sections) will not bring about a sustained improvement in the public power supply, which is the objective of the rehabilitation program. Organizational and management problems at the NPA can be divided into two categories: (a) problems arising out of the institutional framework within which NPA operates, and (b) managerial and operalional deficiencies within NPA itself. This section outlines some suggestions for significantly altering the organizational processes in these areas. Institutional Framework 3.8 The power subsector is regulated by the MEP, which was created in 1974 for establishing and implementing policy in the subsector. In carrying out its role, the MEP appoints members of the Board of Directors of NPA and reviews its annual budget and reports. NPA was incorporated under the Nitional Power Authority Act of 1982. The authority is run by  - 43 - the Board ot rs who appoint the Ceneral Manager. The Board is composed of fis :o seven members, with the General Manager an ex-officio member. 3.9 NPA's urganizational structure is that of a conventional utility consisting of the General Manager and six departmental heads covering finance, commercial engineering, generation, transmission and distribution, provincial operations, and personnel. In addition there is a special project group for Bumbuna. 3.10 The main institutional problems faced by the NPA are: (a) a lack of autonomy arising out its relationship with the MEP; (b) uncer- tainty about power sector policy; and (c) poor morale amongst skilled staff due to low public service pay scales. (a) Relationship between NPA and MEP. While the 1982 Act is intended to provide N4PA with a high degree of autonomy through an independent Board, in practice this is not the case. MEP officials tend to interfere with the day-to-day management of NPA and the Ceneral Manager has effectively no financial independence. Interference extends to important operational areas such as management decisions to disconnect supplies to non-paying customers. The General Manager's financial independence is undermined by the need to get the Minister's approval for expenditures of even a few hundred Leones (less than $50). At the same time, there are diffZculties in getting Covernment approval to raise tariffs. Part of the reason for interference is the physical proximi.y between the NPA and the MEP; NPA shares its headquarters with the MEP. A factor of greater significance has been the absence of a strong Boa-:d, which could shield NPA management from political and other pressures. The Minister can hire or fire any member of the Board, as and when he sees fit, and in the past many members have been civil servants or members of parliament. The Secretary of the Board is the permanent secretary of the MEP. The result of the structure is that the NPA Board is highly susceptible to MEP interference. Since the assessment mission in February 1986, changes have taken place in NPA's Board: a new Board has been appointed with a full-time chairman, and the General Manager has been re- placed. These changes have improved relations between the NPA and the MEP and have given NPA more autonomy than it has had in the past few years. While these changes are in the right direction and will enhance the effectiveness of NPA management, the basic problem of autonomy and the absence of members with sound commercial exper.ise remains: the Board is still over- represented by incumbent or ex-civil service personnel, and the limited autonomy which the Board may genuinely enjoy currently is dependent on the good will of the incumbent Minister of Energy and Power.  - 44 - (b) Uncertainty in power sector policy. The effectiveness of NPA management has also been :ompromised by foreign exchange problems which have forced changes in power sector strategy and left NPA without a policy framework in planning and decision- making. Shortages of foreign exchange have not only prevented planned maintenance schedules at power stations due to a lack of imported spares, but they have also been the main factor behind the Government's decision to delay the Bumbuna project (details discussed under long run issues). NPA had based its strategy on the project delivering power by 1986/87 and which would have considerably reduced demand on its operational staff. The delay has led to a temporary refocusing of efforts on diesel plants. In the meanwhile, uncertainty persists about whether the Government will allow the Bumbuna project to resume oir replace it by alternative schemes. The change in direction Lnd, more importantly, the continuing absence of an agreed policy on the subject has caused considerable confusion within NPA and lowered morale. (c) Staff remuneration and morale. NPA employs about 1,500 staff, divided approximately two-thirds to one-third between the Western Area and the Provinces. The authority is overstaffed in terms of its unskilled employees and has been taking steps recently to try and reduce manning levels. This has proved difficult as most employees are union members. The reverse problem applies to skilled eta'f, which NPA has difficulty in retaining, because of low public sector *,ay scales which are outside its control; salaries for technical staff are less than half what they could earn in the private sector. This has contributed to low morale and the utility has problems in keeping the caliber of staff it needs. The issue of remuner- ation to skilled workers is now being looked into by NPA, but no concrete steps have as yet been taken. Managerial and Technical Problems Within NPA 3.11 While external factors have not been favorable, both NPA and MEP recognize the need to strengthen NPA's management, planning, financial, and technical capabilities. It is difficult to assess the contribution made by management weaknesses to NPA's current problems. The institutional difficulties described above, combined with foreign exchange problems, have made management's task difficult, if not impossible. 7/ On the other hand, the mission was made aware of poor management practices in certain operational areas, such as the Kingtom 71 Senior management strongly feels that existing circumstances do not provide a fair test of their capabilities.  - 45 - power station and the Provincial Systems, which were unrelated to institutional or resource constraints. 3.12 The Kingtom power station is the nucleus of power supplies in the Western Area, yet there are no graduate (or equivalent) engineers on the station staff. Training of technical operations and maintenance staff has also been neglected. In the Provinces, there is a strong feeling of isolation and neglect reinforced by the recent difficulties in obtaining fuel and spare parts (see next section). In part, this is due to a lack of direct communications with the head office as there are no telephone connections. However, another reason is the limited contact with head office staff responsible for Provincial Systems: field trips made by responsible officials are few and far between. Coordinating a large number of widely dispersed provincial stations from Freetown is a costly operation which has added to NPA's financial and management burdens. Recommendations 3.13 To respond to these institutional and organizational problems, the mission recommends the following action program. (a) As a first step, and before the implementation of any rehabilitation measures, the following changes should be made by the Government regarding the institutional framework within which NPA operates: (i) The NPA Act should be reviewed and modified so as to permit NPA to manage its affairs without undue interference from the MEP. The Act should be made more explicit about the terms on which members of the Nh" Board are appointed and their qualifications. For example, limits should be set on the number of incumbent or ex-civil servants on the Board. The Act should also require the presence on the Board of respected members of the business community and representa- tives of consumers. To protect the Board's independence, the Act should specify a minimum period (preferably two years) for which members are appointed and during which they cannot be fired or replaced. The Government, while agreeing to the reconmendation in principle, has indicated that political constraints will, at present, prevent changes to the NPA Act. However, it has agreed to these changes as a target for the future. The mission acknowledges these difficulties but feels strongly that without changes to the NPA Act, an enduring solution to the problem of lack of autonomy cannot emerge.  - 46 - (ii) A mechanism should be devised to ensure that: (a) NPA can regularly set tariffs at levels which allow it to recover its costs on an agreed basis; and (b) it has access to sufficient foreign exchange for importing spare parts and carrying out routine equipment maintenance. (iii) In return for greater legal and financial autonomy NPA should be required to: (a) run on a ^ommercial basis achieving a target rate of return; and (b) be fullv accountable to the Minister of Energy and Power for its performance in providing reliable power supplies. (iv) The MEP should focus its efforts on power sector policy formulation and coordination. One of its first tasks should be the development of a policy statement outlining the GovernmentCs strategy for the sector in the short and longer term. (v) Remuneration to skilled NPA staff should be raised to make them competitive with the private sector. This will require a premium to be paid above public service scales as a compensation for technical skills. (b) Once the institutional changes have been implemented, the next step is for the Government and the NPA to agree on the scope and financing of a program to strengthen the organizational processes at NPA. The following components should be part of the program: (i) The senior management at NPA needs to be strengthened. The decision on the need for replacing members of the management team at the head office should be made by the new Board of Directors. However, at Kingtom, a plant supe-intendent should be appointed as soon as possible. The person selected should have a proven track record of managing a power stations with slow speed diesels. (ii) Planning sections covering the engineering and corporate financial functions should be introduced in NPA. These planning functions should be set up to report directly to the Ceneral Managzr. (iii) The Covernment should actively review the viability of NPA's Provincial Systems operations. Options to be considered should include: (a) decentralization; and (b) selling Provincial Systems to the private sector or local community groups, when such possibilities arise. (iv) The new technical training center (built with the help of the Federal Republic of Cermany) for secondary school leavers should be fully supported by NPA in terms of funds  - 47 - and staffing. Consideration should be given to the adoption of an externally recognized technical qualification at the end of the course. (v) A management development program should be designed for middle tier management. (vi) For technical staff at the intermediate and senior levels on-the-job training is required. At the more senior levels of technical staff, training should include diagnostic evaluation and preventive maintenance techniques. At the intermediate levels, training should include: diesel injector and turbocharger maintenance; pump maintenance; cable and overhead line construction and maintenance tech- niques; and maintenance of transfoomers, switchgear and protective devices. Concentrated courses of short duration run by visiting technical experts from manufacturers or specialized utility training staff are also recommended. Physical Rehabilitation of the Public Power Supply 3.14 The operational deficiencies associated with the decline in power supplies arise out of two related factors: breakdown of plant and equipment, and the economic problems experienced by Sierra Leone over the recent past. Specifically, the principal causes of generacing plant breakdowns have been a lack of preventative maintenance and a chronic shortage of spare parts due to foreign exchange limitations. Lack of fuel has not been a major problem except in the Provincial Systems. These run on gasoil, of which there has been a major shortage. This section evaluates the operational problems underlying power shortages, their implications for NPA and its customers and the options available to ease these problems. 3.15 Operational Problems. Generation in NPA's main grid, and the Western system covering Freetown, has declined sharply over the past few years from 132 GWh in FY 1981 to an estimated 85.3 GWh in FY 1986, or some 36Z. At the height of the crisis, in September 1986, NPA had only 6 MW of capacity to meet the system's peak demand, estimated at 26 MW. The capacity available at the time of the mission, 28 MW, was less than one-half of the installed capacity of 56.4 MW. Severe load shedding has been experienced since February 1983, although it has been generally planned according to a predetermined schedule based on available generation capacity. The supply situation has improved recently, with about 18 MW of capacity available in March 1987. However, the situation remains precarious: there is no spare capacity to cover for plant breakdown and supply is being maintained only by running some units on base load, for 24 hours, even though they are not designed to operate in this manner, thereby increasing the risk of future supply interventions.  - 48 - 3.16 Experience in the fourteen isolated centers in the rest of the country (the Provincial System) has been even worse with combined output in FY 1986 of 7.8 GWh, less than one-third of FY 1982 levels of 25 GWh. Since 1985 supply been restricted to a maximum service of 16 hours/day in all but one center, with some areas suffering more than others; genera- tion in Lungi has fallen by only 8% since 1981, whereas in Koidu the reduction was 63%. 3.17 Another symptom of NPA's malaise is a high level of losses: 2I% in the Western Area and an average of about 40% in the Provinces with some areas reporting losses of as much as 50%. 3.18 Generating plant breakdowns. The Western Area is supplied from two diesel power stations located in Freetown (Table 3.2). The larger and newer diesel station, Kingtom, has a capacity of 38.2 MW, comprising five slow speed fuel oil engines commissioned between 1964 and 1980. At the time of the -nission, only three of the five units were available for service and all of these were derated to a maximum total output of 15 MW because of cooling water and other technical problems. The two larger operating engines (9.2 MW Sulzer units) were 3 to 5 years overdue for major overhaul and could deliver a combined output of only 10 MW against a rated capacity of 18.4 MW. Of the three Man units, only one was in service. Its output has more recently declined from 5 to 3 KW. The capacity at Kingtom was significantly boosted by the installation of four 3 MW mobile, high speed Deutz units. These units, which were commissioned just before the mission, run on fuel oil but switch to the more expensive gas oil at high temperatures. Another drawback is that they are not designed to run as base load units, as NPA has been running tlhem. The second station, Falconbridge, has a total capacity of 6 MW from 5 units commissioned between 1960 and 1976. The units all burn gasoil. At the time of the mission only two units were available for service providing a total output of less than 3 MW. Table 3.2: POWER PLANTS ON WESTERN AREA GRID Kingtom Falconbridge Total Unit no. 1-3 4-5 6-9 1-3 4-5 (Man) (Sulzer)(Deutz,mobile) Nameplate rating (MW) 19.8 18.4 12.0 3.0 3.2 56.4 Available a! (MW) 5.0 10.0 12.0 2.4 0 29.3 After rehabilitation (MW) 17.7 18.4 12.0 2.4 2.8 53.3 Commissioning date 1964-71 1978-80 1986 1962 1976 Speed (rpm) 130.5 150 1000 428 750 a/ At the time of the mission in ear.y Rebruary. Source: NPA.  - 49 - 3.19 As can be seen in Table 3.2, the available capacity during February 1986 was 28.6 MW. This capacity was reduced significantly during the remainder of 1986 because of major overhauls carried out on the two Sulzer units. 8/ Work has also been completed on the modifi- cation to the sea water cooling system on the Sulzer units at a cost of $1.1 million. This work, which is part of NPA's planned rehabilitation of all units at Kingtom and Falconbridge, has been carried out under the Power Sector Engineering and Technical Assistance Project funded by the IDA (Credit 1265-SL, 1982) and the OPEC Fund (Loan 292-P). The effects of NPA's rehabilitation plan, if realized, are shown in Table 3.2. The available capacity will then be raised to 53.3 MW. Allowing a reserve plant margin of the largest and next largest unit, the capacity would be adequate to meet a system demand of up to 37.5 MW. 3.20 In the fourteen isolated Provincial Systems under the control of the NPA, the total plant capacity of 12.5 MW supplied an aggregate demand of 8.6 MW in 1981. Since that time, available capacity has declined because of a lack of spare parts. Three new 500 KW units have been installed at Port Loko, Makeni, and Koidu for more than two years but have not been commissioned due to a contractual dispute. NPA should attempt to resolve this dispute as a matter of urgency and arrange for these sets to be commissioned. 3.21 Transmissior./Distribution and System Losses. There is no major technical problem related to transmission and distribution in the Western Area Grid except for switch gear. All of the seven 11 kV indoor type primary substations on the network are fitted with oil circuit breakers. These have been placed under great strain because of regular load shedding since 1983 leading to contact erosion and, in some instances, malfunctioning of the breaker. Apart from these maintenance problems, the substations are in good condition except for "Wilberforce" where the switch gear is of an obsolete design and in need of replace- ment. 3.22 Losses on both the Western Area and Provincial System3, as already mentioned, are high: non-technical losses arising from illegal connections, tampering with meters, faulty meters, and meter reading or billing errors are considered to contribute 10-15% (out of a total of 27%) losses in the Western Area and 20-25% (out of a total of 40%) losses in the Provincial Systems. Both the split between technical and non- technical losses and the source of non-technical losses are urgently in need of further study. There is a special investigative unit at NPA but its activities are restricted by lack of funds and equipment. 8/ The capacity, as mentioned earlier, has now recovered to about 18 MW following the completion of repairs to one of the Sulzers.  - 50 - 3.23 Factors Underlying Plant and Equipment Breakdown. There are a number of common factors underlying all operational problems in plant and equipment in the NPA system. The first of these is a shortage of spare parts resulting from difficulties in obtaining foreign exchange. In many instances, plants have been idle for long periods because of the unavailability of relatively minor parts. A related problem is lack of maintenance. This is both because of the absence of spare parts and pressure on NPA staff in the current period of shortages not to take plants out of service. At the time of the mission, the two Sulzer units were respectively 25,000 and 20,000 hours overdue for major overhaul. Later in the year, NPA obtained the spares to carry out the overhauls, purchased under a UK grant. The overhaul of one of the units has been completed, and the other is in progress. More spares are being purchased under IDA Credit 1265-SL. 3.24 Prolonged outages have occurred at Kingtom due to unexpected technical problems. Four out of the five units have had problems asso- ciated with stator faults, foundation problems, broken tie bolts, and cooling water failure. These have each resulted in outages of between one and four years. It is difficult to assess how many of the technical problems could have been avoided through a better choice (if that was open to Sierra Leone) of the make of the generating sets before instal- lation. It is clear, however, that the delay in resolving these problems has been exacerbated by shortages of qualified technical staff. 3.25 Shortages of Fuel. While the main constraint on the Western System has been the shortage of serviceable generating capacity, for the Provincial Systems it has been the lack of fuel. The fuel supply problems have already been discussed in Chapter II. Suffice it to say that the problems are much more severe in the Provinces. Part of the reason why the latter have borne the brunt of the fuel shortage problem is their exclusive use of gasoil. In thie case of the Western System, Kingtom operates on fuel oil, of which there was a surplus even in 1985. The other reason is the lower priority given to the Provinces as illustrated by the following statistics for gasoil deliveries in 1985. Table 3.3: GASOIL ORDORED AND SUPPLIED (1985) Fuel Ordered Fuel Supplied % Supplied Western Area 722,00C gallons 578,000 gallons 80 Provinces 2,747,000 gallons 733,000 gallons 27 Source: NPA. The priority given to fuel deliveries to the Provincial System by the NPA head office should be urgently reviewed. This problem has been aggra- vated by a lack of communication facilities between the Area Engineers at the Provincial Centers and the NPA head office.  - 51 - 3.26 Cost and Benefits of Rehabilitation. In 1985, the Bank estimated the costs of rehabilitating the Western Area System at about $12 million. Taking account of the work carried out since that time on the generating plants at Kingtom (including $1.1 million for the modifications to the cooling system paid for by IDA), the mission estimates that a further $10 million is needed for rehabilitating the Western system. Details are provided in the investment section below. 3.27 Since 1983, there has been a proliferation of private generation ranging in size from a few kilowatts for a private dwelling to hundreds of kilowatts installed at commercial and industrial establish- ments. A brief survey carried out by the mission estimated that there is probably 25-35 MW of private generation capacity in the Western Area (Annex 7). The import and use of these small sets, which run on gasoline or gasoil, has added to the countries foreign exchange problems. 2.28 It is estimated (see Annex 7) that the use of private generating sets produced 40 GWh in FY86, and cost the economy $1.5 million more than if the power had been -enerated by the NPA (based on the difference between the fuel cost of generating power at Kingtom of USC3.4/kWh and the autoproducers cost of USC7.1/kWh). 9/ When the capacity cost of private generating plant is taken into account (based on a 10-year life, an annuity rate of 12%, and a cost per kW of $800), the annual cost rises to $4 million. These estimates suggest that the cost of unreliable supplies to the economy may have been quite high. A rehabilitation program, which is successful in raising the reliability of public supplies, would pay for itself by discouraging the use of private sets. A saving of even $3 million a year would repay a $10 million investment program in 3-4 years. Concerns About Existing Rehabilitation Plans 3.29 A serious risk attached to existing rehabilitation plans concerns the viability of repairs to the three MAN units at Kingtom. These units are about 20 years old and are approaching the end of their normal working lives: the three units have done 70,000, 85,000, and 95,000 hours respectively compared to a usual maximum of between 100,000 and 120,000 hours. In addition, all three units have experienced frequent and prolonged breakdowns in the past due to both foundation problems and overheating of stator and rotor windings. While repairs have been and can be carried out in the future, the problems are likely to reoccur if there are fundamental defects in these units, which are of a prototype design. An independent expert should be called in to review the technical and economic viability of keeping these units in service. 9/ The comparison with Kingtom is relevant because most of the demand would have been supplied fLom base load plant--if it had been in service.  - 52 - Contingency plans should also be made for new generation capacity to replace these units. 3.30 Recommendations. NPA should consider adopting the following strategy to improve system operations: (a) Rehabilitate all the generating plants at Kingtom and Falconbridge, with priority given to the overhaul of the Sulzer units As the overhaul of the Sulzer units has brought avail- able capacity below the natural demand level, it is vital that NPA proceed with completing the overhauls and rehabilitation program as a matter of extreme urgency. (b) Carry out an urgent independent review to determine the economic and technical viability of keeping the three MAN units at Kingtom in service. Contingency plans should also be made for new generating capacity to replace these units. (c) Carry out a loss reduction study on the Western Area system, including an investigation into measures to reduce non- technical losses on the Provincial system. (d) Commission the new diesel units at Port Loko, Makeni, and Koidu. (e) Improve communications by the installation of a high frequency radio system between Freetown and the main Provincial Centers. (f) Initiate replacement of the 11 kV switchgear at Wilberforce substation. Tariff Policy, Demand Management, and Financial Rehabilitation 3.31 Together with a strategy for rehabilitating the supply of electricity, the problem of power shortages should also be addressed through better demand management, including changes in tariff policy. In the long run, tariff increases would help to alleviate shortages by dampening demand, but because of the magnitude of suppressed demand, the more immediate impact of tariff increases is likely to come from an improvement in supplies through a strengthening of NPA's finances. At the time of the mission, tariffs were well below both the short and long run costs of providing power and thus contributed to NPA's precarious finances. This section evaluates the role of tariff policy in the current crisis and emphasizes the urgent need for a review of this policy.  - 53 - Demand and Supply Side Implications of Tariff Policy 3.32 Tariff policy has direct implications for both NPA, the supplier of electricity, and its customers. 3.33 On the demand side, the problem faced by the consumer is that of suppressed demand. As shown in Table 3.4, after growing steadily over the 70s, generation declined at an average rate of 12% p.a. between FY82 and FY86 and falling to about the same level as in 1971. This decline has seriously inconvenienced residential consumers and adversely affected economic activity across a wide spectrum of users from tourist hotels to commercial and industrial organizations. A measure of suppiessed demand and, therefore, of the welfare loss associated with supply diAruptions is provided by the proliferation of private generating sets imported and used at considerable cost to the owners. The issue on the consumption side is the extent to which demand management measures, such as tariff increases or better allocation of available public supplies, can help to significantly alleviate the problem of suppressed demand. Table 3.4: NPA GENERATION AND SALES Growth rate FY1971 FY1982 FY1986 1971-82 1982-86 (GWh) (GWh) (GWh) (%/year) (%/year) Generation) Western area 76.0 133.5 85.3 5.3 -10.6 Provinces 16.5 24.9 7.8 3.8 -25.2 Total 92.5 158.4 93.1 5.0 -12.4 Sales Western Area 60.0 98.1 62.3 4.6 -10.6 Provinces 12.6 14.9 4.7 1.5 -25.1 Total 72.6 113.0 67.0 4.1 -12.2 3.34 On the supply side, past tariff policy has directly contributed to NPA's deteriorating finances. In FY85, total billings at 28 million Le were well below operating costs of 39 million Le. Net losses for the year amounted to 7 million Le ($1.2 million at the official exchange rate) in spite of a Government subsidy of 6 million Le. Net losses are estimated to have increased to 9 million Le in FY86. These losses have been mainly caused by declining sales and delays in Government approval of tariff increases to compensate for increased costs. NPA is increa- singly short of cash and, currently, is even unable to pay for fuel supplies, lubricants, or urgently needed spares. The issue here is the extent to which tariffs need to be raised to allow NPA to meet its finan-  - 54 - cial objectives without providing a cover for inefficient operations. At the same time, a mechanism needs to be agreed upon which would allow an estimation of tariff charges on a regular basis to reflect changes in NPA's cost--for example, to take account of changes in exchange rates and oil prices. 3.35 Suppressed Demand: Extent and Causes. An analysis of historical data suggests that demand, prior to the inception of supply problems in 1982, was significantly influenced by changes in income: the short run (one year) impact of a 1% increase in GDP is a 1.03% increase in demand. Eventually, over the long run, demand increases by 2.4% (Annex 8, page 3). Electricity tariffs also appear to have a statisti- cally significant linkage with demand but the exact magnitudo is subject to some uncertainty: the estimates of short run price elast city range from -0.17 to -0.34 (Annex 8, page 1). The lower end of the range implies that the effect of tariff changes on demand is relatively weak, at least in the short run; a 1% increase in tariffs would only reduce demand by 0.17% in the short run and by 0.4% in the long run, assuming the ratio of short to long run effects is the same as for income. Using the linkages estimated between demand and CDP and ignoring price effects, suppressed demand--the gap between demand and available supply--reached a level of 81 GWh (Annex 9) in FY86, about the same as total generation in the Western Area. 3.36 Tariff Increases and the Relative Economics of Public and Private Generation. The mission's nalysis suggests that the large increases in tariffs, as introduced recently to allow NPA to recover its costs, may make it difficult for the utility to win back customers (especially those with access to privately generated supplies), and return sales to previous levels. Based on the mission's estimates discussed earlier, autoproducers generated about 40 GWh in FY86, half the level of suppressed demand, the rest was lost. The short run cost of autogeneration (based on fuel costs and ignoring capital costs as sunk) ranged from LE 0.6/kWh to LE 1.6/kWh in 1985, depending on the amount of fuel purchased at black market prices (Annex 11). These costs provide a measure of the willingness to pay during 1985 for alternative supplies by industrial, commercial, and relatively well-off consumers, who account for the dominant proportion of sales. These costs were, on average, more than twice the level of NPA's average tariff existing at the time of Le 0.49/kWh, and higher than that recommended by the Bank in that year (1985), Le 1.0/kWh. If replacement cost for private generation capacity is taken into account, the difference between estimated willingness to pay and existing or contemplated tariffs would be further magnified. With low, and as we suggest below, subsidized tariffs, electricity supplies from NPA were highly competitive relative to self-generation. However, the relative economics of private and public supplies have changed recently with the adoption by the Government of a large increase in the averag., tariff to Le 3.78/kWh (with large commercial and industrial consumers paying Le 6.1/kWh, small commercial consumers paying Le 3.7/kWh, and residential consumers being charged Le 1.7/kWh). The marginal cost of private generation, based on the most recent (April  - 55 - 1987) gasoil price of Le 47/IG, is estimated at less than Le 3.5/kWh. With the marginal costs of private generation for industrial and commercial consumers (who own more than 80% of private generators, see Annex 7) now well below the level of tariffs, NPA will have to demon- strate its ability to provide reliable supplies if it aims to quickly win back customers. 3.37 The seriousness of the problem is better appreciated by noting that even these revised higher tariffs are now below economic costs, following the depreciation of the Leone and the need to be raised (see discussion below on supply costs and tariffs), thereby further reducing the competitiveness of public supplies. A major factor behind the high cost of electricity supplies, and their lack of competitiveness, is the high level of losses. These need to be curtailed to make a significant reduction in NPA's costs. At the same time, the prices of gasoil, the main fuel used in private generating sets, have often been subsidized. The Government should at least ensure that gasoil prices at the retail level are continuously maintained at levels which reflect the economic cost of supplies, as discussed in Chapter II. NPA's sales are unlikely to be choked off by the large tariff increase, given the magnitude of suppressed demand, especially from customers who do not have self- generation facilities. However, in the long run, a return to a stable pattern of growth in electricity sales, at prices which allow cost recovery, requires an improvement in NPA's performance (both in terms of loss reduction and supply reliability), and a Government pricing policy for petroleum products which does not discriminate against public supplies of electricity. 3.38 The Effect of Tariff Increases on Inflation and Affordability of Electricity Supplies. Quite separate from the issue of substitution between private and public generation, increases in tariffs raise questions regarding the impact on inflation, income distribution, and the affordability of electricity consumption by residential users. Based on available evidence, however, none of these concerns provides a basis for holding back economically justifiable tariff increases. The results of the limited household energy survey carried out by the mission (discussed in more detail in Chapter IV) show that electricity consumption in urban areas accounts for about 4% of the monthly budget of households. The figure is approximately the same for all the income classes surveyed, suggesting that households in all income classes would be adversely affected to the same extent by tariff increases. Po-icy makers should not, therefore, shy away from tariff increases on the mistaken grounds that they will bear excessively on the poor. 3.39 The effect of tariff increases on inflation is more difficult to assess, especially as no data is available on the proportion of domestic production costs accounted for by energy in general, and electricity in particular. However, there are some grounds to suggest that tariff increases would not add significantly to inflation. Firstly, all producers of domestically consumed goods and services cannot pass on higher prices of electricity to consumers; there are Government controls  - 56 - on the prices of some products, while imports determine local prices on many traded goods. So the ripple-through effect on the prices of other goods and hence on inflation is likely to be attenuated. Secondly, in spite of the sharp rise in average tariffs during 1986 (from Le 0.49/kWh to Le 3.78/kWh), the increase in NPA's revenues (about 330 million Le even assuming sales of 100 GWh) will still be less than 2% of the estimated GDP for 1987. If GDP is taken as a proxy for final demand in the economy, the increase in tariffs does not appear to add significantly to inflationary pressures: experiences in other developing countries whose degree of "openness" (as measured by the ratio of exports to GDP) is the same as for Sierra Leone (where exports account for about 30% of CDP) suggests that an increase in energy prices yielding revenues equiva- lent to 2% of final demand would increase consumer prices by about the same magnitude. Hence even the 800% increase in tariffs during 1986 is not likely to increase inflation significantly above existing levels. 3.40 Lastly, the large increase in tariffs raises issues of affordability of electricity consumption in Sierra Leone. At the time of the household survey, the domestic tariff was Le 0.41/kWh, and electri- city consumption accounted for, as mentioned earlier, about 4% of the monthly budget. With the increase in the domestic tariff to Le 1.7/kWh, electricity's share of the household budget would have increased to 16% assuming that electricity consumption and wages remained largely un- changed. An increase of this magnitude would have severely strained household budgets. The consumers' response to such a shock is difficult to assess- The econometric estimates, discussed earlier, suggesting a low sensitivity of demand to price changes in the short run are not valid for the large and sudden increases experienced in 1986. Consumers could well respond by reducing consumption of electricity through reduced energy usage (belt tightening measures) as well as by substitutirn of other energy forms where economic, such as using a low grade light from a fuelwood stove instead of an electric lamp; it is not clear that substi- tution of kerosene would save money, given the large increases in its prices as well. However, while the price of electricity relative to the price of labor (wages and salaries) may now appear high, it is a neces- sary signal of the costs of, what is for all practical purposes, a luxury good in the Sierra Leone context. 3.41 Other Options for Managing Demand. The normal daily load curve on the Western System exhibits two peaks of approximately equal size, one during the day from 9:00 a.m. to 4:00 p.m. and one in the evening from 8:00 p.m. to 11:00 p.m. This cycle gives a daily load factor of aJout 83%. There is, therefore, little obvious scope for improving demand (load curve) management: a tariff, or other signal, to industrial and commercial consumers not to consume during the evening peak would quickly have the effect of shifting the peak into the day. There are other possibilities for managing demand such as negotiations with large consumers for providing interruptible supplies. However, under existing crisis conditions at NPA, such options for demand management should be low in priority.  - 57 - Supply Costs and Tariffs 3.42 In the absence of an agreed system expansion plan, the basis for evaluating the economic cost of electricity supply is assumed to be the option which will generate power at the least expected cost--the Reduced Bumbuna Project followed by thermal development--as evaluated in a later section. These estimates are indicative of a lower bound on future supply costs. The mission has made an approximate estimate of the range of average incremental costs for commissioning and operating new generating capacity and transmission and distribution extensions associated with this expansion option. Details and assumptions are in Annex 17. 3.43 Marginal capacity costs of rehabilitating the existing system and installing and operating new generation, transmission, and distribution has been estimated at: (a) $630-$950/kW/year, depending on the rate of growth of demand, for the medium voltage consumer on the interconnected system; and (b) $330/kW/year for the medium voltage consumer on the diesel isolated systems. For low voltage consumers the capital charge is 20% higher. Marginal energy costs have been assessed on the basis of: (a) fuel (and lubricating) costs that are estimated for the two alternative projections of fuel prices, discussed in Chapter I, and then averaged; and (b) an allowance for losses. Finally, average incremental cost (marginal capacity plus marginal energy costs), have been assembled on the basis of an assumed operating regime of 5,870 hours/year (67% use) for the interconnected system and an average 2,200 hours/year (25% use) for the isolated syste.n (Table 3.5). Table 3.5: NPA LONG RUN SUPPLY COSTS a/ (Le/kWh in 1986 prices) Base Case Scenario Accelerated Growth Scenario Medium Low Medium Low voltage voltage voltage voltage Marginal Energy Costs Interconnected system 2.0 2.2 2.1 2.3 Isolated system 3.2 3.6 3.2 3.5 Virginal Capacity Costs Interconnected system 8.7 9.2 5.8 6.1 Isolated system 11.4 12.8 6.1 6.5 Average Incremental Costs Interconnected system 10.7 (22.3) 11.4 (23.8) 7.9 (16.5) 8.4 (17.5) Isolated system 14.6 (30.4) 16.4 (34.2) 9.3 (19.4) 10.0 (20.8) a/ The table is based on costs estimated In USS at constant 1986 r,rices (see Annex 17), and the figures in brackets are in USf/kWh at 1986 prices. The Leone values have been derived by converting the dollar estimates using an exchange rate of Le 48/S. The toble can be re-estimated for any other exchange rate to a first approximation by con- verting back into dollars at Le 48/S and then reconverting at the new exchange rate. This approximation is valid as over 90% of costs are incurred in foreign exchange.  - 58 - 3.44 The indicative economic costs suggest that on a kilowatt-hour bas;A: (a) delivered costs of electricity vary by more than a factor of two, from about Le 8-16 in 1986 terms; (b) supply costs are sensitive both to the size of the system being supplied and the rate of growth of demand (linked to the two scenarios); (c) low voltage supply costs are higher than at the medium voltage level; and (d) off peak costs (i.e., marginal energy costs) are in the range of 20% to 30% percent of average incremental costs for medium voltage consumers. This pattern of supply costs should form the basis of NPA's future tariff structure, if the current system is eventually to be replaced by the least cost option involving a mixed hydro-thermal system. 3.45 Tariff Structures. NPA's tariff schedule comprises residential, commercial, industrial, street lighting, and temporary con- sumers. There is a minimum charge per customer and charges per kWh increase after the first 100 kWh of consumption. Tariffs for industrial customers both at low voltage (up to 15 kW) and high voltage (over 15 kW) differentiate between energy, demand, and reactive en rgy charges. De- tails of tne schedules existing at the time of the mi sion are given in Annex 10. This tariff structure has a number of anomalies that have, in the past, added to NPA's financial problems and provided incorrect signals to consumers. A detailed review of this structure would need a full tariff study. There are, however, some obvious areas where improvements can be made, as revealed by comparing these tariffs with: (a) structure of average incremental costs, as discussed above; and (b) NPA's financial requirements. The following comments focus, first, on ways of enhancing the incentive effects of tariffs and, secondly, the likely trends in average tariff levels taking account of more recent changes implemented in the second half of 1986. 3.46 In terms of structure of existing rates, the most obvious distortion is the differential between low voltage and medium voltage tariffs. The average level of tariff for low voltage consumers, at the t- of t.he mission, was around Le 0.45/kWh, which was not much different from the average medium voltage rate of about Le 0.43/kWh. There should be a greater differential between these categories to reflect the signi- ficantly lower costs of supplying customers who accept power at higher voltages. 3.47 Another structural anomaly was in the make-up of domestic tariffs. The monthly kilowatt-hour charge for domestic consumers was in the form of an increasing two block tariff: Le 0.41/kWh for the first 100 kWh (the lifeline rate), and Le 0.47/kWh for higher levels of consumption. The problem lies in the band-width of the lifeline tariff. The household survey carried out by the mission suggests that the average domestic consumption is 55 kWh/month. A band-width of 100 kWh/month is giving away benefits to relatively well-off consumers. A splitting of the lower block (0-50 kWh, and 50-100 kWh) is called for to reduce this subsidy. Neither this structural anomaly nor the previous one appear to have been fully addressed in the recent tariff revisions.  - 59 - 3.48 Thirdly, the level of tariffs at the time of the mission was below both the economic and financial costs of supply. Tariff levels have been a continuous problem at NPA. The Government, for political and social reasons, has been unwilling to allow NPA to raise tariffs, except as a result of insistence from the Bank group. At the time of the mission, the average tariff was Le 0.49/kWh. This was half the level of Le 1/kWh that the Bank had recommended in 1985, (and equivalent to USE 16/kWh) to enable NPA to recover its financial costs. As NPA's tariff is applied unifotmly to all parts of the country, the rate of Le 0.49/kWh was insufficient to even recover the fuel costs of the isolated systems, estimated at Le 0.52/kWh in 1985. (The provision of a Government subsidy to NPA for recovering losses on provincial losses does not eliminate the distorted signal provides to consumers.) Recently, and as already mentioned, the Government, as part of the SAL negotiations, agreed to raise the tariffs to Le 3.75/kWh (or USC 15/kWh, based on an exchange rate of Le 25/$, valid at the time of the last increase during October- November 1986). Further increases will be necessary in future to bring tariffs into line with the financial and economic costs of supply; the rapid depreciation of the Leone in recent months necessitates additional large increases. The Bank's most recent estimates suggest that tariffs need be raised to about USC 20/kWh (see Annex 17) to allow full cost recovery and an 8% return on revalued assets, as agre_d in earlier discussions between the Bank and the Government. If tariffs are set at this high level, then they will be at the upper end of the range of rates (USc 10-20/kWh) currently existing in other West African countrias. The need for such high rates is in part a reflection of the relativesy high losses faced by NPA and suggests the urgency of measures to imlrove NPA's operational efficiency. While losses are likely to decline ir. future as a result of efficiency improvements, it is expected that they will remain at sufficiently high levels (15-20%) to maintain the p:essure on long run supply costs; the long run marginal cost of supply to customers at the grid is estimated to be in the range of USC 16.5-23.8/kWh (Table 3.5). 3.49 Other Financial Weaknesses at the NPA: Billing and Revenue Collection. An additional faccor which has contributed to NPA's finan- cial weaknesses is billing and revenue collection. Bills, at the time of the mission, were two months late because of delays in their preparation; due to the lack of an in-house computer to assist in billing, bills are sent to Monrovia, Liberia, for preparation or are done by hand. Delays in billing have led to cash flow difficulties. A related problem is that of late payments by larger organizations anLd government departments. In addition to tariff changes, financial control has to be strengthened if cash flow is to improve. Recommendations (a) NPA should implement a tariff study to design a more logical tariff structure that would help to meet NPA's financial objectives, reduce the existing anomalies and distortions, and make suggestions on how tariffs should be changed at a time of rapidly declining exchange rates and during a transition to a hydro-based system (the study should also define means for  - 60 - ensuring that only reasonable costs can be recovered by NPA through tariffs); (b) NPA should carry out a study to determine the most appropriate computer hardware and software for in-house billing and proceed without delay to procurement; and (c) NPA should develop an action plan for reducing arredrs. Long Term-Issues 3.50 The main long term issue revolves around the options available for meeting the future demand for power in Sierra Leone. In considering alternative development strategies, the government's long-term objectives are: (a) the development of a hydro-based power system to reduce its reliance on imported energy; and (b) interconnection of provincial centers to improve the reliability of power supplies in the interior of the country. This section considers two aspects of the long run power supply issue: (a) the viability of the Bumbuna hydro electric scheme relative to other options to meet future demand in the Western Area; and (b) development options for provincial centers. As a background to this discussion future levels of power demand estimated under the two growth scenarios, AG, and BC are summarized in Table 3.6. Details of these projections are given in Annex 12. The forecast demand is quite consistent with neighboring countries experiences for the period beyond 1990, as suppressed demand will be eliminated.  - 61 - Table 3.6: DEMAND FORECAST Western Area Provinces Mines Total s/ AG BC AG BC AG BC AG BC Maximum Demand (MW) 1986 22 22 3 3 10 10 35 35 1990 30 24 12 10 12 10 54 44 1995 48 29 22 11 53 10 83 50 2000 75 38 37 14 14 11 126 63 Generation (GWh) 1986 85 85 8 8 64 64 197 197 1990 167 141 41 35 73 61 284 257 1995 280 169 64 42 83 65 430 296 2000 444 221 92 55 91 70 630 366 Sales (GWh) 1986 62 62 5 5 58 58 165 165 1990 131 108 27 23 66 55 227 206 1995 238 135 49 31 75 59 365 245 2000 378 177 78 44 82 63 541 304 Growth in generation (% p.a.) 1986-2000 12.i 7.1 19.1 14.e 2.5 0.1 8.7 4.3 1990-2000 10.3 4.6 8.4 4.6 2.2 1.4 8.3 3.8 a/ Total figures for generation and sales include 20 GWh under BC and 3 GWh under AG for smalI auto producers; for 1986, an estimate of 40 GWh is included urder both scenarius. 3.51 The main assumptions underlying these forecasts are: (a) over the remainder of the 1980s, sales are supply constrained but demand is driven by GDP growth beyond 1990; (b) the effect of the early tariff increases, assumed under AG, are not reflected ir, the demand figures until che early 1990s as suppressed demand, and are not eliminated until FY88/89; (c) in the case of BC, no price effects are assumed since, under this scenario, tariff increases are both late and limited; (d) technical and non-technical power losses are assumed to decline to 20% (14% tech- nical, 6% non-technical) in the year 2000 under BC and to 15% (12% tech- nical, 3% non-technical) in the AG scenario; (e) mining power demand is linked to mining activity levels projected under the two scenarios; and (f) autogeneration declines under both scenarios following the rehabili- tation of the Western Area System--it falls to half its current level (20 GWh) under BC and to a negligible value (3 GWh) under AG. Bumbuna vs. Other Development Options 3.52 The viability of the Bumbuna hydroelectric scheme is the central long term power issue. The main issue is whether Bumbuna is to be preferred to alternative thermal or hydro options. If Bumbuna is the preferred option, another issue is whether Phase 1 of che full Bumbuna Scheme (as originally designed, or the Reduced (version of the) Project  - 62 - should be constructed. On the basis of the availab.e data, the Reduced Bumbuna Project is the least cost solution over a flairly wide range of load growth forecasts and fuel price assumptions. However, there may be room for improvement in the design and sequencing of the project which, as currently envisaged by the consultants to the Bumbuna scheme, is not optimal. At the same time, the implementation of the scheme will require the resolution of macroeconomic financing constraints in terms of the availability of both foreign exchange and of fiscal revenues required for meeting local costs. 3.53 Timing of Planned System Expansion. Once the rehabilitation of the Kingtom and Falconbridge power stations is completed, the plants there should be capable of a total output of 53.3 MW with all units in service. This should be sufficient to meet the demand on the Western Area system until FY93 on the AC forecast and FY97 on the BC forecast (Annex 13). However, in the event that it is uneconomic to restore the MAN units at the Kingtom to full service, then timing of the power system expansion will have to be brought forward. 3.54 Options for Power System Expansion. Tne following development options are technically viable for supplying power to the Western Area: (a) The Mano Rivet ichene. The Mano River flows along the southern border of Sieria Leone with Liberia. A potential hydro site on che river was identified in 1981 (see map) which could support a 180 MW project at a cost of US$356 million ($1980/kW). The scheme would involve a large storage reservoir providing over- season regulation of the river flow and energy production of 760 GWh in an average year. The capacity of the scheme is much larger than could be absorbed by Sierra Leone for sometime. Hiowever, all the capacity could be absorbed by Liberia if it could afford the scheme. The site is also significantly closer to Monrovia and the Liberian grid than it is to Freetown and the Western Area grid. As such, it has been considered mainly as a project for Liberia, at least for the next 20 years. Any joint venture between Sierra Leone and Liberia for the development of the Mano River Scheme would have to overcome the technical problem of different f-equencies; Liberia's power supplies operate on a 60 cycle per second system compared with 50 in the case of Sierra Leone. There are also political difficulties involved in developing a scheme straddling the border. (b) Thermal Options. At present, Sierra Leone relies totally on imported oil for power generation. Known lignite deposits are of limited size and not a serious candidate for future thermal generation. The thermal options available for future power plants in the Western Area are diesel engines burning fuel oil, steam plant burning either coal or fuel oil, and gas turbine plant burning gasoil. Steam plant is not economic against diesel at unit sizes below 20 MW but, at larg-r unit sizes,  - 63 - steam plant burning imported coal would be attractive for base load duty, even at the current depressed prices of oil. Existing diesel plant would be retained for peaking duty if a new base load steam plant were to be construActed. These could eventually be replaced by peaking gas turbinei. (c) The Bumbuna Falls Hydroelectric Project covers the development of the potential of the Seli River upstream of the village of Bumbuna about 180 km from Freetown, for an ultimate total power of 305 MW producing in an average year 1,250 GWh of firm energy. Since both the power and energy represent about ten times the present demand of the Western Area system and neighboring provincial load centers, the project would be developed in phases with each phase constructed to match the demand at the time. The optimum physical development sequence involves four phases, with the third phase effected in two steps. The first phase would comprise a storage dam upstream of the Bumbuna Falls, a power tunnel through the hillside to a shaft power station in a ne ,hboring valley, thereby developing a net head of 114 m. The installed capacity in Phase 1 would be two 26.7 MW units. The Bumbuna reservoir storage would be small, insufficient to provide full over-season regulation of the river flow. The reductions in energy available during the dry season are shown in Annex 14. In Phase 2, a storage reservoir, capable of providing over-season storage, would be constructed at Yiben, about 30 km upstream of Bumbuna, and a third unit would be installed in the Bumbuna power station, bringing the capacity to 80 MW. In Phase 3a, the Yiben dam would be raised by 15 m and three 15 MW turbines (fitted with 35 MW generators) installed in a power station at the foot of the dam. The Yiben dam would be raised by a further 24 m in Phase 3b and the Phase 3a units uprated to 35 MW by the installation of new turbine runners. The final phase (Phase 4) would involve the construction of a second power station at Bumbuna with three 4G MW units. Table 3.7 sets out details of the project and illustrates the effect on the cost of energy of the increased storage capacity in the second and subsequent phases of the project. (d) The Reduced Bumbuna Project comprises a revised Phase I design involving a higher dam--a 10 meter increase in height--and a power station, with a single 47 MW unit, at the foot of the dam utilizing the spillway tunnel, thereby eliminating the long pressure tunnel. Economies were alsc made in other aspects of the design which would lead to reduced security of supply from Bumbuna, e.g., the elimination of the bottom outlet gates and the installation of a single generating unit. These economies resulted in a substantial reduction in the estimated capital cost, see Table 3.7, at the expense of a lower net head (and  - 64 - therefore smaller energy production per unit of water) and reduced security of supply, as discussed later. However, in spite of lower energy production, there is a 23% reduction in the average lifetime energy cost compared with the original Phase 1 design. The capital cost estimates presented in Table 3.7 allow for two 23.5 MW generating units in the first phase of the Reduced Project, thus improving supply reliability and providing better matching to the forecast system power requirements. Table 3.7: BUMBUNA HYDRO PROJECT Annual Energy Capital Average Installed Cost c/ Cost of Capacity Firm Average 1986 Energy a/ (MW) (GWh) (GW1h) (US$ m) (USf KWhj Phase I b/ 53.4 206 350 159.05 7.7 Phase 2 80 368 534 134.60 9,2 Phase 3A 125 718 886 94.79 7.4 Phase 3B 185 1 148 1 212 80.59 6.4 Phase 4 305 2 248 1 458 91.71 6.4 Total Project 305 1 248 458 560.74 6.4 Reduced Project Phase 1 b/ 47 157 290 98.47 5.9 d/ e/ Present value of capital and operational cos5S per unit of energy generated on average. Estimate based on 12% discount rate, 50 year life, and 1986 costs, does not include distribution costs. b/ Excludes S22.7 million for works already completed. c/ Cost figures are in financial terms and differ from the economic cost estimates given in Annex '4, Figures are inclusive of transmission costs. d/ Note that average costs are between one-third ano one-fourth the level of estimated LMRC given in Table 3.5. The divergence is partly due to the differences between marginal and average costs. In addition, the average cost estimate does not take account of distribution costs, as wel l as the technical and non-technical losses of the power utility, which are expected to remain relatively high even afier efficiency improvements have been implemented. Source: Bumbuna Feasibility Reporls, 1980 and 1984 Supplement, and mission calculations.  - 65 - 3.55 The Problem Relating to the Bumbuna Scheme. The hydroelectric potential of Sierra Leone was examined in a report for the United Nations prepared in 1971. 10/ A total 22 potential sites were identified and preliminary designs and cost estimates prepared. The 1971 Report recommendations identified three sites as worthy of detailed investigation. These were: Bumbuna Falls (72 MW) 40 km north east of Makeni Benkongor Falls (15.2 MW) 40 km west of Koidu Goma (18 MW) 30 km north of Kenema 3.56 Bumbuna, designed to supply the Western Area, the missing load, and towns along the transmission route to Freetown, was identified as the most attractive scheme and a full feasibility study was prepared in 1980 which recommended an estimated 305 MW project to be developed in five phases. This scheme involved an additional upstream reservoir not envisaged in the 1971 UN report, thus allowing much greater output. Subsequently, work commenced on Phase 1 of the project with contracts let for the construction of camps and access roads and the main diversion tunnels; this work was completed in 1985. Meanwhile, the Bank prepared a report expressing reservations about the economic viability of Bumbuna against the background of the depressed Sierra Leone economy. 11/ As part of the ensuing discussions, two main questions were raised by the Bank: the ability of the country to finance a project costing US$165 million (Phase 1 cost at 1980 prices, excluding escalation and interest during construction, including cost of civil, works now completed) and the high specific cost of Phase 1 relative to the subsequent phases of the project. 3.57 The first phase of the project was then scaled down to 47 MW (from 71 MW) in the form of a Reduced Project at World Bank request. The aim was to minimize capital expenditure wit},out detracting from the tech- nical viability of the subsequent development. However, even the Reduced Project has been deferred because of continuing doubts abotit its viabil- ity, timing, aid affordability by the country. 3.58 Opti,ons Evaluated in the Analysis. An evaluation of thL least cost plan has been made by making a comparison of the most favorable schemes to supply the future demand for power in the Western Area under each of the following options: 10! The strengthening of the Sierra Leone Electricity Corporation, November 1971, Motor-Colombus. 11/ Macroeconomic impact of the Bumbuna Hydroelectric Project, November 22, 1983.  - 66 - - pure thermal scheme; - full Bumbuna development; - reduced Bumbuna Phase 1, followed by pure thermal; and - Bumbuna Phase 1, followed by pure thermal 3.59 The analysis attempts to identify the scheme that will offer the greatest likelihood of providing the required amount of power at least-cost, taking account of a range of load growth and oil price assumptions. Internal rate of return (IRR) calculations which were not carried out because of time constraints should be done as a follow-up exercise. A second point to be made is that the optimal development sequence, following the Reduced Bumbuna Project, has not been worked out by the project's consultants. In this analysis, it has been assumed that it will be followed by thermal schemes. For the sake of simplicity and to make a comparison easier, a fourth option was introduced with the original Phase 1 design of the Bumbuna scheme followed by a thermal development. 3.60 The Procedure for Comparison. For each option, an optimal sequence of new plant additions "plantings") was arrived at to meet the growth in ;>mand for power under the AG scenario to the year FY2020. The planting sequence under each option was adjusted, using a computer simulation of the system's operations, so as to minimize the present value of total capital and operating costs subject to the satisfaction of certain reliability conditions for the power supply. 12/ The basic optimization was carried out at a discount rate of 12X and using the Bank's internal base price projections for oil and coal prices (Annex 3.9). 3.61 The above exercise was repeated for demand growth under the BC scenario. In this manner for each option, two optimal development paths were derived, one under AG and the other under BC conditions. A sensitivity analysis was also carried out to see how the discounted value of total costs for each path would vary with an alternate oil price development (see Annex 15) and different discount rdtes. 3.62 Economic Comparison of Options. The results of the optimization exercise are shown in Table 3.8 below. The figures reveal that the Reduced Bumbuna scheme (followed by thermal plants) is economically the most attractive of the four options examined. Out of the 16 combinations of oil price, discount rate, and load growth assumptions examined, the Reduced Bumbuna is the least-cost solution in twelve cases and does not deviate significantly from the least-cost 12/ The minimum reserve plant margin, designed to give reasonable supply security, was the use of the largest unit on the system plus 15% of the demand for the thermal scheme and largest unit plus 5% of demand for the more reliable hydro schemes. An average was used for mixed schemes.  - 67 - option in the other four cases. The largest deviation from least-cost occurs at the highest discount rate and under the alternative low oil price evolution. It is interesting to note, however, that the full Bumbuna development is economically inferior to the pure theralal option (in 12 out of 16 cases). This, in part reflects the high capital costs for the full Bumbuna Scheme as well as the use of coal as a low cost thermal generation option. A discount rate of 8% or less would be needed to make the full Bumbuna option attractive. More details of the analysis including planting options and estimates of costs are given in Annexes 13 and 14. Table 3.8: COMPARISON OF BUMBUNA OPTIONS (present value in millions of USS, of capital and operating costs estimated at 1986 prices) Accelerated Growth Base Case Discount Base oil Alt. Base oil Alt. oil Option Rate price price price price Full Bumbuna Project 8 300 295 162 161 10 237 234 127 125 12 193 191 100 99 14 161 160 81 80 Thermal 8 339 303 168 158 10 238 215 116 109 12 174 158 83 78 14 130 119 61 57 Reduced Bumbuna Project a/ 8 303 (3) 280 b/ 130 b/ 123 b/ 10 221 b/ 207 b/ 96 D/ 91 bl 12 168 b/ 159 (1) 74 b/ 71 b/ 14 132 (2) 127 (8) 60 b/ 56 b/ Phase I of Full Project 8 329 308 148 142 10 246 234 114 110 12 193 185 90 87 14 156 151 74 71 a/ Figures in brackets give the increment above the least cost solution. b/ Denotes where the Reduced Bumbuna Project is the least cost. Source: Mission calculations. 3.63 Effect of Investment Options on the Balance of Payments. The relative effect on the balance of payments of three of the four main options (the pure thermal scheme, the full Bumbuna option, and the Reduced Bumbuna scheme) over the period 1990 to the year 2000 are shown in Annex 16 for both scenarios, AG and BC. In assessing the possible impact on the foreign exchange balance, average loan terms of 12 years repayment period and a 10% interest rate have been used. The outflows for servicing the loans in the case of the two Bumbuna options are partially offset by the gains from the decline in oil imports due to the use of the hydroelectric schemes. The results show that, over that  - 68 - period, the Reduced Bumbuna scheme will lead under the AG scenario to greater net outflows than the pure thermal scheme but the level of excess outflows is relatively small--on average $1-2 million per annum--in the early years when the amortization burden is assumed to be high based on a 12 year loan and could be treated as an insurance premium for greater supply reliability compared to oil imports. It should be noted that the balance of payments impact for the Bumbuna schemes will be more favorable with better financing terms than assumed in the analysis (i.e., 10% rate of interest with repayment over 12 years). The results under the BC scenario are broadly similar. 3.64 Financial Constraints and the Impa-t of the Bumbuna Scheme on the Sierra Leone Economy. The Bumbuna scheme represents a "lumpy" or relatively large investment for the Sierra Leone economy and has in the past, as discussed earlier, raised legitimate concern about its likely macroeconomic impact. At the macroeconomic level, there are three specific issues relating to Bumbuna: (a) the feasibility of financing such a lumpy investment from the limited foreign and budgetary resources available to the Government, given other development priorities; (b) the ability of the economy to service the debt arising from the investment; and (c) the secondary or general equilibrium consequences of the invest- ment. A thorough evaluation of these questions, while essential for establishing the priority and timing of the scheme, is outside the scope of this report. Nevertheless, based on available information, some observations are given below to place these important concerns into context and indica,te directions for future work on these issues. 3.65 The first point to be made is that the scaling down of the project from its original design will significantly reduce the potential macroeconomic impact. The Bank's earlier reservations about the project, detailed in its 1983 report on the subjoct as 'iscussed above, related to Phase 1 of the full Bumbuna scheme, costing approximately S220 million in 1982 dollars. The reduced Bumbuna project, after taking account of work already completed, is estimated to cost about $100 million at 1986 prices, less than half the amount of the original scheme. The analysis of the previous section suggests that the balance of payments impact of the reduced scheme is not significantly different from that of the alternative, less lumpy investment in a number of smaller thermal generation facilities, after account is taken of the savings in oil imports. Hence the debt service burden would not bc greatly improved by avoiding investing in the scheme. 3.66 Besides the external account, Bumbuna will increase aggregate demand and add to the Government's fiscal burden, but these effects appear to be relatively weak. The local component of the reduced Bumbuna investment is estimated at Le 204 million at 1986 prices (or $13.9 million at the average exchange rate of Le 14.7/$). It is spread over six years, and peaks in the fourth year at Le 50.4 million (see Annex 9, page 3). Under the AG scenario, where coristruction begins in 1988, the peak expenditure would be incurred in 1991. Based on latest internal Bank projections, the domestic expenditure on Bumbuna in 1991  - 69 - will add to the fiscal burden by an amount equivalent to 4.3Z of total Government revenues, including grants, for that year (estimated at Le 1,185 million, at 1986 prices). This percentage assumes that all local expenditures are financed by Government revenues and not from foreign borrowing, which would lower the fiscal burden. Moreover, the financial rehabilitation of NPA would enable it to fund a substantial part of the local costs of Bumbuna from internally generated revenues. Local expenditures on Bumbuna will also add to aggregate demand but, here again, the size of the impact appears to be modest: increased expenditures during the peak period, 1991, are less than 0.J% of GDP, equivalent to Le 10,453 million at 1986 prices (internal Bank projec- tion). These effects will be even smaller if the construction of the scheme is delayed as is the case in the BC scenario. 3.67 The second point concerning the macroeconomic impact of Bumbuna is that, in spite of the scaling down of the project, the "lumpiness" of the project will lead to financing problems: the project would initially require the allocation of a significant amount of scarce foreign exchange thereby crowding out imports needed for other projects. A measure of the difficulty is provided by comparing import requirements during project construction with the Bank's latest projection of the capital inflows to Sierra Leone, in the form of grants and medium and long term loans. Project related imports, estimated to total $84.6 million at 1986 prices, will as in the case of local expenditures, be spread over six years, peaking in the fourth year (Annex 14). Under the AG scenario, imports will peak in 1991 at $24.3 million, in current prices. This expenditure will be financed from a total of only $70 million available from grants and net medium and long term borrowing. New forecasts of capital inflows are unlikely to change the qualitative conclusion that financial difficulties will be encountered if Bumbuna is constructed during the late 1980s and early 1990s. However, the emphasis in Government policy should be on investing in an economically and technically sound project which the country can maintain. The financial problems should be recognized and managed, but should not be the reason for cancelling or inordinately delaying the project. In the long run, efficiency criteria must dominate in project selection. The focus of future work on the issue of financing anu project selection should be on (a): evaluating the economic rate of return (ERR, for the project (the equalizing diLcounL rate for the project--the discount rate at which the Reduced Scheme and its thermal alternative supply electricity at equal cost--lies between 12% and 14%, as revealed by Table 3.8, but is not a reliable measure of the ERR); (b) assessing the magnitude of suppliers' credit for the project which, as it represents additional funds, will reduce the crowding out effect; and (c) evaluating options for spreading out foreign exchange payments and for better management of the financing problem. 3.68 The third and last point to be made is that any analysis of the second round, or general equiiibrium, effects of the Bumbuna investment should take account of changes currently taking place in economy. For example, the shift to a system of flexible exchange rates, if fully implemented, will reduce, if not eliminate, the adverse effects on  - 70 - exports from any price increcses that result from domestic spending on the scheme; in a system of fixed exchange rates, rising domestic prices of traded goods would reduce exports and increase imports by making locally produced goods less competitive. At the same time, structural changes to the economy, as envisaged in current negotiations at the Bank, will help to reduce wage rigidity and dampen what is, potentially, the most important second round effect: the impact on the price of labor, both skilled and unskilled. The increased flexibility in labor markets will temper any increase in wages that occur as a result of increased demand for unskilled labor arising from the project. The effect on the demand for skilled labor and hence on its wages would also be reduced by the use of expatriate staff by contractors. Further, in the longer run, Bumbuna will reduce the demand on skilled labor, as it would be relatively easy to maintain, compared to coal or oil-fired power generation stations. Taking these factors into account, it is by no means clear that the secondary effects of the reduced Bumbuna scheme would produce significantly adverse consequences for the economy. 3.69 Shortcomings of the Reduced Bumbuna Design. There are two main shortcomings in the design, as it currently stands. Firstly, in introducing economies, the risk of supply interruptions has increased; double circuit transmission lines have been replaced by a single circuit line, and a single generating unit has been introduced in place of the two Francis units, in the original Phase 1. At the same time, bottom outlet gates have been eliminated at the end of the spillway tunnel so that, apparently, there i, -) back up system to control the flow of water in the tunnel if the inlet gates were to malfunction. Secondly, the optimal development sequence for Bumbuna, following tt.e Reduced project, needs to be worked out; in this analysis, it was assumed that there would be no further development of Bumbuna. Both these shortcomings need to be studied before a final decision on Bumbuna is made. 3.70 Interconnection with Neighboring Countries. One of the problems with the full Bui.buna Project is the high cost of the first phase, relative to the amount of firm (or dependable) energy available. In the second phase, the construction of the Yiben storage reservoir increases the firm energy available from 206 GWh to 368 GWh. An interconnection with, say, Guinea would make available a larger market for the output of Bumbuna thus providing two benefits: (a) the energy available during the wet season, which could not be absorbed by the NPA system, could be absorbed by Guinea, provided the Guinea system does not also have a surplus of hydro energy during the wet season; and (b) Phase 2 and subsequent phases of the Bumbuna Project could be brought forward, thus enabling the lower energy costs to be realized earlier (see Table 3.7). 3.71 A transmission interconnection with Conakry would involve, as a minimum, a single-circuit 132 kV line from Lunsar, via Kambia, a distance  - 71 - of some 190 km. The approximate capital cost of the interconnection would be US$10 million, or an annual charge of US$1.4 million. 13/ In order to justify tha interconnection by fuel savings to Guinea, this would require a minimum annual transfer of about 40 GWh fron Bumbuna, assuming replacement of diesel generation using heavy fuel. Under Phase 1 of both the full and reduced Bumbuna options, there is surplus hydro energy during the wet season (but not during the dry season) in the ini- tial years. Under the AG scenario, the surplus energy figures are given in Table 3.9. On the basis of these figures, an interconnection with Guinea would appear to be worth more detailed study. If the surplus energy could be sold to Guinea, with suitable take or pay contracts, then the economics of the Full Project will become more attractive. Table 3.9: BUMBUNA SURPLUS ENERGY IN WET SEASON (AG SCENARIO) Fiscal_Year Full Bumbuna Reduced Bumbuna (GWh) (GWh) 1993 107 73 1994 88 54 1995 66 34 1996 44 17 1997 28 7 1998 49 0 1999 21 0 2000 2 0 Development_Options for Provincial Load Centers 3.72 The Role of Small Hydro Schemes. In addition to the 1971 UN study on the hydroelectric potential of Sierra Leone, two other studies have been carried out in recenr years. Details of the schemes identified in these stud. - are summarized below in Table 3.10. The schemes are shown in the n,. at the end of the report. The majority of these schemes are run of the river and will require supplementary diesel generation during the dry season. In these circumstances, the question arises as to whether any of these schemes could be economically developed to supply residential and mining Loads in the Provinces and eventually form part of an integrated national or regionaL work or network. 13/ Assuming 30 year life, 12% discount rate and O&M costs equivalent to 1.6% of the capital cost.  - 72 - Table 3.10: DETAILS OF SMALL HYDRO SCHEMES Scheme Installed Capital Equalizing Type Location Capacity Cost a/ Cost Discount Rate (MW) (USSa) (USS/kW) Gbangbaia Moyamba 1.0 3.4 3 410 10.4 b/ (Run of River) Gandorhun BO 20.0 65.4 3 270 10.0 b/ (Run of River) Mawaloko Kabala 0.5 d/ 3.2 6 420 9.3 bl (Rur. of River) Benkongor Falls Kono 10.8 22.0 2 040 29.5 c/ Betmai I'alls Magburaka 3.6 13.5 3 750 12.8 c/ (Run of River) Singimi Falls Moyamba 7.2 28.0 3 890 20.8 c/ (Limited Storage) Kambatimbo Falls Komadugu 0.8 6.6 8 250 5.2 c/ (River Diversion) a/ 1985 Costs. b/ Return against diesel alternative; no real increase ii fuel prices. c/ Return 3gainst diesel alternaive; 2% then 4% (after year 5) escalation of all costs. d/ Installed in stages. Source: CEDSI/CANREDE Ltd., March 1983; SNC, December 1985; and mission calculations. 3.73 An appreciation of the problem facing most small hydro sites in Sierra Leone is obtained from the experience of the Goma scheme identi- fied by the original UN study and being constructed by Chinese help. The site at Coma, north of Kenema, is being developed as a 4MW station to feed Kenema and local villages initially, but will also supply Bo when the Bo-Kernema transmission project is completed. At Coma the dependable output is only 0.5 MW during the dry season compared with the 4 MW installed capacity. As a back up tor the dry season, a 5 MW heavy fuel oil plant is to be constructed at Bo. The 4 MW of hydrogeneration available from Goma for most of the year will take economic preference over oil generation. The mission estimated that the load factor (utilization factor) for the fuel oil plant will be of the order of 10%, hardly high enough to justity the additional capital and operating costs, and tht transportation difficulties associated with transporting fuel oil from Freetown. 3.74 No details of the economic evaluation of the Coma project were available, but Table 3.10 gives an initial evaluation of the viability of  - 73 - other schemes. The two schemes that appear most attractive are the Ben- kongor Falls (10.8 MW) and the Singimi Falls (7.2 MW). The equalizing discount rate (rate above which thermal options are of lower ccst) for these two projects is above 20x. This is partly because they are not run of the river and partly because large mining loads nearby would absorb most of the output of the schemes from the outset (gold and diamond mines in the case of Benkongor and rutile and bauxite mining in the case of Singimi) giving themr a high load factor. The total cost of the two projects is an estimated $50 million and their capitalized value, including fuel and maintenance should not exceed $56 million. Benkungor would also produce more energy than local demand and could be either scaled down or connected with a national grid in future. Both schemes anpear attractive, especially as they will save on diesel for powel generation currently being imported by the mines, and their feasibility should be reviewed at an early date. 3.75 Interconnection of Isolated Power Plants. Th. numerous isolated power systems in the Provinces impose a considerable strain on the resources of NPA in terms of technical staff, maintenance, fuel supplies, and administration. NPA is fully aware of these problems and is committed to interconnection where it is technically cnd economically feasible. A more detailed policy needs to be identified for future interconnections to prevent haphazard development. In principle, interconnections should be made whenever they become economically feasible. However, the design of interconnection schemes should conform to an agreed national standard. 3.76 The Marampa iron ore mine at Lunsar, which is now closed, has a total of 10.7 MW of medium speed diesel generating plant lying idle. At present, the plant is in good condition but it will deteriorate if it is not used. In view of NPA's current capacity shortages, it is suggested that discussions be held with the mine authorities with a view to purchase of power either on a regular or standby basis. The mine could be connected to the Western Area grid by advancing the construction of the section of the Bumbuna-Freetown transmission line between Freetown and Lunsar. 3.77 Supplies from Neighboring Countries. Should the Mano River Project be constructed, eventually there might be a case for interconnec- tion of the Sierra Leone and Liberia systems. However, different system frequencies, and the relatively large distances from the Mano project site to Freetown, would tend to negate the case for full interconnec- tion. If the project is constructed initially to serve mainly Liberia, a small part of the output could supply the more remote southern parts of Sierra Leone, which are a long way from the Western Area grid. Recommendations 3.78 The analysis carried out in this Assessment indicates that the Reduced Bumbuna project is attractive, relative to alternative thermal options over a wide range of load forecast and fuel price assumptions.  - 74 - However, there was insufficient data to decide on the optimum design of the project in terms of sequencing (i.e., development strategy to follow the reduced version of Phase 1). At the same time, there are questions relating to the physical design which need to be addressed (for example, the need for extra gates) before the project is taken further. A number of interesting small hydro schemes have elso been identified which may meet a high proportion of the demand of the mining sector with consequent substantial savings in fuel imports. The implementation of some of these hydro prospects may make interconnection of a number of Provincial Sys-ems economic. However, the viability of the identified small hydro schemes and possible regional interconnectiors need further investigation. 3.79 The mission recommends that NPA should address the main unanswered question raised above before making a final decisicn with regard to Bumbuna and the development of Provincial Systems. The following steps are recommended for identifying a least-cost strategy for meeting future power demand in Sierra Leone: (a) prepare , comprehensive load forecast for the whole country, including a survey of the needs of auto-producers; (b) determine the optimum development sequence of the Bumbuna project and finalize the design of Phase 1; (c) investigate the prospects for export of surplus energy from Bumbuna to Guinea and the benefits of mutual power and energy exchanges on a firm or standby basis; Cd) in conjunction with (b) and (c), determine the viability and optimum sequence of development of the small hydro schemes at Singimi Falls and Benkongor Falls (examine also the possibility of joint ventures with mining companies); 'e) examine the possibility of interconnection of isolated Provincial Systems where it is economic to do so and outline technical standards to be adopted in interconnection schemes. (f) review the generating capacity of the major mines and evaluate tile feasibility of using unutilized capacity to supply local demand; and (g) identify optimum transmission and distribution development, taking account of the least-cost generation development and the findings cf the proposed loss reduction study,  - 75 - 3.80 In addition to the above issues, NPA should: (a) investigate the feasibility of an interconnection between the Marampa mine and the Wes*.ern Area system to utilize the idle generating capacity at the mine; and (b) reconsider the location of the proposed heavy fuel burning diesel station at Bo. Investment Trends 1981-1986 3.81 NPA.'s investment u 3gram over the past five years has been severely limited because of a chronic shortage of foreign exchange, a lack of internally generated investment funds because of inadequate tariff levels, and the curtailment of the Bumbuna ptoject, which was to have been commissioned in 1986/87. The total investment over the period FY81 to FY86 is estimated at approximately US$45.9 million, made up as shown in Table 3.11. The figure includes Government investment in electricity which has been passed on to NPA, either as a direct grant or in the form of equity. Table 3.11: ESTIMATED INVESTMENT IN NPA ASSETS FOR FY 1981-86 item Cost (USS million) Bumbuna Project 22.7 Kingtom Cooling System 1.0 Kingtom Rehabilitation a/ 4.2 Kingtom Mcoile Diesels 5.0 Goma Hydro Project 11.4 Provincial Generation 1.3 Training and Miscellaneous 0.3 Total 45.9 a/ Includes some transmission and distribution work. t/ Based on 2,132 kW of capacity added s:nce FV8I at USS60/kW. Source: NPA, Government accounts, mission estimates.  - 76 - Table 3.12: ESTIMATES OF NPP INVESTMENT REQUIREMENTS FY87-91 BASE CASE SCENARIO (1986 USS million) Fiscal Yea- 1987 1988 1989 1990 1991 rota BoAKenema Project 1.80 3.60 3.60 9.00 Kingtom rehabilitation 0.25 1.18 1.87 0.55 1.33 5.18 Kingtom spare parts 1.58 1.58 Transmission and distribution 0.9Q 0.94 0.47 0.47 0.46 3.28 Provincial systems 0.50 0.50 0.50 1.50 General plant 0.11 0.11 0.11 0.11 0.44 Master Plan Study 0.10 0.10 0.20 Tariff studies 0.02 0.10 0.12 Engineering studies 0.18 0.28 0.46 Technical assistance 0.12 0.67 0.38 1.17 Communications 0.11 0.13 0.43 0.75 1.42 Loss reduction study 0.10 0.10 Total 5.19 6.82 7.18 2.23 3.03 24.45 Note: Further investment in Bumbuna not needed until FY*2. Source: World Bank and mission estimates. Table 3.13: ESTIMATES OF NPA INVESTMENT REQUIREMENTS FY37-93 ACCELERATED GROWTH SCENARIO (1986 USS million) Fiscal Year 1987 1988 1989 1990 1991 1992 1993 Total Bumbuna project 1.79 22.80 20.13 23.81 23.56 6.38 98.47 Bo/Kenema project 1.80 7.20 9.00 Kingtom rehabilitation 0.25 1.18 1.87 0.55 1.33 5.18 Kingtoin spare parts 1.58 1.58 Trarsmission and distribution 1.88 1.40 2.68 2.63 1.33 9.92 Provincial systems 0.50 0.50 0.50 1.50 Ge;leral plant 0.22 0.22 0.40 0.65 1.49 Master plan study 0.10 0.10 0.20 Tariff studies 0.02 0.10 0.12 Engineering studies 0.18 0.28 0.46 Technicni assistance 0.12 0.67 0.38 1.17 Communications 0.11 0.13 0.43 0.75 1.42 Loss reduction study 0.10 0.10 Tctal 6,13 12,78 28.70 24.81 28.25 23.56 6.38 130.61 Note: Investment requirements after FY91 not identified, apart from Bumbuna. Bumbuna costs relate to Reduced Project (Phase 1). Source: World Bank and mission estimates.  - 77 - Future Requirements 3.82 NPA's investment requirements in the short term will cover the rehabilitation of the generating plant at Kingtom and Falconbridge, togetner with a restocking of spare patcs for both the Western Area power stations and the Provincial stations. A limited amount of funds are required to provide technical assistance to assist in the commissioning of the new diesel units at Port Loko, Makeni, and Koidu, and for the repair of the damaged alternator at Makeni. Tables 3.12 and 3.13 show the investment requirements under the AG and BC scenarios. Under AC, investment requirements 'or he Reduced Bumbuna Project are included as it is assumed to be commissioned by FY93 in that scenario. Under 1C, the Bumbuna scheme is commissioned in 1997 and no expenditures are envisaged before 1991. Investment will also be required in transmission expansion and distribution rehabilitation, in order to meet the forecast load growth under both scenarios. Some of the investments have already been identified under a power system rehabilitation project to be funded by the World Bank and the European Economic Community. The remaining items cover technical assistance for training, engineering studies, purchase of a computer billing system, and a communications system between Freetown and the Provincial centers.  - 78 - IV. WOODFUELS/HOUSHHOLD ENERCY Reasons for Concern 4.1 Available data on household c,nsumption patterns reveals a high dependence on fuelwood, which nakes consumers vulnerable to increases in its price, and, at tuie same time, creates pressures on the country's woodfuel resources. Household energy consumption is about 85 of total final energy use in the Sierra Leone economy. About 97% of this consump- tion is supplied from woodfuel resources which dominate household energy use in both rural and urban areas. At the moment, woodfuel supplies at the national level appear adequate in physical terms to meet an estimated demand of 3.9 million cubic meters per annum, but there are growing signs of regional shortages near urban areas; prices of fuelwood in Freetown have gone up by 150% in real terms betwee., 1983 and 1985 and are now a; levels ($10-15/ton at a Le 25/$ exchange rate) comparable with neighboring cointries, such as Ghana, experiencing acute shortages near urban areas. Increases in fuelwood prices have contributed to higher energy costs for a majority of urban households. Based on a limited consumer survey in Freetown, the mission found that energy expenditures absorb as much as 40% of the income of poor urban households (Table 4.1). Table 4.1: HOUSEHOLD ENERGY EXPENDITURES AS A SHARE OF MONTHLY INCOME Monthly Income Class Fuelwood Charcoal Kerosene Electricity Total 0-300 Le 22.3% 3.2% 9.2% 4.9% 39.6% 301-500 Le 8.2% 2.2% 7.6% 2.6% 20.6% 501-800 Le 8.3% 1.5% 4.6% 2.3% 16.7% + 800 Le 0.9% 2.6% 3.7% 3.0% 10.2% Source: Mission estimates from survey of 128 households (details in Annex 18). 4.2 There are two main issues regarding household energy in general and woodfuels in particular: the first worry concerns the affordability of household energy supplies; the second pertains to the sustainability of current patterns of woodfuel consumption over t . next ten to fiiteen years. The affordability issue arises because of the mission's findings, mentioned above, that a significant proportion of the income of urban households is being allocated to energy purchases and at a level which is not sustainable. These results, while based on a small sample, are con- sistent with the findings of earlier studias (see Annex 18). At the very least, they are indicative of a growing and serious problem and underline the need for the Government to give greater attenticai to this problem and consider policy options for rrducing household energy costs. This issue is particularly relevant at a time when the Governmern has significantly increasea the prices of petroleum products and power supplies.  - 79 - 4.3 The second worry relates to the likelihocd and nature of the fuelwood crisis over the next ten to fifteen years if the present pace of growth in woodfuel demand is sustained. Very little work has been done on existing consumption patterns and potential supplies of woodfuels. Lack of both anal sis and reliable data makes it too difficult to assess whether the existing regional shortages would develop into a major fuelwood crisis in the foreseeable future, and if so what policy options are available to prevent such a crisis. 4.4 This chapter is composed of four sections. The first two sec- tions provide the framework for discussing the above concerns which are closely interlinked and cannot be evaluated independently. Section one examines the characteristics of household energy use in order to throw some light on the factors underlying high energy costs and th. choicC between fuelwood, charcoal, and i,iodern fuels. Section two focuses on the supply and demand for woodfuels and reviews the problem of regional scarcities and the supply side factors, including institutional weak- nesses, which have contributed to the increased prices of fuelwood in urban markets. The third section considers a possible woodfuels and household energy strategy for addressing the affordability and sustain- ability issues as part of an integrated approach for providing household energy at least cost. The final section outlines an investment and tech- nical assistance program for implementing the proposed strategy over the next Live years. Characteristics of Household Energy Use 4.5 Pattern of Household Consiumption: The Dependence on Fuelwood. As shown in Table 4.2, 98% of household consumption is primarily for cooking and the remainder is for lighting or cooling. Woodfuels supply 97% of total consumption, with fuelwood being the overwhelming choice for household cooking, both in urban and rural areas. 4.6 While the picture conveyed by Table 4.2 is generally valid for all parts of the country, there are some significant differences between the pattern of urban and rural consumption. Fuelwood remains a "free goo,i" for the vast majority ot rural households in Sierra Leone, with most families gathering their wood requirements from local forest resources in the process of clearing land for slhifting agriculture. Per capita fuelwood consumption in the rural areas is estimated to be of the order of 600 kg per annum and accounts for virtually all household energy use. In the urban areas, such as Freetown where there is a price and a flourishing market for fuelwood, per capita use declines to an average of about 400 kg per annum. In the case of other fuels, the rural-urban contrasts are greater. The estimates for annual per capita consumption of other fuels in urban and rural areas are respectively: charcoal, 50 kg and 10 kg; kerosene, 18 liters and 5 liters; electricity, 55 kWh and 5 kWh. These estimates are within the range experienced in other West African countries. The most notable exception is the low level of market  - 80 - penetration by charcoal in urban areas; for example Freetown consumpLion estimates of 50 kg per capita are well below estimated urban area use in Liberia (141 kg) and Senegal (230 kg). Table 4.2: ESTIMATED HOUSEHOLD ENEFGY CONSUMPTION, 1986 ('000 toe) Cooking and Lighting and Ironing cooling Total (%) Fuelwood 728 -- 728 (92%) Charcoal 38 -- 38 (4.8%) Kerosene 4 17 21 (2.7%) Electricity - 3 3 (0.4%) LPG 1 __ I (0.1%) Total (%) 771(98%) 20(2%) 791 (100%) Source: Mission estimates and earlier surveys arranged by University of Sierra Leone. 4.7 The relatively low levels of urbanization in Sierra Leone means that only 15-20% of national fuelwood consumption is commercialized; the remainder is not traded. Conservation and substitution measures need to differ between the areas where fuelwood is a traded good and those where it is not. In ateas where fuelwood is traded, the key factors affecting future demand for fuelwood (and with it household energy costs) are: (a) the retail price of fuelwood compared with other household fuels; (b) the efficiency and costs of cooking stoves for wood and other fuels; and (c) household income and purchasing power. 4.8 Relative Prices, Inefficient Stoves, and the Comparative Cost of Using Alternative Fuels. In spite of the recent sharp increases in the prices of fuelwood and the low efficiencies of existing stoves, fuelwood remains competitive with respect to other household fuels. Table 4.3 shows that, at the time of the mission, the marginal cost of cooking anergy delivered to the pot was approximately the same for wood and charcoal but higher for other fuels: electricity and kerosene (pur- chased by the "pint" from street venders) were between 35% to 40% more expensive than fuelwood purchased in bundles. These differences will be magnified by the recently announced large increase in the price of kerosene and electricity. (Kerosene purchased by the gallon at subsidized official prices has the lowest financial cost per unit of utilized energy in Table 4.3. However, this option was not available to most householders due to the widespread shortages of petroleum produk.ts sold at official prices. The increase in the official prices of kerosene would, in any circumstances, remove this anomaly.) There is, therefore, no incentive to switch away from woodfuels for cooking until these rela- tive costs shift in favor of modern fuels, or, given consumer preference for the convenience value of modern fuels, the cost differential is at  - 81 - least significantly eroded. This appears unlikely in the foreseeable future with the Government committed to a policy of full-cost pricing and the long run costs of both petroleum products and electricity expected to stay at or above current prices in Sierra Leone. Consequently, the use of kerosene and electricity will be largely restricted, as at present, to lighting needs except in the higher income brackets. Table 4.3: COMPARATIVE CvSTS OF COOKING FUELS IN FREETOWN, FEBRUARY, 1986 Costs per unit EnLrgy Cost per unit of Conversion Cost per unit of Fuel purchased content energy purchased efficiency energy utilized (avg) (MJ/unit) (SLO/MJ) I%) (SLIZ/MJ) Fhelwood Split bundles 0.32 Le/kg 15.1 MJ/kg 2.1 11 18.6 Whole bundles 0.27 Le/kg 1.8 15.9 Charcoal Small bags (2 kg) 1.03 Le/kg 27.5 MJ/kg 3.75 21 17.8 Large Bags (15-22 Kg) 0.31 Le/kg 3.0 14.2 Heaps b/ 1.57 Le/kg Not Av. Kerosane lmperiai gallons 4 Le/gal 159.1 MJ/gal 2.5 35 7.1 of pisfS a/ 12 le/gal 7.5 21.4 Electricity 0.49 Le/kWh 3.6 MJ/kWh 13.6 60 22.7 a/ Recycled beer bottles, approx. 13.7 "pints" per imperial gallon. b/ Used primarily for ironing. Source: Mission estimTates. 4.9 The importance of conversion efficiency for the competitiveness and the costs of traditional fucls is illustrated in Table 4.3. In the table, fuelwood is by far the cheapest fuel in terms of costs per unit of energy purchased but is almost twice as expensive to use as kerosene pur- chased at official prices due to low conversion efficiencies. Experience in othe; West African countries suggests that the efficiency of the traditional three-stone fuelwood stove (or, more accurately, pot support for cooking over an open fire) is about half the level of available improved wood stoves. The introduction of such stoves would signifi- cantly cut back energy costs and reduce the amount of fuelwood used for cooking; if fuelwood prices were to continue to increase then the improved stoves would at least prevent an increase in the already high burden of energy costs to poor households. 4.10 While consumer preference may exert a decisive influence, the lack of market penetration by charcoal can partly be explained by its relatively high user cost. Charcoal is less costly to use than fuelwood only when purchased in bulk (large bags), an option not open to a majority of consumers because of cash flow constraints. An additional  - 82 - factor which may deter consumers from using charcoal is the cash outlay needed for purchasir. charcoal stoves. 4.11 Household Incomes and Energy Costs. In the event that improved woodfuel stoves fail to penetrate the fuelwood deficit urban areas in sufficient numbers, the real cost of cooking energy is likely to continue to rise as a result of increasing fuelwood prices (see discussion in next section). The mission's household survey results suggest that a majority of urban households already spend between 20% and 40% of their incomes on energy purchases, mainly on fuelwood. Thus, most households do not have the purchasing power to maintain their fuelwood consumption in the face of rising fuelwood prices. The only alternative is less cooking, i.e., a painful reduction in basic welfare. 4.12 In urban areas, it is evident that households would be prepared to spend a high proportion of any additional i.icome on increased cooking. The introduction of more efficient stoves would have the same effect as reducing the price ot wood. The incomre mad( available by this price change will, to some extent, be spent on increased cooking. Improvement in combustion efficiency, therefore, may not result in a fully commen- surate reduction in fuelwood consumption. While the full cost reduction represents a very real improvement in basic household welfare, it is essential that deforestation projections are not rigidly linkeci to improvements in combustion efficiency. 4.13 Taxes and Subsidies. Fuelwood prices are determined by the cost of cutting and transport and do not include a component for the economic cost of the wood itself, i.e., the cost of reafforestation or, in its absence, the cost of deforestation in terms of decreasing soil fertility. While higher taxes on fuelwood would be justified in economic terms, enforcement problems nmoy well prevent the delivered price being raised up to the level of the full economic cost of supply. Fuelwood consumers will, therefore, continue to be indirectly subsidized by a price distortion. This economic fact of life should be explicitly recognized in any discussion of subsidies on stoves or reduced taxes on alternative cooking fuels (kerosene). Woodfuel Supply and Demand 4.14 Supply and Demand at the National Level. Sierra Leone has large and diverse forest resources. According to latest estimates 87% of the Sierra Leone's total area (7.2 million ha) is classified as natural forest land. However, only 5% of land area (365,000 ha) is closed high forest, located largely in the remote eastern part of the country but with some reserves in the Western Area peninsula near Freetown. The rest of the country is now covered in secondary forests in various stages of degradation (261,000 ha), forest regrowLh from bush fallow agriculture (3.7 million ha), and savannah woodland (1.7 million ha). There are 286,000 ha of mangrove located near Freetown.  - 83 - Table 4,4: FOREST RESOURCES AND ANNUAL WOOD INCREMENTS Estimated Estimated total Forest/land type Area annual yield annual yield (1000 ha) (m3 'ha) ('000 m) Closed high forest 365 2-4 770-1,500 Secondary forests 261 1 260 Forest regrowth (bush fallow) 3,774 2-3 a/ 7,500-11,0',0 Savannah woodland 1,619 0.4-1.2 600-1,9uO Coastal woodlands (mangrove) 286 0.3-0.6 100-200 Plenta'ions 10 3-15 30-150 Overall tree cover 6,315 9,260-15,010 a/ For most areas. Sour.e: FAO/UNDP/MANR Land Bureaus Survey Report (1979), misszon estimates. 4.15 Table 4.4 presents estimates of the annual increment in forest cover. The total annual increment from all sources is estimated at between 9 and 15 million cubic meters (n, ) with the bulk of supplies coming from rorest regrowth that develops as part of the bush fallow cultivation cycle. Due to the lack of reliable and up-to-date information on forest inventories and yields, especially from bush fallow, the estimated total annual incren,ent is only an approximate measure of sustainable wood supplies. 4.16 Aggregate wood production is well in excess of estimated fuelwood consumption for the country as a whole. As compared 3with an annual increment in the natural forest cover of 9-15 million m , total woodfuel consumption shown in Table 4.5 is around 4 million m3 (or 1.1 m /capita/year), indicating an apparent surplus. 14/. However, in practice, wood available for consumption is less than annual incremental supplies. Wood availability depends on economic access to the resour- ces. In Sierra Leone, significant wood resources are concentrated far from the large consumption centers in sparsely populated areas which lack appropriate access infrastructure to exploit the resources economically. In the case of supplies from bush fallow, economic access is often limited to 2-5 mile distanuces from transport routes and availability is low with current prices ar.d existing transport infrastructure. 4.17 In summary there is a physical surplus of wood supplies at the country level with sustainable physical production exceeding annual con- sumption. However, such a comparison between physical supply and demand 14/ Other uses o roundwood (including poles and sawlogs) are around 0.2 million m and co not materially affect the physical 5urplus.  - 84 - obscures the true extent and nature of fuelwood scarcities, as it does not reveal the extent to which supplies are economically available to meet demand at specific locations. This observation is reinforced by increasing signs of regional scarcities as discussed next. Table 4.5: WOODFUEL DEMAND (1986) Fuelwood Charcoal Equivalen* tuelwood demand ('000 tons) ('000 tons) ('000 tons) b/ ('000 m ) c/ Household a/ Freetown 194 24 437 625 Other areas ,928 32 29 3,213 2,122 56 2,686 3,838 Industry 5 4 42 61 Commer'ce 6 0.5 11 15 Tobacco curing 25 - 25 36 Total 2,158 60.5 2,764 3,950 a/ Annual fuelwood (charcoal) consumption per capita. Freetown 400 kg (50 kg); other areas, 600 kg (10 kg). b/ Efficiency of converting wood to charcoal, 10% on a weight basis. c/ Wood density assumed to be 0.7 tons/m3, air dried. Source: Mission estimates. 4.18 The Emergence of Regional Scarcities: Causes and Trends. In spite of an apparent surplus of physical wood supplies at the national level, there are indications of severe and growin.g regional shortages combined with a deterioration in the size and quality of forest resources. As shown in Annex 19, fuelwood prices in the Freetown area have increased more than 220% in real terms since 1978 and woodfuel merchants can be observed travelling more than 100 miles to Moyamba Junction and even beyond Kenema for supplies. Approximacely 60%-70% of Freetown supplies are now estimated to come from outside the Western Area. Fuelwood shortages, in the form of active markets and rising prices, have also emerged in the tobacco growing areas (near Makeni and Bo) where fuelwood is used by farmers for tobacco curing, as weil as in Kenema and Koidu. The effect of regional scarcities is now being felt in rural districts: wood has become a traded commodity and accounts for a significant proportion of the farmers income (as much as 80% in some villagec) in catchment areas supplying consumption centers. 4.19 There are three main causes of the destruction of the natural forest cover and the emergence of regional shortages: (a) urbanization and over-cutting of wood to supply the large and growing consumption centers; (b) forest clearing and degradation associated with the need for  - 85 - agricultural land; and (c) lack of commitment to tree protection or for planting new trees. 4.20 Bad forestry management practices have been triggered by uncertainty about land-holding rights and, therefore, about the long term ownership of the forest cover. Moreover, the policing role of the forestry services has reinforced the villagers loss of interest in forest cover protection and development. 4.21 As a combined result of these factors which largely derive from population pressures, there is a continuing erosion of forest cover and shortages continue to magnify. In the Western Area region where these shortages are most pronounced, a conservative 3.5% per annum increase in the populatign should raise total w,odfuel demand from its current level of 650,000 mT to about 1 million m in the year 2000. To generate even half the increased amount from the Peninsula area would require new fuelwood plantations, exceeding 10,000 ha, to be added to the Western Regions current forest reserves (around 18,000 ha) whose function is mainly protective and cannot be used to meet future demand. 15/ 4.22 In spite of the uncertainty in these estimates, it is clear that there is an urgent need to initiate measures to improve wood supply and limit demand growth in the Western Area if the shortages are not to magnify two tc threu fold over the coming 10-15 years. Similar measures may be needed in the vicinity of Bo, Makeni, Kenema, and Koidu Town where prices have been rising, but little is known of the supply-demand balance. Househcld Energy/Woodfuel Strategy 4.23 Need for Intervention. The analysis in the previous sections points to the urgent necessity of limiting wood demand growth and improving supplies in regions where significant woodfuel imbalances have emerged, especially in the Peninsular area. Without some form of govern- ment intervention, the costs of energy supplies to urban households will continue to rise in real terms, and the deterioration of the forest cover will accelerate, not only in the deficit areas but the rapidly expanding zones which supply the shortfall. A strategy designed to address these problems should focus on improving the situation in the worst affected areas: the Peninsula, as well as the areas around Makeni, Bo, Koidu and Kenema. There is also a need for an integrated approach to household energy/woodfuel problems: given the country's limited resources, it is both efficient and desirable to allocate available funds between various 15/ Based on a mean annual increment of 18 m3/ha.  - 86 - supply and demand side programs in a manner which maximizes their com- bined impact and encourages the provision of household energy at least cost. 4.24 In formulating a household energy/woodfuel strategy the following components should be included: (a) actions on demand manage- ment; (b) supply side measures; (c) changes in the institutional frame- work; and (d) evaluation of the potential from other energy sources. Management of Household Demand--The Focus on Fuelwood 4.25 Conserving Woodfuels and Reducing Costs Through Improved Wood Stoves. The ,nost common form of cooking is over the traditional open three stone fires, which have poor thermal efficiencies estimated to be in the range from 8% to 15% according to fire tending practices. This method of cooking is widely used in the rural areas and by more than an estimated 60% of Freetown households. In principal, substantial savings of fuelwood appear feasible through the widespread use of low cost, more efficient models of woodstoves. Hence, replacing the three stone stove with woodstove models that are financially attractive to users and burn less fuelwood must be the focus of any household manago